OS2 World.Com Forum

Subject  :  Scitech, SNAP and OS/2...
Author  :  Ben Dragon
Date  :  17 Nov, 2006 on 01:27
Well, I guess everyone has heard the latest news by now. In case you haven't Scitech has ceased development of their SNAP video drivers and is selling the code... all of it... and moving onto other things.

What does this mean for eCS/OS2 users?

Well, I foresee several possibilities:


    1. Someone buys the code and
    a) drops eCS/OS2 support
    b) or continues eCS/OS2 support and either enhances it or not...

    2. No one buys the code and we're left out in the cold or... the least likely...

    3. No one buys the code and they release it as open source.



I'm hoping for the 1b option myself, or, as a distant 2nd, the last option.

But what will we, as a community, do for upcoming video hardware support if we get left out in the cold WRT video drivers?

I don't think anyone would like to go back to the situation that we had in the past prior to Scitech and SNAP, (or SDD).

There is a major new MB setup coming that, supposedly, won't require drivers as we now know it, but that hasn't happened yet and no one knows what's actually going to be in the final production version. But if it does happen as billed it will be OK for us... if and when that happens, but what will be do in the meantime?

WRT to the short term SNAP will work with what's out there, but for the long haul...

Well... I guess we could pass the hat... :\



Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  Blonde Guy
Date  :  17 Nov, 2006 on 06:31
I'd add a possibility.

Someone buys the SNAP source code and makes it open source.

Now think about all of the possible bidders for the the OS/2 SNAP products. What kind of bid would it take to be the top dog?

I think there are OS/2 World bounties bigger than that. We might be the new owners for the equivalent of lunch money.

---
Expert Consulting for OS/2 and eComStation


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  melf mikaelelf@os2ug.se
Date  :  17 Nov, 2006 on 09:06
I'm not familiar with costs of software source code like this. What could one expect the price would be? Is solutions like SNAP, which supports lots of cards, of interest for a grahic card manufacturer or would such a buyer just buy and drop the hole thing of reasons of competition? Could Serenity afford to buy the source code, do they want to, would it be good?

Well, this announcement is a little bit shocking but maybee not unexpected, feels like we are hanging in the air again.

---
/Mikael


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  flywheel flywheel@worldonline.dk
Date  :  17 Nov, 2006 on 14:37
AFAIR SNAP Graphics for eCS is based on GRADD - which AFAIR still is closed source - therefore I do not believe that SNAP will end up as open source.


melf (17 Nov, 2006 09:06):
I'm not familiar with costs of software source code like this. What could one expect the price would be? Is solutions like SNAP, which supports lots of cards, of interest for a grahic card manufacturer or would such a buyer just buy and drop the hole thing of reasons of competition? Could Serenity afford to buy the source code, do they want to, would it be good?

There will be a minimum price, most likely including the GRADD license.

If one manufactorer is to buy it - the larger he is the better. AMD/ATI seems to me to be the perfect place for SNAP to be.

(With S3 and Intel we would only get support for integrated solutions, nVidia already got their wrapper - that works great, Matrox I would not trust to sit the right way on a toilet seat, the rest is to small - since many of us use multiple boot and also have other considerations when buying Graphics cards).

The best solution (For the consumers) would be a consortium of a handfull of manufactorers - bringing support to platforms like eCS, Linux, ZetaOS and so on.


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  Ben Dragon
Date  :  17 Nov, 2006 on 16:53

flywheel (17 Nov, 2006 14:47):
AFAIR SNAP Graphics for eCS is based on GRADD - which AFAIR still is closed source - therefore I do not believe that SNAP will end up as open source.

OK. Just a clarification here. Scitech are not selling just the SNAP for eCS/OS2 code... they're selling ALL the SNAP code... as one complete unit. This includes the Windows and the Linux versions. Only the eCS/OS2 version requires GRADD and I don't believe that Scitech has access to the GRADD sources.

And if what I read on that is true, the eCS/OS2 portion of SNAP comprises the smallest part of their market.

IBM owns GRADD. If Scitech finds a bug they submit a report to IBM and IBM in turn, makes the fixes and gives the updates to Scitech and they, in turn, distribute the fixes with their driver distribution.

I think that it's highly unlikely that the sale will have anything at all to do with GRADD or the licensing of it. Can anyone add anything to that?


flywheel (17 Nov, 2006 14:47):
There will be a minimum price, most likely including the GRADD license.

If one manufactorer is to buy it - the larger he is the better. AMD/ATI seems to me to be the perfect place for SNAP to be.



Not really. There sure as hell isn't any way that nVidia will allow AMD/ATI to have access to their sources given that they kept Scitech so completely in the dark.

So, SNAP would end up being just another ATI graphics driver with support for additional non-ATI video cards that presently exist in SNAP

ATI no longer makes graphics drivers for eCS/OS2 and they haven't for a long, long time. So why should they buy code to drivers for their own cards to work on another OS that they all ready dropped support for?

If memory serves me correctly ATI has some input for drivers that support their cards under Linux, but I'm not sure just how great their involvement in that is, but it think they provide the necessary information specific to each card, to those that develop the driver.

So, I can see AMD/ATI being interested in SNAP for Linux, and I can remain optimistic about them keeping in the eCS/OS2 code should they buy SNAP though I can't say that my optimism is factually grounded upon past practices, but based, rather, in blind hope.

Another large company would fit the bill... but I have no idea who that would be.


flywheel (17 Nov, 2006 14:47):
The best solution (For the consumers) would be a consortium of a handfull of manufactorers[sic] - bringing support to platforms like eCS, Linux, ZetaOS and so on.

Don't forget, any such setup would require continued investment and development of the product in order to make it a viable effort. Also, the likely reason that SNAP is being dumped is the difficultly in getting access to the necessaries from the video card manufacturers for new cards and the difficulty in keeping up with the new direction of video... that being of 3D particularly so WRT the desktop and still continue to make a profit. SNAP is a 2D driver.

Still, I sure hope someone picks it up and continues to develop it. In the meantime, if SNAP doesn't sell, then I can't see Scitech releasing it as open source, any time in the near future.

Buy it to kill it is a likely scenario.
Buy it to free it is not... though one can hope.


---


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  BigWarpGuy
Date  :  17 Nov, 2006 on 21:33
How much would it sell for? Would it be possible to raise the funds - as a group effort - to pay/buy it?

---
BigWarpGuy
- - - - - - - - - - - - - - - - - -
OS/2-eCS.org
Director of Communications
- - - - - - - - - - - - - - - - -
supporting the past OS/2 user and the future eCS user
http://www.os2ecs.org

Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  kb0uov
Date  :  17 Nov, 2006 on 22:19
>How much would it sell for? Would it be possible to raise the funds - as a group effort - to pay/buy it?

I think we're all curious as to how much it will bring. If there was a way for us
to raise the money, the question would be could we afford the talent needed to add new chipsets, and any fees the manufacturers required for access to the tech specs? I think (if the biddings goes too high) we would be better off trying to buy the os/2 specific part of snap from the high bidder.
I think it would be great if scitech kinda followed jp software, and released a version of their software without any limitations.

kb0uov


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  flywheel flywheel@worldonline.dk
Date  :  17 Nov, 2006 on 22:58

OK. Just a clarification here. Scitech are not selling just the SNAP for eCS/OS2 code... they're selling ALL the SNAP code... as one complete unit. This includes the Windows and the Linux versions. Only the eCS/OS2 version requires GRADD and I don't believe that Scitech has access to the GRADD sources.

I know - but since I am an eCS user I focus on the eCS part. But my guess is that a GRADD license will be involved - I don't know if that is a wishfull thinking.


So, SNAP would end up being just another ATI graphics driver with support for additional non-ATI video cards that presently exist in SNAP ATI no longer makes graphics drivers for eCS/OS2 and they haven't for a long, long time. So why should they buy code to drivers for their own cards to work on another OS that they all ready dropped support for?

ATI got some serious problems with their Linux support - it was in that context SNAP Graphics could be a subject. The Opensource drivers lack stable 3D acceleration. The commercial drivers are far behind the nVidia ones. I have tried installing the ATi drivers, every time something goes wrong. Once I succeded but then it insisted on using the VGA output instead of the DVI. After a week I gave up and invested in SNAP Graphics for Linux. It took about 5 minutes and everything was running smoothly. The installation process and true PnP is killer features - when compaired with the official commercial ATI drivers.


The possibility for supporting more operating systems is due to the nature of the SNAP, and there's already a fully functional OS/2 support (As far as I can see the OS/2 product is the most developed.)



Don't forget, any such setup would require continued investment and development of the product in order to make it a viable effort. Also, the likely reason that SNAP is being dumped is the difficultly in getting access to the necessaries from the video card manufacturers for new cards and the difficulty in keeping up with the new direction of video... that being of 3D particularly so WRT the desktop and still continue to make a profit. SNAP is a 2D driver.

Any form of high technological support requires continued investment and development of the product in order to make it a viable effort.
Many card manufactorers are having problems with delivering easy-to-use drivers for non-Windows systems.
The consortium could be responsable for developing the nukleus drivers. This is the beauty of the SNAP split driver system. The competitors won't have to see as much as one line code. You can release the hardware dependent driver in binary form.


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  Pete losepete@ntlworld.com
Date  :  18 Nov, 2006 on 02:14
Hi All

Just to throw in my 2 pence worth...

There is a post from Roderick, of Mensys fame, in the scitech newsgroup which indicates that Mensys/SSI are trying to work out a solution with Scitech to what will quickly become a problem for their product, eComStation.

If the above does not happen then I think we all have to develop religous tendencies quickly and start praying that whoever gets SNAP decides to donate the non-profitable bit (bits ?) to the open source movement - including the OS/2 stuff.

Regards

Pete


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  miche
Date  :  18 Nov, 2006 on 15:15
IMO the best solution is to have it as open source.
None hypothetical buyer will invest money in the os/2 version...
We need the help and the effort of the linux and windows programmers to update the product, add new hardware support and so on...

We should find some 'sponsor' to buy it and release as open source.


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  Sebadoh sys3175@optonline.net
Date  :  19 Nov, 2006 on 04:55
It won't matter.. without continued development of drivers it will soon be useless.. and most likely anyone who purchases it will develop it in a way in which new drivers cannot work with the older os/2 nucleus.


kb0uov (17 Nov, 2006 22:19):
>How much would it sell for? Would it be possible to raise the funds - as a group effort - to pay/buy it?

I think we're all curious as to how much it will bring. If there was a way for us
to raise the money, the question would be could we afford the talent needed to add new chipsets, and any fees the manufacturers required for access to the tech specs? I think (if the biddings goes too high) we would be better off trying to buy the os/2 specific part of snap from the high bidder.
I think it would be great if scitech kinda followed jp software, and released a version of their software without any limitations.

kb0uov



Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  Kim
Date  :  19 Nov, 2006 on 09:59
I might be a slow reader, but has there been mentioned any kind of amounts what SciTech expects to get paid for the code? Also, are bundling all code together regardless OS?

Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  flywheel flywheel@worldonline.dk
Date  :  19 Nov, 2006 on 14:19

Kim (19 Nov, 2006 09:59):
I might be a slow reader, but has there been mentioned any kind of amounts what SciTech expects to get paid for the code? Also, are bundling all code together regardless OS?

No, no minimum limit has been mentioned.
But naturally they'll try selling everything in one package - if that fails piece by piece - if that fails - I guess it ends up in a filecabinet in the cellar.

IMO they can't release it for opensource without making it totally useless - most of the drivers are made on the background of NDAs (Except the latest few generations of nVidia drivers that has been reverse engineered - and might not be that much worth as opensource) and there's the GRADD business to think of.


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  warprulz daniel.lee.kruse@gmail.com
Date  :  21 Nov, 2006 on 04:38

flywheel (19 Nov, 2006 14:19):
<snip>
But naturally they'll try selling everything in one package - if that fails piece by piece - if that fails - I guess it ends up in a filecabinet in the cellar.

Hope not


IMO they can't release it for opensource without making it totally useless - most of the drivers are made on the background of NDAs (Except the latest few generations of nVidia drivers that has been reverse engineered - and might not be that much worth as opensource) and there's the GRADD business to think of.

I have to agree. I suspect there is too much proprietary, non-SciTech code in there. Too bad SNAP couldn't be opensourced. Is there a univid like there is for sound (uniaud)? I think that would be our best bet for future video drivers.

If GRADD was completely IBM developed couldn't they opensource that like they have done with some of their other code (JFS, module loading, etc.)? There probably is too much non-IBM, proprietary code in there, too, relating to graphic chip implementations. Like I wrote above - univid.


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  Sebadoh sys3175@optonline.net
Date  :  21 Nov, 2006 on 06:45
There isn't.. because honestly no operating system without official support for video chipsets is viable as a platform. Honestly thinking that OS/2 isn't really viable as a graphics platform, but it is better then most non-supported os'es simply because it does have snap. most have to make due with simply vesa fallback mode of most video cards which is getting increasingly poor as time progresses. This doesn't fair well for us. I think they could release snap as opensource (I also believe they will not) they may have problems releasing the actual chipset drivers but the core system could defienlty be released. And gradd support wouldn't even matter.. since if they released the support driver compiled it wouldn't even need gradd components. Gradd is also a publicly available source (however I am not sure about the latest public release level. Also gradd isn't even that much of a concern if scitech decided on the opensource route since OS/2 again would not be a primary concern.

Oh well.. been fun while it lasted, and who knows what I will end up doing if and when my ati 9550 dies.


warprulz (21 Nov, 2006 04:3:

flywheel (19 Nov, 2006 14:19):
<snip>
But naturally they'll try selling everything in one package - if that fails piece by piece - if that fails - I guess it ends up in a filecabinet in the cellar.

Hope not


IMO they can't release it for opensource without making it totally useless - most of the drivers are made on the background of NDAs (Except the latest few generations of nVidia drivers that has been reverse engineered - and might not be that much worth as opensource) and there's the GRADD business to think of.

I have to agree. I suspect there is too much proprietary, non-SciTech code in there. Too bad SNAP couldn't be opensourced. Is there a univid like there is for sound (uniaud)? I think that would be our best bet for future video drivers.

If GRADD was completely IBM developed couldn't they opensource that like they have done with some of their other code (JFS, module loading, etc.)? There probably is too much non-IBM, proprietary code in there, too, relating to graphic chip implementations. Like I wrote above - univid.



Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  flywheel flywheel@worldonline.dk
Date  :  21 Nov, 2006 on 14:18

No, no minimum limit has been mentioned.
But naturally they'll try selling everything in one package - if that fails piece by piece - if that fails - I guess it ends up in a filecabinet in the cellar.

This wasn't meant as fatalistic as it might seem.
IMO no one is interested in buying the whole package (Except if someone like AMD wants to pull something), and I guess that no one is interested in buying the OS/2 subpackage.
IMO that is a bit of an advantage. That will hopefully force ScitechSoft to try to sell the packages seperately, and hopefully we'll end up with a small discount on the OS/2 package. (Then again I guess the licence fees could trip everything)

What I would really like is that all of it will get sold - even the BeOS alpha code (Perhaps Magnussoft and their partners would be interested) - a partnership across platforms would make SNAP much stronger, than one single OS/2 instance.

Most of this might seem as blabber - which it is, I'm of to bed - been a long day


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  cytan
Date  :  21 Nov, 2006 on 15:31
Hi all,
After installing Scitech SNAP 3.1.8 in 02 Oct, my OS/2 box which I use at work everyday, it just had an uptime of about 49 days (wrap around is seen!):
********************************
E:/@rachel%uptime
8:28am up 0 days, 3:01:54, load: 70 processes, 276 threads.
E:/@rachel%xuptime
rachel: uptime is 49 days, 19:00 hours and 34 seconds
E:/@rachel%
***********************************

I've never been able to get an uptime of more than 2.5 weeks without the video or WPS hanging. Finally after years of waiting, Scitech finally got something for me that works and now it may be going away ;-(. This is sad!

cytan


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  obiwan
Date  :  21 Nov, 2006 on 19:13
Don't panic guys. Nothing is decided yet, and we've been through countless episodes like this, and it has worked out anyway. There are plenty of reasons to be optimistic about a change in SNAP. We do not yet know the outcome of this. The very worst imaginable case is we have to use something else for video support, which, while painful, is not bad in the end, because we need more features than what SNAP currently gives us, and it gives us another reason to work on it.

Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  JLKT
Date  :  23 Nov, 2006 on 01:12
This news is indeed depressing. Now that we don't have SNAP (I know nothing is concrete at the moment), what other video drivers can we use and rely on?

I don't mean to drag on and on but without good video driver support, os/2 is indeed dead.


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  flywheel flywheel@worldonline.dk
Date  :  24 Nov, 2006 on 12:35
Is MGL and the SDK's included ?

Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  cyberspittle cyberspittle@yahoo.com
Date  :  24 Nov, 2006 on 20:41
I'm really not sure how bad the news really is. Iv'e used SciTech for quite some time on unsupported chips using their VBE mode. In the future, I just won't have the option of an update with my chipset being supported. I guess I better get used to VBE mode. Hopefully, the speed of GPUs will continue to compensate for non-native drivers. I'm sure eBay will be albe to provide OS/2-eCS users for some time with video cards ...

Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  obiwan
Date  :  28 Nov, 2006 on 02:43

JLKT (23 Nov, 2006 01:12):
what other video drivers can we use and rely on?

Some folks have been working independently on their own OS/2 video drivers. SciTech has worked so well that these efforts have been largely isolated and unnoticed. If SNAP does decline, the good work of independent driver developers will attract recognition, support, and hopefully improved collaboration. This is good because SciTech was never going to expand the feature set of SNAP on OS/2.

Most Xorg drivers are open source, so much of what we need to produce good drivers is available to us.


without ... os/2 is indeed dead.

Fill in the blank with any of the numerous features we have been momentarily without in the past decade. If any one thing could "kill" OS/2 it would have already. We should be past using the "D" word by now.


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  Cris
Date  :  28 Nov, 2006 on 10:05

obiwan (28 Nov, 2006 02:43):
Some folks have been working independently on their own OS/2 video drivers.

Obiwan... could you be more specific on that please?
I'd like to know which card I'm going top buy next time.

Thankyou!
Bye
Cris


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  obiwan
Date  :  28 Nov, 2006 on 18:05

Cris (28 Nov, 2006 10:05):
Obiwan... could you be more specific on that please?
I'd like to know which card I'm going top buy next time.

What I saw was activity on the os2ddprog list. Some people were working on an accelerated driver based on GENGRADD. Looked like the main target was a Matrox card. Still, I would stick to cards that are known to work well with SNAP until these other projects become more mature.

My point was simply that we are not completely in the dark if SNAP fades (and that is the worst it can do - it won't suddenly disappear). Obviously it would be very nice to see SNAP for OS/2 continue development. I just think we need to be optimistic, whatever the outcome here.


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  Fahrvenugen
Date  :  28 Nov, 2006 on 18:29
I have to agree, while I'd love to see SNAP continued, if the product is sold, I have my doubts that the OS/2 code base will be continued.

The one thing which has come to mind though, and I don't know if SSI has thought about this possibility. But, speaking hypothetically, if SNAP is sold to a company which is not interested in the OS/2 code base, I wonder if it would be possible for SSI to work out an agreement with whomever ends up owning SNAP to allow for the continued support (and development) of the OS/2 product through SSI hired developers.

I'm sure there would be legal issues and non-disclosure (NDA) agreements to be worked out, but essentially I could see the possibility of something like this:

-Company X buys and owns the SNAP code, and provides support for any operating systems they want, with the exception of eCS (and/or OS/2)

-Company X allows SSI / SSI hired developers (through a NDA agreement) to have access to the SNAP code (particularly the OS/2 code) for the purpose of providing SNAP on eCS

-SSI developers are allowed to use the current OS/2 code base, along with future enhancements made to the main SNAP source code tree (support for newer video cards, etc), within SNAP for eCS

-Any code developed by SSI developers that isn't eCS specific and can enhance the overall product is also available to Company X developers for inclusion in the main SNAP product.

I have no clue if anything like this would be possible, but it might be worth SSI considering.


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  obiwan
Date  :  28 Nov, 2006 on 19:05
That is probably the optimal solution, and I would guess that is what Serenity hopes to negotiate.

Of course, if OS/2 SNAP does end up in the hands of Serenity, that is very good because Serenity has the most interest in its maintenance and development. A possible downside is it could be expensive, but that can be managed.

Still, another reason for optimism!


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  abwillis abwillis1@gmail.com
Date  :  28 Nov, 2006 on 20:00
I doubt that SNAP can be pieced out... SNAP if it is of value derives it value from the technology in such a way that I don't think that someone interested in buying it would want someone else to have. The only thought I had around this, if someone (maybe SSI) were to _license_ the technology for the OS/2 code before it was sold with an agreement that would go with the sale. In order to keep the license fee low the agreement would be such that SSI can only develop OS/2 drivers so long as the company that is purchasing is not creating OS/2 drivers (non-compete clause). SSI would be able to able to develop drivers for a period of X (maybe 5?) years so long as the buyer is not creating OS/2 drivers. If they create drivers for say 3 years SSI could not compete during that time but could resume for the remainng time of the license. I say this would lower the cost as SSI goes in with the position that they may never have a chance to use the license (the company could potentially build OS/2 drivers for the whole time). This protects the buyer in that they have a non-compete clause so that SSI will not compete with sales of OS/2 drivers and are not licensed for linux or Windows drivers. The time frame is important so that it goes with the sale so that another VPC doesn't occur. Whether new Snap technology the new company develops gets to be used by SSI would be negotiable but for costs I would assume no.

Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  obiwan
Date  :  28 Nov, 2006 on 21:14
Correct me if I'm wrong. Isn't the point of SNAP that there are two components, a platform side and a hardware side, so that once a driver is written for a card it works on any platform that has SNAP? So SSI would own/control only the OS/2 implementation of SNAP, and have the ability to use (with proper licensing) any drivers written by the new owners of SNAP. Presumably the reason this is an issue is because the buyer will not be interested in maintaining the OS/2 component, so the arrangment would not be competitive, but mutually beneficial.

Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  Sebadoh sys3175@optonline.net
Date  :  29 Nov, 2006 on 07:19
That would be assuming that anyone who purchases it a) maintains compatability with the binary format they are currently using b) doesn't buy it for propietary purposes, and even gives a damn enough to allow continued licensing of the OS/2 version.


obiwan (28 Nov, 2006 21:34):
Correct me if I'm wrong. Isn't the point of SNAP that there are two components, a platform side and a hardware side, so that once a driver is written for a card it works on any platform that has SNAP? So SSI would own/control only the OS/2 implementation of SNAP, and have the ability to use (with proper licensing) any drivers written by the new owners of SNAP. Presumably the reason this is an issue is because the buyer will not be interested in maintaining the OS/2 component, so the arrangment would not be competitive, but mutually beneficial.

Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  Sebadoh sys3175@optonline.net
Date  :  29 Nov, 2006 on 07:26
True.. declaring it dead is a little over the top. On the other hand lets look at it this way. The OS/2 core components (or atleast what is needed to make a modern os useful)

Printer Drivers: Development ceased dec 2006 (well actually before that but lets use the "official" date.

Video Drivers: Snap is frozen, most 3rd party and oem drivers are at various degrees of non-functionality.

Sound Drivers: UniAud! Yeah! I get to listen to the same 15 second sound sample everytime I turn my computer on randomly! yeah uniaud (and alsa too for to that end)

Ethernet: Yeah genmac2 now we can run some but not all windows drivers, atleast we have SOME wifi drivers now.

Tcpip: Frozen December 2006 (actually quite some time before that).

Kernel: Frozen December 2006 (again, actually might as well go back to 2004 on this one)

Wps: Now this is a fun one.. yeah guys.. we're gonna go back to 1996 here =)


Hmmmm.. there is a necromancer somewhere which has a corpse of OS/2 walking around for quite some time =) lol..

hey.. have a sense of humor.. sing the doom song with me.. "doom doom dooom.. doom doom doom..."


obiwan (28 Nov, 2006 02:43):

JLKT (23 Nov, 2006 01:12):
what other video drivers can we use and rely on?

Some folks have been working independently on their own OS/2 video drivers. SciTech has worked so well that these efforts have been largely isolated and unnoticed. If SNAP does decline, the good work of independent driver developers will attract recognition, support, and hopefully improved collaboration. This is good because SciTech was never going to expand the feature set of SNAP on OS/2.

Most Xorg drivers are open source, so much of what we need to produce good drivers is available to us.


without ... os/2 is indeed dead.

Fill in the blank with any of the numerous features we have been momentarily without in the past decade. If any one thing could "kill" OS/2 it would have already. We should be past using the "D" word by now.



Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  obiwan
Date  :  29 Nov, 2006 on 19:05

Sebadoh (29 Nov, 2006 07:19):
That would be assuming that anyone who purchases it a) maintains compatability with the binary format they are currently using b) doesn't buy it for propietary purposes, and even gives a damn enough to allow continued licensing of the OS/2 version.

Generally, investors who buy the rights to software are looking to (1) cash in on licensing sales wherever possible, and (2) spend as little as possible on development. So these two assumptions are more reasonable than we might anticipate (accustomed as we are to being dismissed by software licensors).

Retension of binary compatibility would be easiest on us, reducing the need for upgrading, but actually what I had in mind was that if SSI ends up controlling the OS/2 component it would be for the purpose of maintaining compatibility with future changes.


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  obiwan
Date  :  29 Nov, 2006 on 19:28

Sebadoh (29 Nov, 2006 07:27):
... atleast what is needed to make a modern os useful

More precisely, what makes it hardware-compatible. That's actually quite distinct from useful. I'm quite the fan of using emulators to extend the usefulness of software beyond the realm of hardware-compatibility, although I admit I'm not too excited about having to do that with OS/2 in particular.

The other point touched on is development activity on components such as the WPS. Of course, OS/2 is and always has been a different way of computing from the "carrot-on-a-stick" chase of Windows and the "release-early, fix-often" [sic] model of Linux. With the general quality and stability of IBM's original OS/2 base, and most software developed for it early on, the date of the latest update is not as relevant. However, note that the WPS is peculiarly extensible, and many independent new enhancements have been and are being made to it, some incorporated into eComStation, so the date of the last IBM-branded change is not really fair.

Still, though, you point out well that this move by SciTech is a reminder of the challenges of continued maintenance of this OS, and the requirement of a long-term vision in the OS/2 community, along the lines of Voyager and/or OSFree, discussed in another thread.

BTW: Sound looping is sometimes an indication of an IRQ conflict or failing hardware.


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  Ben Dragon
Date  :  29 Nov, 2006 on 20:32
The main problem that I see with Scitech selling SNAP is that it's highly unlikely that anyone (or company), would pick it up... I mean with the purpose of maintaining and selling it as Scitech did.

I mean, Scitech isn't selling SNAP because it's making them too much money. They're selling it because the product can no longer make them enough money to keep the company going and keep employees paid.

Therefore, unless someone, (or some company), is a hell of a lot better at finding/creating a market for the product than Scitech was, then SNAP is dead in the water and we are SOL... that is to say without a universal video driver in the manner that SNAP has been.

That's the bad news.
The good news comes from that as well.

That is to say...
The less viable the product is, the cheaper it will sell for on the open market. If it turns out to be not very viable as a product,
(as opposed to as a driver), then there is an increased likelihood that Mensys/SSI will be able to afford to, and will choose to buy it.

Let us assume for a moment, that they can afford it purchase it and that the price is no longer a major factor, then Mensys/SSI must determine whether or not the community has the expertise to develop the driver. And, if it does, then Mensys/SSI must determine if those with the expertise will do it for free, or if they'll want to be paid.

I, personally, don't know if the community has the expertise, but I do know, if I were doing it, (I lack the skills), I would want to get paid for it.

So that begs yet another question; Assuming Mensys/SSI do buy the product, will it make them enough money, (in additional eCS/OS2 licenses sold), to warrant employing someone to maintain it?

If the answer to this question is "yes", then we might get our favourite universal video driver back after all.

Let us all cross our fingers and hope that that ends up being the case, for then the drivers are ours and we won't find ourselves in this situation again at some point down the road.

---


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  ecsguy
Date  :  30 Nov, 2006 on 05:25
Best solution is to get OS/2 and eComStation abstracted from the hardware completely.

If we had an open source virtual machine like QEMU, you could be taking advantage of dual core CPUs right now. You could have RAID disk storage.

Right now eComStation is always trailing. With a Linux host layer provided by QEMU we could stay even with Linux.
http://fabrice.bellard.free.fr/qemu/


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  obiwan
Date  :  30 Nov, 2006 on 07:47
Huh? Dual core CPU's are utilized by the SMP kernel. RAID is possible on OS/2.

OS/2 can run as a guest on QEMU. I do it. It's a bit slow but workable.


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  zman zirkle@wizard.net
Date  :  30 Nov, 2006 on 09:20
i suspect he means you can't do it cheaply with what most os/2 users are running.

my server has an old; but still awesome, mylex dac960 with 32mb cache.

my current desktop is now a real success. the ecs2.0 beta team has really gotten things to fall nicely into place for me. the latest acpi allows all my hardware to work. eveything is running on high irq! six usb ports (2 usb2.0) work. lucide and open office beta2 is good here to! technically if you buy a pentium d you'll have to -- oh i know this is so hard for most users -- PAY and get rid of that ancient patched version 4 or even the worse older stuff.

everybody can take this as an ecs2 beta report as well.


system:

pentiumd 3ghz
1gb ram
ati x850 video (screen is 1920x1080p)
32" philips lcd hdtv monitor
intel d955xbk mobo
intel 100mb lan (no driver for on board 1gb)
pci soundblaster
hauppauge win tv pvr350
2 sony dvd-rw drives
1 sata 15k rpm harddrive
virtual pc run windows xp

this is my primary tool, not a mule. os/2 beta beats ms ga most any day!


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  Cris
Date  :  30 Nov, 2006 on 11:23

ecsguy (30 Nov, 2006 05:27):
Best solution is to get OS/2 and eComStation abstracted from the hardware completely.

If we had an open source virtual machine like QEMU, you could be taking advantage of dual core CPUs right now. You could have RAID disk storage.

Right now eComStation is always trailing. With a Linux host layer provided by QEMU we could stay even with Linux.
http://fabrice.bellard.free.fr/qemu/


I don't like this approach at all, and I suspect a lot of eCS users would jump boat if this was the direction chosen for the future of OS/2. I would.

Bye
Cris


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  BigWarpGuy
Date  :  30 Nov, 2006 on 18:55

Cris (30 Nov, 2006 11:23):

ecsguy (30 Nov, 2006 05:27):
Best solution is to get OS/2 and eComStation abstracted from the hardware completely.

If we had an open source virtual machine like QEMU, you could be taking advantage of dual core CPUs right now. You could have RAID disk storage.

Right now eComStation is always trailing. With a Linux host layer provided by QEMU we could stay even with Linux.
http://fabrice.bellard.free.fr/qemu/


I don't like this approach at all, and I suspect a lot of eCS users would jump boat if this was the direction chosen for the future of OS/2. I would.

Bye
Cris



I totally agree with you. I would rather switch to ROS than do something like that.

---
BigWarpGuy
- - - - - - - - - - - - - - - - - -
OS/2-eCS.org
Director of Communications
- - - - - - - - - - - - - - - - -
supporting the past OS/2 user and the future eCS user
http://www.os2ecs.org


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  Sebadoh sys3175@optonline.net
Date  :  30 Nov, 2006 on 21:11

obiwan (29 Nov, 2006 19:33):

Sebadoh (29 Nov, 2006 07:27):
... atleast what is needed to make a modern os useful

More precisely, what makes it hardware-compatible. That's actually quite distinct from useful. I'm quite the fan of using emulators to extend the usefulness of software beyond the realm of hardware-compatibility, although I admit I'm not too excited about having to do that with OS/2 in particular.

The other point touched on is development activity on components such as the WPS. Of course, OS/2 is and always has been a different way of computing from the "carrot-on-a-stick" chase of Windows and the "release-early, fix-often" [sic] model of Linux. With the general quality and stability of IBM's original OS/2 base, and most software developed for it early on, the date of the latest update is not as relevant. However, note that the WPS is peculiarly extensible, and many independent new enhancements have been and are being made to it, some incorporated into eComStation, so the date of the last IBM-branded change is not really fair.

Still, though, you point out well that this move by SciTech is a reminder of the challenges of continued maintenance of this OS, and the requirement of a long-term vision in the OS/2 community, along the lines of Voyager and/or OSFree, discussed in another thread.

BTW: Sound looping is sometimes an indication of an IRQ conflict or failing hardware.


Don't think that is the case since even when the sound is looping the computer runs perfectly stable, in fact since I rarely turn my speakers on anymore I didn't even notice it was happening =). The newest RC seems to have fixed this, atleast so far ::fingers crossed::

And yes, WPS/Som is a great thing and it is highly customizable, it is for this reason I have only OS/2 installed on my computer. On the other hand there are annoying and nagging bugs which have never been fixed and will never be fixed. Put that on top of the fact that the more you extend it the more you run into shared memory shortage and overloading the WPS process. It is amusing in a computer with 2GB of ram, I still get out of memory errors, actually alot more often then I used to on my old machine with 256megs and I have 1/4 of the apps running which I used to have. (I ran a 16 node BBS on a 128mb machine including custom made mail daemons, (i wrote the bbs software myself), a web gateway, 12 telnet nodes and quite often I when online with it because it had the best video card of any of my pcs).. today, PSI and Firefox are enough to run me outta ram =)

Anyway.. I was just in a sarcastic mood and was having a bit of fun.. I mean.. nothing is perfect, I love OS/2 and it's faults.. because atleast I know where the faults lie.


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  Sebadoh sys3175@optonline.net
Date  :  30 Nov, 2006 on 21:18
However, it is quite likely that a linux or embedded system developer will buy it at a pretty steep price (alot more then SSI or mensys could afford) and either kill it, or just develop it for their own purposes. I am sure it would be alot easier then maintaining seperate xserver for each video chipset. And they would undoubtably add accelerated 3d into the mix.. This all would not find it's way into the OS/2 version.. because like I said if this scenario plays out, the OS/2 version will be cut out of the mix due to the fact that the developer would not be at all interested in cross platform capabilities or spending money on continued OS/2 development with no return in investment. The best case scenario they let SSI work on the OS/2 versions.. however without the help of the original coders this would be a very slow and tedious process and much would get broken along the way undoubtably, more likely they would only be integrating a rather huge code base of working device drivers into their own system.



Ben Dragon (29 Nov, 2006 20:32):
The main problem that I see with Scitech selling SNAP is that it's highly unlikely that anyone (or company), would pick it up... I mean with the purpose of maintaining and selling it as Scitech did.

I mean, Scitech isn't selling SNAP because it's making them too much money. They're selling it because the product can no longer make them enough money to keep the company going and keep employees paid.

Therefore, unless someone, (or some company), is a hell of a lot better at finding/creating a market for the product than Scitech was, then SNAP is dead in the water and we are SOL... that is to say without a universal video driver in the manner that SNAP has been.

That's the bad news.
The good news comes from that as well.

That is to say...
The less viable the product is, the cheaper it will sell for on the open market. If it turns out to be not very viable as a product,
(as opposed to as a driver), then there is an increased likelihood that Mensys/SSI will be able to afford to, and will choose to buy it.

Let us assume for a moment, that they can afford it purchase it and that the price is no longer a major factor, then Mensys/SSI must determine whether or not the community has the expertise to develop the driver. And, if it does, then Mensys/SSI must determine if those with the expertise will do it for free, or if they'll want to be paid.

I, personally, don't know if the community has the expertise, but I do know, if I were doing it, (I lack the skills), I would want to get paid for it.

So that begs yet another question; Assuming Mensys/SSI do buy the product, will it make them enough money, (in additional eCS/OS2 licenses sold), to warrant employing someone to maintain it?

If the answer to this question is "yes", then we might get our favourite universal video driver back after all.

Let us all cross our fingers and hope that that ends up being the case, for then the drivers are ours and we won't find ourselves in this situation again at some point down the road.



Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  obiwan
Date  :  30 Nov, 2006 on 21:47

zman (30 Nov, 2006 09:20):
if you buy a pentium d you'll have to -- oh i know this is so hard for most users -- PAY and get rid of that ancient patched version 4 or even the worse older stuff.

I know this is getting off-topic, but why will we have to?


virtual pc run windows xp

There is a VPC for OS/2?


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  rudi
Date  :  01 Dec, 2006 on 13:07

obiwan (28 Nov, 2006 18:05):

What I saw was activity on the os2ddprog list. Some people were working on an accelerated driver based on GENGRADD. Looked like the main target was a Matrox card.


The only activity in the field of graphics drivers on the PDD list was a long time ago. It was not targeted to create an accelerated video driver, but to add motion video acceleration. In the end these efforts lead to the creation of WarpOverlay!, which is abandoned now. So I don't see any reason to be optimistic.


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  ecsguy
Date  :  01 Dec, 2006 on 18:13

zman (30 Nov, 2006 09:20):
i suspect he means you can't do it cheaply with what most os/2 users are running.

Yes, exactly.

I have run RAID on OS/2 in the past. I could run it today by buying surplus hardware.

BUT I can't tell customers to buy surplus hardware.

What about even new home users!.

I talked to a guy yesterday that has a new Gateway brand system with Intel dual core (two 3 GHz cores), 1GB memory and two 230GB SATA drive. He won't consider eComStation until it can take advantage of most of his hardware.

You can't keep scrounging parts forever.
Dual core and RAID are very inexpensive today compared to the past.

There are newer SATA disk controllers chipsets coming. eComStation and OS/2 might lack a driver. Again another wait.

Always waiting.

With QEMU, eComStation and OS/2 users can take advantages of new hardware advances much sooner.

Running the OS in a virtual is going to be the standard setup for all OSes anyway. I run eComStation for the OS and apps not the hardware. The hardware is important just like the foundation of a house. BUT it is the OS and Apps that I interface with. Most eComStation and OS/2 users that need to buy a NEW machine today would have better choices if eComStation and OS/2 MCP2 ran in QEMU. Running in them today in QEMU * IS * faster than waiting 6 months or 12 months or forever for another driver or kernel change.


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  obiwan
Date  :  01 Dec, 2006 on 18:20
Independent efforts are isolated and uninteresting when a good system like SNAP is easily available. They become more interesting as it becomes less supported. The efforts have existed, there are people who can do it and would again, whether they are now or not.

I don't advertise all my projects, so I don't assume that silence means there is zero work or research going on.

As Ben Dragon pointed out very well, the economics of it dictate how this falls into place. But in the end it takes work to maintain drivers, and that will equal the time and money we as users put into it. That is a given, whether SciTech maintains SNAP or SSI or someone else or no-one. A change like this is merely a shift, and that can be good because it can mean new things. The only reason it can be bad is if we all drop OS/2 because of it, so there is no more money or time invested.

That's why I think pessimism about this unfinished story is worse, and why I respond. I can't force you to be optimistic, though, I can only encourage you. You will see in it what you want to.


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  obiwan
Date  :  01 Dec, 2006 on 18:38

ecsguy (01 Dec, 2006 18:13):
Most eComStation and OS/2 users that need to buy a NEW machine today would have better choices if eComStation and OS/2 MCP2 ran in QEMU. Running in them today in QEMU * IS * faster than waiting 6 months or 12 months or forever for another driver or kernel change.

As I said, eComStation and OS/2 do run in QEMU. You are free to do that if you want to. I do it for some purposes but couldn't tolerate it as a primary desktop. As others have pointed out, it is not at all desireable to most of us. You keep mentioning dual core processors, but that gains you nothing if it is slowed down in an emulator. Besides, QEMU itself does not take advantage of multiple cores. It would provide some additional RAID capability in a roundabout way, but at the cost of maintaining a host OS that most of us would not want to. For less work you could create backups that would restore a destroyed OS/2 system nearly as fast as software-based RAID recovery, which really should be done anyway even if you use RAID.


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  zman zirkle@wizard.net
Date  :  01 Dec, 2006 on 19:47

obiwan (30 Nov, 2006 21:47):

zman (30 Nov, 2006 09:20):
if you buy a pentium d you'll have to -- oh i know this is so hard for most users -- PAY and get rid of that ancient patched version 4 or even the worse older stuff.

I know this is getting off-topic, but why will we have to?


virtual pc run windows xp

There is a VPC for OS/2?


the acpi in ecs 2.0 is licensed only for ecs 2. it is required to get all the benefits of the system. up until october, i had to run this system in pic mode. all the drivers loaded with low irqs, stomped on the usb controllers, and the video adapter. now they load high (irq 16-23) and are shareable thanks to highirq.sys.

virtual pc was available for os/2 til ms bought out connectrix. it was a product of innotek. of course we got the usual howls -- its too expensive, os/2 is perfect on my pentium 1 nobody should want more, it is so slow, and .... i use it on select pc at corporate and at home. the ms version does not run, yet, on an intel mac; but, parallels does. yea -- no qemu or what ever for me. i want twenty first century multimedia capacity NOW! mac osx is my migration path of choice starting with replacing my tp770x next year!!!


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  ecsguy
Date  :  01 Dec, 2006 on 20:06

obiwan (01 Dec, 2006 18:3:
As I said, eComStation and OS/2 do run in QEMU. You are free to do that if you want to. I do it for some purposes but couldn't tolerate it as a primary desktop.

Several other people have tried without luck to get eComStation and OS/2 MCP2 working in QEMU.
http://qemu-forum.ipi.fi/viewforum.php?f=17&sid=cadb45771a7a9f2f2bd618be8a7c71f4

We could really use help.

Can you post version info and steps in install procedure?

Did you have networking (peer and tcpip) working?


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  obiwan
Date  :  01 Dec, 2006 on 20:50

ecsguy (01 Dec, 2006 20:06):
Several other people have tried without luck to get eComStation and OS/2 MCP2 working in QEMU.
http://qemu-forum.ipi.fi/viewforum.php?f=17&sid=cadb45771a7a9f2f2bd618be8a7c71f4

We could really use help.


Hrm, I thought all the problems had been covered there. I'll read more carefully, and we can take the QEMU discussion to that forum. I'll try to set aside some time this weekend and see if I'm able to help figure out the trouble people are still having.


Can you post version info and steps in install procedure?

Warp 4. I just did the same as was already described on that forum.


Did you have networking (peer and tcpip) working?

No, haven't bothered with it because I haven't needed it yet. QEMU networking is rather a pain with any guest system. Again I'll try this weekend to get some time to attempt to reproduce the problems (and successes) described in the forum. I need to get my FreeBSD image connected anyway. Peer might be a long-shot, but I'll try it.


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  OriA; alanh77@comcast.net
Date  :  02 Dec, 2006 on 03:55
Will eCS run on an Intel Mac?

---
Alan

Nerve Center BBS tncbbs.no-ip.com


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  ecsguy
Date  :  03 Dec, 2006 on 02:17

OriA; (02 Dec, 2006 03:55):
Will eCS run on an Intel Mac?

The processor is not a problem but are there OS/2 drivers for video, NIC, sound, and even disk controller?
This web site has links to the drivers windows users are using for Intel Mac
http://wiki.onmac.net/index.php/Users/XOM/Drivers

QEMU does run on Intel Mac. It is also much easier getting windows to run as a QEMU guest on Intel Mac than to run it directly on the hardware.
http://qemu-forum.ipi.fi/viewtopic.php?t=2002&highlight=intel&sid=008983619606b6564a36d8f822db851d

I tried running eComStation as a guest on QEMU on Intel Mac but eComStation locked up during boot just like on a regular Intel. I was able to run Warp4 (pre fixpak) and no networking in QEMU on Intel Mac. So the basis for running eComStation must be there. Just need some developers to look at the few bumps that are keeping eComStation from running. Also networking is a must since there are no VPC type shared folders

QEMU
http://fabrice.bellard.free.fr/qemu/


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  obiwan
Date  :  03 Dec, 2006 on 20:18
Sorry ecsguy, you are right. QEMU isn't as slow as I remembered it, so it could be a solution for some situations. Also there is the locking on boot, even with Warp 4. I'm thinking it could have to do with the way QEMU emulates a hard drive. It can probably be resolved, but I've spent about as much time on it as I can afford. I'm leaving my notes about it on the QEMU forum. Hopefully I'll be able to get back to it later, or someone else can try my suggestions.

Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  Sebadoh sys3175@optonline.net
Date  :  04 Dec, 2006 on 00:55

obiwan (03 Dec, 2006 20:1:
Sorry ecsguy, you are right. QEMU isn't as slow as I remembered it, so it could be a solution for some situations. Also there is the locking on boot, even with Warp 4. I'm thinking it could have to do with the way QEMU emulates a hard drive. It can probably be resolved, but I've spent about as much time on it as I can afford. I'm leaving my notes about it on the QEMU forum. Hopefully I'll be able to get back to it later, or someone else can try my suggestions.

I for one absolutely hate emulation.. and it actually doesn't do a whole lot for us, even though it is running on newer hardware it doesn't add any capabilities I didn't have otherwise. I still can't use an usb device which doesn't have OS/2 support unless there is some network mapable way to make it work which is a huge kludge. No matter how good the emulator is it is always going to be slower then running it native (i am not talking about a hardware level emulator like VM, just any software emulator) and guys. saying that you need ecomstation 2 in order to use acpi is kinda silly... that is like complaining because you need to use Warp since an app doesn't run in 2.11! =) times change. At some point you need to upgrade and atleast ecomstation 2.0 is in the works.. (even if it isn't really a new OS as 2.x would imply.. it is simply... 1.4 w/ 3rd party extensions =).. warp 4.56 =)


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  Junior jj_vandelft@ananzi.co.za
Date  :  04 Dec, 2006 on 23:01
Hi all,

Been following this forum a lil n found this at GoldenCode web site :

" Device Driver Development. Golden Code can provide full design and development services for OS/2 device drivers. Once armed with the functional requirements of the driver and any associated hardware/hardware interface documentation, this team can create, modify or maintain/support virtually any device driver project. The driver development team is experienced in a wide range of OS/2 driver types including:

* both 16-bit and 32-bit physical device drivers (PDD)
* implementation of network drivers using the Network Device Interface Specification (NDIS) at both the lower (MAC) and upper (protocol) interfaces
* security enabling services installable security subsystem (SES ISS)
* installable file systems (IFS) "

Perhaps they would be prepared to take over or buy snap. Noticed how well people supported netlabs, perhaps the OS/2 community would help in supporting them as well for a good cause. I'm in as much i can possibly give within my means.


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  ecsguy
Date  :  05 Dec, 2006 on 03:23

obiwan (03 Dec, 2006 20:1:
Sorry ecsguy, you are right. QEMU isn't as slow as I remembered it...

The speed of QEMU was the thing that first got my attention. And that is without the QEMU acceleration feature!

eComStation guest on QEMU would be very useable on new machines.

We need a virtual machine environment to complement the direct hardware support we can get. But the VM environment needs to be open source so it does not get bought (killed) again. Since VMWare OS/2 support was killed, we have invested too much (time (5 years!!!), and $$$) to be empty handed. When Microsoft goes to Parallel's and says let's make a deal, what do you think will happen to OS/2 guest support.


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  zman zirkle@wizard.net
Date  :  05 Dec, 2006 on 22:33

ecsguy (05 Dec, 2006 03:23):


We need a virtual machine environment to complement the direct hardware support we can get. But the VM environment needs to be open source so it does not get bought (killed) again. Since VMWare OS/2 support was killed, we have invested too much (time (5 years!!!), and $$$) to be empty handed. When Microsoft goes to Parallel's and says let's make a deal, what do you think will happen to OS/2 guest support.


good points. choice is good and of course i'm for the progress of qemu; but, parallels with its working full usb support does for me what vpc for os/2 came just short of right now. if vpc had made it i'd have been liberated. microsoft offered me a free upgrade to vpc for windows which supports os/2 as a guest. in fact vpc 2004sp1 still supports os/2 as a guest and i downloaded the "server version". i could have solved my os/2 problems by reversing the roles windows and os/2 play in my life and it would not have cost me a penny; but, i'm not quite readty to roll over in spite of the money!

i find linux on the software front to be no better than os/2 for my future computing applications. more and more i am pushed to move to something more mainstream. so if ms makes a deal with parallels it looks like the os/2 guest support would be there as with vpc. at this point beating up on poor old os/2 for microsoft is so over and who cares anymore. in the end, ibm's treatment really was worse and in the long run ibm is not a friend of linux or open source.

i think golden code could keep snap up; but, the issue is the money. ibm paid scitech for snap. now that ibm has cut off funding, scitech lost its number one source of funds for snap. since os/2 users haven't paid for it in large enough numbers .... well you get the rest! looking at what happened with the java mess, i wonder if golden code really cares!


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  ecsguy
Date  :  06 Dec, 2006 on 05:51

zman (05 Dec, 2006 22:33):
good points. choice is good and of course i'm for the progress of qemu; but, parallels with its working full usb support does for me what vpc for os/2 came just short of right now.

Parallels already dropped the OS/2 host. Most of the new features released are only supported in ms win guests with no support for OS/2.

QEMU has basic USB support which should be improved in the next release.

With QEMU, there are no features off limits to OS/2 users. If we find the resources we can always have feature parity on features that are important to us. We don't have a supplier dictating what we can have.


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  ecsguy
Date  :  06 Dec, 2006 on 07:16

Sebadoh (04 Dec, 2006 00:55):
I for one absolutely hate emulation.. and it actually doesn't do a whole lot for us, even though it is running on newer hardware it doesn't add any capabilities I didn't have otherwise. I still can't use an usb device which doesn't have OS/2 support unless there is some network mapable way to make it work which is a huge kludge.

In most cases it would have given you new capabilities you would not have had yet (or a lot sooner). Wireless NICs is one of the best examples. Wireless has been out for many many years and we still can't use OS/2-eCS with most wireless cards. There are a lot of OS/2-eCS users with laptops where they can't use the laptop's wireless card. Dual core came out over two years ago and we are only now getting support for some motherboards.

USB is not a good example. Until recently OS/2 itself did not have decent and stable basic USB drivers. Also QEMU USB support is improving.

Virtual machines might not give 100% isolation from hardware but it can help a lot.

Today it would allow
using dual/quad cores on most motherboards
Video - on almost all cards/motherboards
NIC (wired and wireless) - on almost all cards/motherboards
Sound - on almost all cards/motherboards
SATA - on almost all cards/motherboards
RAID (the inexpensive controllers) - almost all
USB - full USB is almost here

With a virtual machine you could be using new versions of most devices almost instantly instead of waiting years (or sometimes infinity) as is common for OS/2 drivers.

Today with native drivers, every time the manufacturer revs the hardware we have to tweak the native OS/2 driver (if we even have one). Most hardware makers will even change the hardware completely without using new model numbers. So OS/2 users buy something they think works in OS/2 only to find the device is a minor different rev than the one sold yesterday. Even with Scitech we had this problem all the time. Scitech would release support for a graphics chip and then the next day the hardware maker releases a "Pro" or "+" or something else new and improved rev version that the recently released Scitech version can not use.


Sebadoh:
No matter how good the emulator is it is always going to be slower then running it native

Sorry but it is not always slower. The issue for OS/2-eCS most of the time is lack of a driver that keeps you from using the new version of the device at all or maybe not even be able to boot OS/2-eCS at all.

So running OS/2 as a fast QEMU guest would be a lot faster than not running at all. And on today's hardware OS/2-eCS guests would be very fast and very useable. I have run ms winXP as guest on Linux and it worked well. If winXP worked, OS/2-eCS would scream on that same setup.

With a virtual machine any new Intel desktop or laptop that runs ms win or Linux could almost instantly run OS/2 or eComStation with OS/2-eCS being able to use most of the devices.

Wouldn't it be nice to go shopping for a system and know that you could pick most any system on the shelf and you would be able to use it with OS/2-eCS. You could have even run eComStation on an Apple PowerPC Mac. With the new Intel Macs it will be even easier.

QEMU can be downloaded from
http://fabrice.bellard.free.fr/qemu/

But we especial need a OS/2-eCS developer to help the QEMU team see what tweaks QEMU needs so the eComStation and OS/2 MCP with networking can run properly.


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  oli
Date  :  06 Dec, 2006 on 20:02

ecsguy (06 Dec, 2006 07:24):
Sorry but it is not always slower. The issue for OS/2-eCS most of the time is lack of a driver that keeps you from using the new version of the device at all or maybe not even be able to boot OS/2-eCS at all.

So running OS/2 as a fast QEMU guest would be a lot faster than not running at all. And on today's hardware OS/2-eCS guests would be very fast and very useable. I have run ms winXP as guest on Linux and it worked well. If winXP worked, OS/2-eCS would scream on that same setup.

With a virtual machine any new Intel desktop or laptop that runs ms win or Linux could almost instantly run OS/2 or eComStation with OS/2-eCS being able to use most of the devices.

Wouldn't it be nice to go shopping for a system and know that you could pick most any system on the shelf and you would be able to use it with OS/2-eCS. You could have even run eComStation on an Apple PowerPC Mac. With the new Intel Macs it will be even easier.


Sorry but you are seriously delusional, while there are nowhere as many drivers out there as I would like there is not a motherboard or portable computer available that will not take OS/2 even it it needs to be coaxed a bit, this includes the Apple Intel Mac, a number of people are dualbooting between ecs 1.2 and OSX.

They will all run OS/2 or ecs much faster than the same hardware running QEMU under linux etc. Same goes for WinXP under QEMU. the difference is not "minute".

QEMU is not fast, its not BOCHS slow, but its not upto Paraells in speed, its also less stable in it's current incarnation.

Besides QEMU does not run OS/2 at all as it is and Linux is not exactly drowning in quality driver software either even if things are not as bad as they are under OS/2.


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  Smedles
Date  :  07 Dec, 2006 on 03:10

oli (06 Dec, 2006 20:02):
Besides QEMU does not run OS/2 at all as it is and Linux is not exactly drowning in quality driver software either even if things are not as bad as they are under OS/2.

QEMU for OS/2 is being worked on. (Not by me).


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  Sebadoh sys3175@optonline.net
Date  :  07 Dec, 2006 on 04:10
That is quite a statement from a company which has released 2 applications.. I mean, it is nice to see that information on a website but I don't know if i would believe the could support snap. As is the original developers were laboring just to maintain stability. Case in point... power saving doesn't work on DVI interfaces.. very limited overlay support even though they have a long list of supported hardware.


Junior (04 Dec, 2006 23:01):
Hi all,

Been following this forum a lil n found this at GoldenCode web site :

" Device Driver Development. Golden Code can provide full design and development services for OS/2 device drivers. Once armed with the functional requirements of the driver and any associated hardware/hardware interface documentation, this team can create, modify or maintain/support virtually any device driver project. The driver development team is experienced in a wide range of OS/2 driver types including:

* both 16-bit and 32-bit physical device drivers (PDD)
* implementation of network drivers using the Network Device Interface Specification (NDIS) at both the lower (MAC) and upper (protocol) interfaces
* security enabling services installable security subsystem (SES ISS)
* installable file systems (IFS) "

Perhaps they would be prepared to take over or buy snap. Noticed how well people supported netlabs, perhaps the OS/2 community would help in supporting them as well for a good cause. I'm in as much i can possibly give within my means.



Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  obiwan
Date  :  09 Dec, 2006 on 00:14
I am sure many of you knew this already, but I didn't realize that the SNAP SDK had already been released under the GPL. That isn't the whole driver system but that means we have a lot more available to us for further development than I thought we did.

Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  oli
Date  :  10 Dec, 2006 on 19:48

Smedles (07 Dec, 2006 03:10):

oli (06 Dec, 2006 20:02):
Besides QEMU does not run OS/2 at all as it is and Linux is not exactly drowning in quality driver software either even if things are not as bad as they are under OS/2.

QEMU for OS/2 is being worked on. (Not by me).


I know, and that is nice to hear but QEMU linux does not properly run OS/2, which is why trying to coax people from using native OS/2 to using it in an emulator that does currently not work is silly, especially since Paraells is cheap and was originally designed in 1996 specifically to run os/2 and Microsoft's VirtualPC offering is now free and handles OS/2 just fine.


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  ecsguy
Date  :  11 Dec, 2006 on 03:13

oli (10 Dec, 2006 19:4:
I know, and that is nice to hear but QEMU linux does not properly run OS/2, which is why trying to coax people from using native OS/2 to using it in an emulator that does currently not work is silly, especially since Paraells is cheap and was originally designed in 1996 specifically to run os/2 and Microsoft's VirtualPC offering is now free and handles OS/2 just fine.

Microsoft VPC
1. does not support Linux host (ms win2k/XPpro hosts only)
2. no longer supports OS/2 host
3. dropped VNC server (so now OS/2 clients can not connect)

Parallels
1. Dropped OS/2 host
2. after several years still does not have server
Connectix VPC first had VNC server back in v4.x (2001?) VNC is open source. Been asking Parallels for over three years for this simiple but essential feature. Most every competitor has some type of simiple server feature (MS VPC, VMWARE, win4Lin, QENU).
3. Competing against VMWare and Microsoft which are giving away similar products for free. Both are also giving away server products for free too.
Where is server product Parallels repeated promised?
4. How long can Parallels compete against free desktop and server products from MS, VMWare and also open source (QEMU, XEN)?

XEN (another open source program)
Can now run ms winXP (using latest Intel and AMD cpus)

QEMU
1. supported on Linux, win, BSDs, MacOSX, Solaris
If enough OS/2-eCS developers are interested then OS/2-eCS could be supported in the future (don't have to beg any company)
2. SMP support on x86 (up to 255 CPUs !)
3. includes VNC server
4. Runs Warp4. Almost runs eComStation
If enough OS/2-eCS users show interest then I am sure the issues
will be looked at and fixed.

QEMU home page
http://fabrice.bellard.free.fr/qemu/


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  melf mikaelelf@os2ug.se
Date  :  11 Dec, 2006 on 09:05
I don't know much about the tecnicalitys, but with all respect of QEMU I can't see the point of running eCS as a guest OS. Since Senernity started eCS we have seen great sucess when it comes to hardware compability, intallation procedure and now bootable JFS, ACPI and also som core software, like Open Office, Lucide. I would rather shift OS than running eCS as an emulation. The devlopement so far has been promising in my opinion though the latest SciTech thing is problematic. In the longterm future It seems like the voyager projekt is promising projekt, though I really can't judge that. For my sake..no I don't want to run eCS as guest on whatever!

---
/Mikael

Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  ecsguy
Date  :  11 Dec, 2006 on 10:08

melf (11 Dec, 2006 09:05):
I don't know much about the tecnicalitys, but with all respect of QEMU I can't see the point of running eCS as a guest OS. Since Senernity started eCS we have seen great sucess when it comes to hardware compability, intallation procedure and now bootable JFS, ACPI and also som core software, like Open Office, Lucide. I would rather shift OS than running eCS as an emulation. The devlopement so far has been promising in my opinion though the latest SciTech thing is problematic. In the longterm future It seems like the voyager projekt is promising projekt, though I really can't judge that. For my sake..no I don't want to run eCS as guest on whatever!

I am glad you have all the drivers you need and the time to wait for new ones. When I do my personal machine I spend months picking the hardware so I have drivers for everything.

But please consider that there are other people that might not have the time. This is especially true with anyone trying to support a business with software based on OS/2-eCS. We can not afford to be futzing around and waiting for years for drivers. We can not tell our customer to buy used hardware. We can not tell our customer to wait for who know how long.
Businesses develop apps for OS/2-eCS but they have to be able to sell enough systems to pay the bills and feed our families. In the last couple years I know of a couple great developers that had OS/2-eCS apps who quit. They didn't want to. They loved developing for OS/2-eCS but as is proper their families come first. The hardware market changes too fast. We can't stop that and we have never been able to keep up with drivers and kernel changes. A open source virtual machine gives us another option. Without a virtual machine, eComStation will continue the slide into a hobby market only OS.

Every major OS has several commercial and open source virtual machine options. Every piece of hardware always has drivers for windows yet there are still many virtual machine options for windows. Drivers is not the only reason to want the option of virtual machines or emulators. One of the common processes in the IT server area is to run multiple OSes on one physical server. VMWare and Xen are leaders in that area. Right now neither of those support OS/2-eCS. So here is another reason that IT depts will not consider OS/2-eCS. This in fact happened just two months ago with a customer I was working with. The customer runs all servers as virtuals. They were considering the OS/2-eCS app until they found it could not run as a virtual guest on their servers.

eComStation is one of the only OSes without a fully working open source virtual machine. Besides QEMU, the XEN server virtual machine uses some QEMU code for the guests. If QEMU can run OS/2-eCS then we might also be able to get OS/2-eCS to run as a guest on XEN.


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  Cris
Date  :  11 Dec, 2006 on 15:22

I am glad you have all the drivers you need and the time to wait for new ones. When I do my personal machine I spend months picking the hardware so I have drivers for everything.
But please consider that there are other people that might not have the time.

There is one thing about virtual machines that everybody (among those who push the usage of VMs) seem to ignore:
the fact that the HOST operating system has all kind of drivers for all kind of peripherals doesn't do you ANY good when it comes to the GUEST operating system.

First of all the virtual machine has its own "emulated hardware" that need not bare any resemblance to the physical HW you're running it on; if you buy the latest and greatest peripheral to hook onto the physical box, you won't see it from inside the emulated box, unless it has specifically designed to emulate that too. And even if the virtual box correctly emulates the hardware, you still very often need a specialized driver for the GUEST operating system to take full advantage of it.

So where does this leaves us? In no better position than running on the bare HW, which is preferable anyway.

Bye


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  ecsguy
Date  :  11 Dec, 2006 on 19:31

Cris (11 Dec, 2006 15:22):
There is one thing about virtual machines that everybody (among those who push the usage of VMs) seem to ignore:
the fact that the HOST operating system has all kind of drivers for all kind of peripherals doesn't do you ANY good when it comes to the GUEST operating system.

We are not ignoring anything. In the last five years we have in fact used virtual machines to run OS/2-eCS many times on hardware we otherwise could not use.


Cris (11 Dec, 2006 15:22):
First of all the virtual machine has its own "emulated hardware" that need not bare any resemblance to the physical HW you're running it on; if you buy the latest and greatest peripheral to hook onto the physical box, you won't see it from inside the emulated box, unless it has specifically designed to emulate that too.

That might be true for many external peripherals (like printers) but it is usually NOT TRUE for internal components.

The items I listed in an earlier post are actual hardware types changes where eComstation could and can take advantage of the new hardware without a single change in the emulated VM guest or need for OS/2-eCs driver

dual core
video card
wired NIC
wireless NIC
SATA
sound

This is not theory or hype. Virtual machines are a proven technology that has been in use for over FIVE years. The VM market is one of the fastest growing markets in IT. There are good reasons why that is happening. There were really good business reasons why EMC spent 635 million dollars to buy VMWare, Microsoft bought VPC, etc


Cris (11 Dec, 2006 15:22):
And even if the virtual box correctly emulates the hardware, you still very often need a specialized driver for the GUEST operating system to take full advantage of it.

Sorry but that is not correct.

For many devices the host hardware and the emulated hardware do not need to be the same. This is a pretty standard feature of most VMs.

Also in most cases the new thing is really not that new. The new hardware is usually a revision or update of the previous model. Most of this is done for cost savings reasons by the hardware makers. Most of the time there is ZERO new functionality but the old driver no longer works. A virtual machine where Linux or mswin has the new NIC/LAN driver can still pass the tcpip packets to OS/2-eCS guest. Same with SATA controller and video. For over two years, virtual machines were helping other OSes take advantage of dual core performance even though the guest did not know it was there.

I don't understand why there is so much negative push back on this. This is not some pie in the sky futuristic theory. This is proven fact that is being used everyday on tens of thousands of machines around the world.

The issue I am bringing up is that OS/2-eCS is not able to FULLY take advantage of this powerful and common technology that every other major operating system has available. I have done a lot of testing of eComStation with QEMU. A far as I can tell just a few days to couple weeks of developers time would get us fully running on QEMU. That would be a very big return on time invested.


Cris:
So where does this leaves us? In no better position than running on the bare HW, which is preferable anyway.

Over the last 5 years there are tens of thousands of systems that have been moved to virtual machines. It was because they WERE in a better position using virtual machines than without it.

Both computer scientists and computer business market analysts are predicting that in the near future virtual machines will be standard on most desktop and server machines.


Here is a sample of the future of desktop computing -

Virtual Machine Monitors: Current Technology and Future Trends
http://www.stanford.edu/~talg/papers/COMPUTER05/virtual-future-computer05.pdf

.. excerpt ..
As the pervasive use of virtual machines moves from the server room to the desktop, their effects on computing will become even more profound.
Virtual machines provide a powerful unifying par-adigm for restructuring desktop management. The provisioning benefits that VMMs bring to the
machine room apply equally to the desktop and help solve the management challenges that large collections of desktop and laptop machines impose.

...
Suppose, for example, that a user's desktop machine is running multiple virtual machines simul-taneously. The user might have a relatively low-security Windows virtual machine for Web browsing, a higher-security virtual machine with a hardened Linux virtual machine for day-to-day work, and a still higher-security virtual machine comprising a special-purpose high-security operating system and a dedicated mail client for sensitive internal mail.

..
Admini-strators can publish virtual machines and control how these virtual
machines can be used. The Collective project explored in depth the idea of bundling applications into virtual appliances. The idea is to provide file
servers, desktop applications, and so on in a form that lets users treat the virtual machines as a stand-alone application. An appliance maintainer handles issues like patch management, thus relieving normal users of the maintenance burden.


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  Terry tgindy@yahoo.com
Date  :  11 Dec, 2006 on 21:06
eCS-OS/2 continues to be very effective in pre-emptive multi-tasking (vs. co-operative multi-tasking), networking, stability, extensibility, and the quintessential workplace shell.

And, interesting how eCS-OS/2 continues the need to be just a little more more effective in three "Vs" of development - Virtualization, jaVa, and Video.


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  obiwan
Date  :  11 Dec, 2006 on 21:26
As this is a thread about Scitech selling SNAP, any discussion of virtual machines is in the context of using them to compensate for lack of video drivers.

As the forum and site are about OS/2, the context is computing as preferred by OS/2 users. As it is being posted in December 2006, it is as the technology stands now, in the backdrop of OS/2's history so far.

I say that because it explains the strong opposition to virtual machines here, in the face of it being a generally well-accepted technology.

Video drivers are important to us because the visual desktop environment is the key experience to us that we like and want to perpetuate. Virtualizing the display in particular (as opposed to daemons, data access, and other background tasks) necessarily means a slowdown because the performance is directly linked to applications having native access to the acceleration features of the video hardware. Without that, the user experience is dampened, and there is very little point in running the OS. Virtualization could mean running the WPS desktop inside a window on some other GUI, which also utterly defeats the appeal.

OS/2 was never designed to be a catch-up platform, that looks at what everyone else is doing and tries to do it too. It was from its start an innovative reaction to what was wrong with everything else. OS/2's appeal to everyone involved declined when it became a game of catch-up. The day it becomes nothing more than an imitation of everything else is the day no-one wants to bother with it anymore. So you will have to forgive us for needing something more than "Everyone else is doing it!" to gain acceptance.

What makes OS/2 great is OS/2 itself. Throughout its history, the best performance, stability, and user experience has been achieved when things are done "the OS/2 way."So there has always been a healthy debate about whether the latest trend deserves the effort required, and whether OS/2 benefits, or rather is dilluted or even harmed, by its implementation.

Nevertheless, many of these efforts have been heroicly undertaken and achieved. Projects such as Odin and EMX have greatly enriched and perpetuated OS/2 as a unifying desktop, and at the same time caused what their detractors predicted, a transition and exodus of a huge segment of the OS/2 users most interested in those features, to the respective platforms.

Every platform has its advantages and limitations. They cannot all be everything to everyone.

That being said, most of us prefer OS/2 in all respects except one. That one thing is going to move us to migrate, try to fix it, or complain; or some combination thereof.

Migration hurts OS/2, efforts to fix it help, and complaining often helps find disagreement.

There is a place for virtualization of OS/2, but probably not as a video driver solution. Its overall benefit will be disputed by many. At this stage the job of implementing it belongs to the few who want it badly enough. Same as anything else, like new video drivers.


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  ecsguy
Date  :  12 Dec, 2006 on 12:49

obiwan (11 Dec, 2006 21:26):
As this is a thread about Scitech selling SNAP, any discussion of virtual machines is in the context of using them to compensate for lack of video drivers.

Agreed that is the main context. However the lack of other drivers (esp wireless NICs) is very similar to the lack of video drivers. So since the use of VM technology can help with lack of video drivers and NIC drivers and possibly other drivers, increased investment in VM technology that works with OS/2-eCS can help many OS/2-eCS users.


obiwan (11 Dec, 2006 21:26):
As the forum and site are about OS/2, the context is computing as preferred by OS/2 users. As it is being posted in December 2006, it is as the technology stands now, in the backdrop of OS/2's history so far.

I say that because it explains the strong opposition to virtual machines here, in the face of it being a generally well-accepted technology.


Sorry I still don't understand the strong opposition. Having better VM support does not keep anyone who has drivers from running OS/2-eCS natively.

I am an OS/2-eCS user and so is anyone else running OS/2-eCS as a guest under a VM. Many of us would prefer to run with native drivers but under certain situations it is not possible. So our choice is to run OS/2-eCS under VM/emulation or not run OS/2-eCS at all. I would think having more OS/2-eCS users is better for all OS/2-eCS users.

If most current and long time OS/2-eCS users have the drivers they need or can wait for them that is great. However there are many that don't have the drivers they need. Lack of OS/2-eCS driver is usually the biggest issue for new OS/2-eCS users.

What about the hundreds of millions of people whose hardware is lacking OS/2-eCS drivers. All those machines haver either mswin or Linux drivers available. Every year another 50 million machines are made.without OS/2-eCs drivers.

Last week I had to reload windows on one of my customer's machines. The machine is a Intel dual core with 1GB of memory and 120GB harddrive. There is no OS/2-eCS drivers for the video, NIC and SATA controller. Don't know if OS/2-eCS will recognize and use rthe dual core.

So the choices are
1. run windows on the machine
2. run Linux on the machine
3. run OS/2-eCS as guest on Linux
(OS/2-eCs as guest will probably run better than mswin does natively)

Which choice helps the OS/2-eCS community?

Here is another example from early this year.
A family has one computer but everyone in the family wants to run mswin except for one person that wants to run eComStation.

The choices are
1. everyone run mswin (sorry no eCS, eCS user loses
2. run mswin and have eCS as guest.
(Dual boot is not an option because the mswin users will not wait for boots. Remember it seems like win takes forever to boot.)

Which choice is better for peaceful family life?
Which choice helps the OS/2-eCS community?


obiwan (11 Dec, 2006 21:26):
Video drivers are important to us because the visual desktop environment is the key experience to us that we like and want to perpetuate. Virtualizing the display in particular (as opposed to daemons, data access, and other background tasks) necessarily means a slowdown because the performance is directly linked to applications having native access to the acceleration features of the video hardware.

Details are not always important. What is the bottom line? What is the net performance? We usually don't write programs in assembler any more even though a accounting program written in assembler can be much much faster.

When I put a $2 12 year old 4MBPCI video card in a modern machine where I don't have a OS/2-eCs video driver for the onboard video, OS/2-eCS runs great even though I have 0% utilization of the NEW 10 times faster 64MB AGPx8 video card.

OR I could run OS/2-eCS as virtual machine guest. Yes, it is true the video is slower than if I had a native video driver. I would love to have a native driver but I don't. But the video speed in the OS/2-eCS guest is just as fast as the 12 year old PCI card. Net bottom line is that OS/2-eCS still runs fast.


obiwan (11 Dec, 2006 21:26):
There is a place for virtualization of OS/2, but probably not as a video driver solution.

Why not?

Already today virtualization can many times be as fast as native (12 year old video card).

There is still a lot of work being done in the virtualization field. It is entirely possible that in a few years virtualized video will run at 90% of the speed of MODERN NATIVE video.

Connectix VPC had a special video driver for use in OS/2-eCS guest. Once base OS/2-eCS runs fully in QEMU then maybe some developers can look at optimizing the video.


obiwan (11 Dec, 2006 21:26):
Its overall benefit will be disputed by many. At this stage the job of implementing it belongs to the few who want it badly enough. Same as anything else, like new video drivers.

But there are a lot of OS/2-eCs users that do not know that virtual machine technology can allow them to run OS/2-eCS nicely on today's newest machines.

Most OS/2-eCS have friends that have ms windows machines that can not run OS/2-eCS natively. Would be nice to get some of them as OS/2-eCS users.

QEMU
http://fabrice.bellard.free.fr/qemu/


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  ecsguy
Date  :  12 Dec, 2006 on 17:55
Virtualization just got faster. part of QEMU is now in Linux kernel.

http://www.heise.de/english/newsticker/news/82344/from/rss09


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  Cris
Date  :  12 Dec, 2006 on 20:48
ecsguy, I'm sorry but this is complete and utter crap.

1st: virtual machines are getting more and more market share because of enterprises, not because of home users. enterprises want to consolidate hardware resources and make good use of all the power today's machines have. But enterprises DON'T have the needs that home users have (e.g. usually no need for fully accelerated 3D graphics). Home users like VMs because they can try and use other OSes within their favorite OS, but if it was only for this kind of usage you wouldn't see the present rate of development on the VM front.

2nd: VMs don't let you use HW for which you have no drivers. Take the problem of video drivers: in a VM you have an emulated video card that is quite generic, and don't support all of the features of the HOST video card. Yes, I can run OS/2 as a guest if I have the driver for the emulated video card, but it won't use the features of the HOST video card. If a user has to choose among using OS/2 as a guest on an underfeatured emulated video card, and running Windows natively with full 3D acceleration and all the bells and whistles, what do you think the user would choose? Do you think that running OS/2 this way (i.e. as a guest with an emulated video card) is really all that different than running OS/2 natively on that machine with the VESA driver?
Another example: multi-cored CPUs. Yes, I can run OS/2 as a guest on a multi-cored CPU, but does it use the full power of the multiple cores? No. It will use *some* of that because the VM is capable of using multi cores for its internal things, but you won't get the features that are associated with OS/2 running in native SMP mode. Again: what would the user choose? An OS natively using the full power of the multiple cores or an OS running in a VM which doesn't see all the cores?

You say that having eCS running in QEMU doesn't require muc work. OK then, if you feel it is worth your time, you're free to pursue your goal. It surely will benefit eCS-OS/2. But not in the way you think.
You say that you don't understand the resistance to the use of VMs. You are misinterpreting: there is no resistance to the use of VMs in general, but there is great resistance to the view that the future of OS/2 lies in being used as a guest OS.
I will be migrating to another OS soon before OS/2 becomes a VM-only OS. Even those who state that they will use eCS-OS/2 as a guest in VM will soon find themselves using the HOST OS more and more, until the point they will not see the need to use the VM at all.

That's not what I want to see for the future of OS/2. No, thanks.

Bye
Cris


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  ecsguy
Date  :  13 Dec, 2006 on 09:31

Cris (12 Dec, 2006 20:4:
ecsguy, I'm sorry but this is complete and utter crap.

Interesting, all I am doing is describing what myself and others have actually done in the past and how it could be used to solve a new problem (Scitech SNAP) and other recurring problems.


Cris (12 Dec, 2006 20:4:
1st: virtual machines are getting more and more market share because of enterprises, not because of home users. enterprises want to consolidate hardware resources and make good use of all the power today's machines have. But enterprises DON'T have the needs that home users have (e.g. usually no need for fully accelerated 3D graphics).
Home users like VMs because they can try and use other OSes within their favorite OS, but if it was only for this kind of usage you wouldn't see the present rate of development on the VM front.

There are many home users using VM everyday to do something that thier primary OS does not do or because there are multiple people in the family with different needs or desires.

run OS/2-eCS
run ms office
run new flash version
run new acrobat version
run some other win app
....


Cris (12 Dec, 2006 20:4:
2nd: VMs don't let you use HW for which you have no drivers.

Again that is not a true statement and I have already corrected it several times. This ability is a one of the main features of virtual macines.

The user could use a VM to run OS/2-eCS and use the new/improved part. Maybe the user can not use all the new/improved features. But at least the user can use the device to some degree. In most cases it is enough. All the user wanted were the devices basic standard features to work and is not interested in the optional new/improved/updated features.

In these cases the OS/2-eCS users DOES NOT NEED ANY NEW DRIVER IN THE GUEST.

This usually works for
video
wired NICs
wireless NICs
SATA disk controller
sound cards


Cris (12 Dec, 2006 20:4:
Take the problem of video drivers: in a VM you have an emulated video card that is quite generic, and don't support all of the features of the HOST video card. Yes, I can run OS/2 as a guest if I have the driver for the emulated video card,

Not having a driver for the emulated video card in the guest has never been a problem. The first production VM that supported OS/2-eCS was Connectix Virtual PC that shipped over 5 years ago (Oct 2001). OS/2-eCs had a driver then and OS/2-eCS has never lacked a driver for any of the other emulated video cards as far as I know.


Cris (12 Dec, 2006 20:4:
but it won't use the features of the HOST video card. If a user has to choose among using OS/2 as a guest on an underfeatured emulated video card, and running Windows natively with full 3D acceleration and all the bells and whistles, what do you think the user would choose?

From the dozens of OS/2-eCS home computer installs that I have seen and worked on no one had to make a choice that caused them to lose a feature they wanted.

The emulated video is a S3 Trio. That emulated card supported all of the important features that most users rely on today even with the latest "super turbo gazillion megabyte AGP8/PCI16-express" whatever video cards that are in today's bleeding edge machines.

Connectix VPC video (Oct 2001) - S3 Trio 32/64 PCI with 8MB VRAM

What most users seem to find are "must haves"
1. is it fast enough to keep up with their movements
2. resolution (1280x1024)
3. number of colors (64K)
4. refresh (76Hz)
5. work with monitor
6. no corruption
7. no lockups
8. no traps

Options that some user have used
zoom and pan - supported by S3 Trio
TV out - is driver issue that does not matter to host or guests
DVI - is driver issue that does not matter to host or guests

Hardly anybody uses this one
multi head - not supported by emulated driver

The only feature the S3 Trio does not support is dual head. I can count on one finger the number of users I have worked with that used dual head.

The problem with running natively is when no native video driver exists or it does not work properly. In that case the users does not have the "must haves".

In many cases a VM DOES provide a immediate solution to these VIDEO problems.


Cris (12 Dec, 2006 20:4:
Do you think that running OS/2 this way (i.e. as a guest with an emulated video card) is really all that different than running OS/2 natively on that machine with the VESA driver?

Yes, running emulated can be different and better than running a native driver in VESA mode becuase VESA mode does not work properly with all adapter/monitor combinations. The only video mode that almost always works is 8 color VGA (640x480).


Cris (12 Dec, 2006 20:4:
Another example: multi-cored CPUs. Yes, I can run OS/2 as a guest on a multi-cored CPU, but does it use the full power of the multiple cores? No. It will use *some* of that because the VM is capable of using multi cores for its internal things, but you won't get the features that are associated with OS/2 running in native SMP mode.

If OS/2-eCs, when run natively does not recognize or use the dual core then for sure half the processing power is sitting idle. That has been true for the over two years that dual core has been out.

With a VM the OS/2-eCS could have been using both cores from since dual cores became commonly available (over two years ago).


Cris (12 Dec, 2006 20:4:
Again: what would the user choose? An OS natively using the full power of the multiple cores or an OS running in a VM which doesn't see all the cores?

I have not disagreed with that. If OS/2-eCS natively recognizes the dual cores it is probably the best solution. But for over two years there was NO native support.

Again there have been and still are many OS/2-eCS users with wireless cards that were/are unused because they lack native drivers. With VM some of those users are able to use the hardware they paid for.


Cris (12 Dec, 2006 20:4:
You say that having eCS running in QEMU doesn't require much work. OK then, if you feel it is worth your time, you're free to pursue your goal.

I am doing what I can. Since I currently do not have the particular type of programming skills needed I have to assist the community in other ways.


Cris (12 Dec, 2006 20:4:
It surely will benefit eCS-OS/2. But not in the way you think.
You say that you don't understand the resistance to the use of VMs. You are misinterpreting: there is no resistance to the use of VMs in general, but there is great resistance to the view that the future of OS/2 lies in being used as a guest OS.

This not not only the future of OS/2-eCS. It is the future of most desktop OSes. Just like the first OS/2 with WPS (OS/2.2.0) on 3.5 diskettes ran on a Intel 386 with 4MB, the software internals and hardware of tomorrows OS/2-eCS might also be much different.

Most desktop processors that Intel and AMD will ship in 2007 and beyond will have new cpu instructions specifically to help support virtualization on desktop systems (not servers). These processors are shipping today!

-------------------------------

http://www.cgs4u.com/fall_technology_preview.aspx

Fall 2006 Technology Preview

AMD's Athlon64 AM2
... excerpt..
Virtualization on desktop computers allows a single PC to act like multiple virtual machines. AMD Virtualization can enable client computers to seamlessly support multiple operating environments.

--------------------------------

more example of virtualization used on home PCs

http://technews.download3000.com/Columnists/5232/Security-Adviser-Virtual-concerns.html

December 1, 2006, 6:00 am

..... excerpt..
As a long-time traveling presenter and lab teacher, I used to have to fly with two or three PCs. Now, I carry my entire Windows forest or Linux realm on a laptop. I can start up four or five servers in about two minutes. It's like ordering lunch: Two or three Windows servers, a few Linux servers, and a Solaris server to go, please.

Classroom shutdown is now a single power down. No cords to unplug. No server hardware to pack. Getting ready for the next class is a snap: revert and I?m ready to go. Today's young computer teachers have no idea how hard it used to be.

But what's coming on the virtual forefront is even more revolutionary. I know of one company that's going to allow its employees to work from home using virtual images. The company will send the entire corporate image to the employee over a VPN connection, or at worst, on a single DVD.

This means the employee can run their own home computer in an insecure state, and the company doesn?t worry about it because the work image is locked down and reverted at each new restart. Documents and company databases are stored on a centralized storage server. The company?s firewall only allows one map drive connection into their physical environment; all other inbound ports are closed. That's a pretty tight firewall.

--------------------------------

http://oncomputerstips.blogspot.com/2006/03/jack-eats-humble-pie-re-virtualization.html

Wednesday, March 15, 2006

Jack Eats Humble Pie re; Virtualization on Desktop

... excerpt ..
It hasn't been all that long since I poured verbal cold water on the idea of virtualization on the desktop becoming common. I've been studying virtualization both as a technology and as a market force now for several weeks and I find I must admit I was wrong and just how wrong I was.
.....
Basically, it's a one gigabyte flash memory stick with a twist; You can take enough of your home Windows desktop, including settings, software and data, with you to work in your chosen environment and with your choice of tools on any other Windows PC. And when you leave that PC, none of your data or settings stay behind.
...
Imagine that when you buy or build a PC, there is no operating system as we know it installed. Instead, there will be a virtual machine or "hypervisor", which is the current buzz-word for the host OS in a virtualization scheme. And we can carry a device very much like the Verbatim one, only larger, which contains our chosen operating environment, data and whatever software we choose, making it possible to work on our own PC without our own PC! Plug in your USB device, give it a minute (probably a lot less than a minute) to boot up and you're ready to go. Anything too large to be carried with you on the device would be available over the Internet using a built-in VPN setup to access your home machine.
....
If the hypervisor is well implemented, any user could use literally any work environment they chose on any PC. Like Linux? How about the Mac OS X? You could run them all on the same hardware. No problem at all.

---------------------------------

http://www.extremenano.com/print_article/Virtualization+Headed+Toward+Client+Space/175088.aspx

Virtualization Headed Toward Client Space
By Jeffrey Burt
4/5/2006 12:26:00 PM
BOSTON-The paths for both Linux and virtualization, which have gained ground in the data center, are headed directly toward the client space, according to Dell's top technology executive.

In his keynote here at the LinuxWorld Conference & Expo Wednesday morning, Dell Chief Technology Officer Kevin Kettler said that, up to this point, most of the development of virtualization technology has centered on the needs of server users.

"We at Dell think that is about to change," Kettler said. "What we see is an opportunity to really drive virtualization, and the capabilities of Linux as well, into the client."

Virtualization enables users to run multiple operating systems and applications on a single physical machine through the use of virtual machines. The same convergence of developments-from multicore processors to hardware-based virtualization from chip makers Advanced Micro Devices and Intel to declining implementation costs-that are fueling the drive of virtualization in the data center will do the same in the client space, Kettler said.

Virtualization will enable PC users to create multiple dedicated environments on a single machine that separate such aspects as Web browsing, gaming and media centers from one another, Kettler said. Keeping such environments isolated will protect them from one another-if one is infected by a virus, it can easily be removed without harming any of the other environments.

Virtualization also allows a user to run multiple operating systems-from Linux to Microsoft Windows-on a single machine, freeing the operating system from the hardware.

---------------------------------


Cris (12 Dec, 2006 20:4:
I will be migrating to another OS soon before OS/2 becomes a VM-only OS.

That is of course your choice

There might be other OS/2-eCS that decide that what is in the blackbox engine is not important as long as it is reliable and responsive. They want to interact and interface with OS/2-eCS which is what it will still be.

QEMU
http://fabrice.bellard.free.fr/qemu/


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  Dobber
Date  :  13 Dec, 2006 on 13:37
"The only feature the S3 Trio does not support is dual head. I can count on one finger the number of users I have worked with that used dual head"

Hmmm... I guess all these graphics card makers have added multi head cards to their product lines for that one user? Get real. In the real world, it is common to deploy multi head systems to users. CAD, programming, call centers, operation centers... the list goes on. As people multitask more and more on their desktops, screen real estate becomes VERY valuable.


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  Sebadoh sys3175@optonline.net
Date  :  13 Dec, 2006 on 19:23
I guess I must be the one finger you are talking about, I have OS/2 running on an Ati 9250 (might actually be newer I have too many ati cards to remember which is in which pc) running 2 20in lcd's being run at 1600x1200. Tell me an emulator that is going to support that =) And most of the things you say OS/2 lacks while you might be correct about Warp4 have been added YEARS ago. And I don't put 4 year old video cards in my OS/2 machines.. I put current products which are currently being sold. They may not be the MOST recent.. but they are current.. No matter how this is cut we really need to hope someone makes a deal with scitech to continue the OS/2 version or that the company who purchases the product allows SSI to maintain the OS/2 version.

And once again.. hardware virtualization and software virtualization are two DIFFERENT things.. Amd's going the IBM approach of hardware virtualization, a software emulator which is loaded in another OS and hosts yet another OS is just a kludge. It uses too much memory, will always be slower then direct hardware access and will always have significant limitations. Hardware virtualization allows 2 or more OS to run on the same hardware w/o emulation.


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  ecsguy
Date  :  14 Dec, 2006 on 13:57
It is important to remember that what a technology is capable of is different from what a business will decide to provide.

In the article below Parallels mentions that there will be significant video performance improvements soon in the ms win guest. Corp/enterprise customers must be asking for those 3D gaming video features since everybody knows "Home" users hardly ever use virtual machines.

Also there are other people that noticed that even today VM performance can be competitive with "real" hardware.
Parallels- Rudolph: Parallels Desktop for Mac works on Intel-Macs only and offers performance that rivals - and sometimes exceeds - that found on a "real PC".

I personanlly will not hold my breath waiting for these video enhancements to show up in the Parallels VM OS/2 guest.

If business decisions are different than what the OS/2-eCS community wants, there is an alternative. We can start now or keep waiting while falling further behind the other OSes.
QEMU http://fabrice.bellard.free.fr/qemu/

--------------------------------------------------------------------------------------

http://www.insidemacgames.com/features/view.php?ID=486

Parallels to offer virtual OpenGL and DirectX in 2007

Thursday, December 14, 2006

... excerpt ..
Parallels- Rudolph: ...
One of the features missing in Parallels I believe is 3D gaming support for games such as Battlefield 1942, Unreal Tournament, etc.

Inside Mac Games: Do you plan to implement this key feature in the future?

Parallels- Rudolph: This is something that we're already working on. The goal is to have OpenGL and DirectX support in our next version, which should be in beta around the turn of the year...


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  Terry tgindy@yahoo.com
Date  :  14 Dec, 2006 on 18:30
...Just ran across these virtualization articles with a specific Linux perspective that may assist some of our eCS-OS/2 virtualization & video porting perspectives. The third article dwells upon QEMU:

December 5, 2006 - The Penguin in the Sandbox (Part 1)...
http://www.enterprisenetworkingplanet.com/netos/article.php/3647401

"The basic concept in all virtualization technologies is sandboxing."

December 12, 2006 - The Penguin in the Sandbox (Part 2)...
http://www.enterprisenetworkingplanet.com/netos/article.php/3648796

"QEMU and Bochs are emulators that replicate hardware functions in software, even hardware that is not physically present. So you can run unmodified guest operating systems in all manner of virtual hardware configurations. QEMU is GPL except for the QEMU Accelerator Module, which is a proprietary product. Bochs is LGPL."

January 24, 2006 - QEMU: Maybe a VMWare Killer, Maybe Not...
http://www.enterprisenetworkingplanet.com/netos/article.php/3579721

Note (2) QEMU modes: Full system emulation, or user mode emulation.

Also - "Which is brings us to the first QEMU gotcha: QEMU is GPL, but kqemu is closed-source and proprietary. The terms of use are liberal, but because of the license you won't see it in any Linux distributions; you have to install it from sources, then load a kernel module. Whee."

QUESTION #1: What do the (2) QEMU licensing differences, and (2) mode development choices mean for eCS-OS/2 porting?

QUESTION #2: How would this tie into eCS-OS/2 video emulation?


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  oli
Date  :  14 Dec, 2006 on 19:30

ecsguy (11 Dec, 2006 03:22):
Microsoft VPC
1. does not support Linux host (ms win2k/XPpro hosts only)
2. no longer supports OS/2 host
3. dropped VNC server (so now OS/2 clients can not connect)


That was not the point, you where talking about moving from OS/2 to another environment that virtualised a PC so you could run eCS on it, if that was the case Windows with VPC is a that actually supports OS/2, Microsoft actually turned up for Warpstock 2006 to show it running OS/2, QEMU does not run OS/2 so there is nothing to move to, period.




Parallels
1. Dropped OS/2 host
2. after several years still does not have server
Connectix VPC first had VNC server back in v4.x (2001?) VNC is open source. Been asking Parallels for over three years for this simiple but essential feature. Most every competitor has some type of simiple server feature (MS VPC, VMWARE, win4Lin, QENU).
3. Competing against VMWare and Microsoft which are giving away similar products for free. Both are also giving away server products for free too.
Where is server product Parallels repeated promised?
4. How long can Parallels compete against free desktop and server products from MS, VMWare and also open source (QEMU, XEN)?


Again this has nothing to do with your original argument, fact is Parallels runs os/2 like a dream, however running it nativly is still a better option


XEN (another open source program)
Can now run ms winXP (using latest Intel and AMD cpus)


XEN is a hypervisor, also referred to as a virtual machine monitor a totally different technology to QEMU, VPC et al. It should be noted that it also does not support os/2 although that is a moot point since XEN and hypervisors is not what was discussed


QEMU
1. supported on Linux, win, BSDs, MacOSX, Solaris
If enough OS/2-eCS developers are interested then OS/2-eCS could be supported in the future (don't have to beg any company)
2. SMP support on x86 (up to 255 CPUs !)
3. includes VNC server
4. Runs Warp4. Almost runs eComStation
If enough OS/2-eCS users show interest then I am sure the issues
will be looked at and fixed.


Yeh, "runs Warp4" as long as you do not do anything in there in which case it bombs, and so does the installer in most cases giving you a somewhat limited options system as to install it, QEMU does not reliably run any version of os/2 and in fact most versions will not run at all, you where advocating moving to a linux based system running QEMU as an alternative to running it on native hardware, even though even you admit that this is not possible at this moment in time. The availability of VNC or compatible technologies is a moot point since presumably if the people moving their personal system to another OS would primarily be interested in being able to run it on their own system.

Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  oli
Date  :  14 Dec, 2006 on 19:37

ecsguy (11 Dec, 2006 10:0:
eComStation is one of the only OSes without a fully working open source virtual machine. Besides QEMU, the XEN server virtual machine uses some QEMU code for the guests. If QEMU can run OS/2-eCS then we might also be able to get OS/2-eCS to run as a guest on XEN.

XEN is not a virtual machine, see above post

This is getting to be one of the most clueless technical discussion I have partaken in since I ran a Linux school back in the 90's, at the least back then the average age of the students was 16, so allowances could be made.....


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  obiwan
Date  :  14 Dec, 2006 on 19:43

Terry (14 Dec, 2006 18:30):
January 24, 2006 - QEMU: Maybe a VMWare Killer, Maybe Not...
http://www.enterprisenetworkingplanet.com/netos/article.php/3579721

Note (2) QEMU modes: Full system emulation, or user mode emulation.

Also - "Which is brings us to the first QEMU gotcha: QEMU is GPL, but kqemu is closed-source and proprietary. The terms of use are liberal, but because of the license you won't see it in any Linux distributions; you have to install it from sources, then load a kernel module. Whee."

QUESTION #1: What do the (2) QEMU licensing differences, and (2) mode development choices mean for eCS-OS/2 porting?

QUESTION #2: How would this tie into eCS-OS/2 video emulation?


While parts of this article may reveal some things I did not know about QEMU, the author is mincing terminology in a way that shows she is somewhat confused, and she also makes false statements like QEMU has been ported to OS/2. So I would prefer not to waste time with a defective article and refer directly to the QEMU documentation. Once some clear facts are found there, we can discuss the implications for eCS-OS/2 and video emulation.


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  Cris
Date  :  15 Dec, 2006 on 12:12
ecsguy, you're actually making some points that counter your belief in VMs:


ecsguy (14 Dec, 2006 14:29):
It is important to remember that what a technology is capable of is different from what a business will decide to provide.

In the article below Parallels mentions that there will be significant video performance improvements soon in the ms win guest. Corp/enterprise customers must be asking for those 3D gaming video features since everybody knows "Home" users hardly ever use virtual machines.


Yes, and do you know what that means? That you need specialized guest drivers (so called "additions" in VPC speak) to achieve good performance. Windows users can rest assured that the best additions will be developed for them... what about OS/2 users? You can conclude the exercise yourself.

BTW, if just one (big enough) enterprise calls for a feature (and has a sufficiently large number of licenses) you can bet the sofware house will run to implement them. OpenGL can be useful to some kinds of enterprises (for example for graphical simulations). I said there are few enterprises interested, not nobody.



Also there are other people that noticed that even today VM performance can be competitive with "real" hardware.
Parallels- Rudolph: Parallels Desktop for Mac works on Intel-Macs only and offers performance that rivals - and sometimes exceeds - that found on a "real PC".

That's the nicest quote I ever read. Virtualized HW that "exceeds" real HW? You must be kidding man.
What it means is that Intel-Macs exceed the average PC in terms of performance, and so a VM running on them _can_ exceed performance of an average PC. But on the same HW, a VM will always be inferior to the bare metal, or at best on par.
But even if on par, you are left with a virtualized box which does not match your physical box.


I personanlly will not hold my breath waiting for these video enhancements to show up in the Parallels VM OS/2 guest.

Exactly. So?... Forget about running OS/2 as guest with decent performance.


If business decisions are different than what the OS/2-eCS community wants, there is an alternative. We can start now or keep waiting while falling further behind the other OSes.

I'll wait, thanx. BTW, saying that your thoughts=business decisions is at least misleading.

Bye


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  ecsguy
Date  :  15 Dec, 2006 on 19:51

Cris (15 Dec, 2006 13:57):
ecsguy, you're actually making some points that counter your belief in VMs:


ecsguy (14 Dec, 2006 14:29):
It is important to remember that what a technology is capable of is different from what a business will decide to provide.

In the article below Parallels mentions that there will be significant video performance improvements soon in the ms win guest. Corp/enterprise customers must be asking for those 3D gaming video features since everybody knows "Home" users hardly ever use virtual machines.


Yes, and do you know what that means? That you need specialized guest drivers (so called "additions" in VPC speak) to achieve good performance. Windows users can rest assured that the best additions will be developed for them... what about OS/2 users? You can conclude the exercise yourself.


I have not contradicted myself. I said that if the choice is no native driver or using a VM that using VM is probably better. With the VM THERE IS A BASIC driver for video, NIC, disk controller, sound. Those basic drivers has been around for years without needing changes to allow OS/2-eCS guest to access standard basic functionality. With the basic driver in the VM you can at least USE the OS/2-eCS guest with the new device (video, NIC, wireless, sound, disk controller) for the normal basic functions that most people have as minimum requirements. Also the performance of those basic functions is usually adequate for normal usage (probably not games or CAD, etc ). I agreed that the performance was probably going to be slower than a native driver. However IF there is no native driver the comparison is meaningless. With a VM you can get use out of most devices at a decent performance level vs NO USE (no native driver).

I have used dual head before and also set it up for a client. It is very nice but it is not something most users do even on windows. Dual head is not a basic function. 3D was not a common feature in the past and nnow it is (at least for windows users) It is understandable that the OS/2-eCS guest vidoe driver does not support 3D when even after all these years most OS/2-eCS native video does not support 3D functions.


Cris (15 Dec, 2006 13:57):
BTW, if just one (big enough) enterprise calls for a feature (and has a sufficiently large number of licenses) you can bet the software house will run to implement them. OpenGL can be useful to some kinds of enterprises (for example for graphical simulations). I said there are few enterprises interested, not nobody.

And they usually discuss these enterprise features in game magazines?


Cris (15 Dec, 2006 13:57):



Also there are other people that noticed that even today VM performance can be competitive with "real" hardware.
Parallels- Rudolph: Parallels Desktop for Mac works on Intel-Macs only and offers performance that rivals - and sometimes exceeds - that found on a "real PC".

That's the nicest quote I ever read. Virtualized HW that "exceeds" real HW? You must be kidding man.


Nope, I am not kidding and am I am sure Rudolph wasn't either. Especiailly in an interview with a gamer's magazine. Many gamers have hardware that make number/data crunching scientists drool.

It is in fact true. Performance of VM guests on today's desktop machines can be competitive and SOMETIMES CAN BE FASTER than native.

Here is another one that most people think is impossible but does in fact happen - "accessing data across a network can be faster than using a standard local disk drive. There are many factors that affect performance. The interactions of these factors do not always give the results we usually assume."


Cris (15 Dec, 2006 13:57):
Exactly. So?... Forget about running OS/2 as guest with decent performance.

Nope, not necessarily true. Depends on the users hardware, some users already can get decent performance TODAY running OS/2-eCS guest in a VM. Hardware technology is improving so fast that what seemed impossible yesterday is doable today and is a common everyday thing tomorrow. Yesterday I bought a bowl of soup ($40, shark fin) that cost more than the price of a new PC DVD+-RW burner ($37) today.

10 years ago it was impossible for a PC to run a VM because the hardware was too slow. 5 years ago the hardware was fast enough for servers. Now today everyday common machines that are sold to even home users are fast enough to run OS/2-eCS guest at DECENT speed. If the user has native OS/2-eCS drivers then that is optimal. If they DO NOT HAVE native drivers they could run OS/2-eCS as guest in VM. And the home machines sold in 2007 will run the VM even better.


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  dlundh dlundh@home.se
Date  :  29 Dec, 2006 on 16:07
Before just going out on a limb and proclaiming that running OS/2 in a virtualized environment is too slow and basically crap, have you actually tried it?

I run eCS 1.2R under Parallels on my Mac mini (1.83GHz Dual Core, 2GB RAM) and I have to say that performace is very snappy. It would probably be even better if I splashed the VM with a lot of RAM but OS/2 does just fine without being RAM-greedy unlike Windows 2000 where I add a lot more RAM to get the same level of performance. And while I don't know what kind of speed OS/2 would have on the bare metal of my machine (and I never will, there are no drivers, period) I'm more than satisfied with the speed I get thru Parallels.

It just works, and it's fast enough for what I do. Now, YMMV, but to dismiss it as crap without at least giving it a try sounds pretty dumb to me.


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  melf mikaelelf@os2ug.se
Date  :  29 Dec, 2006 on 20:28
I really don't know where this discussion is taking us, sometimes it seems pointless. For my sake it is simple; I don't want and I can't see a point in running my favourite OS as a guest on whatsoever. I've run w2k as guest on eCS on SVISTA for the purpose of using dreamweaver and photoshop and its really acceptable, and probably eCS runs better as guest, but in the end...an OS as a guest is just a guest OS and not a working solution. And again, though I can't judge it: voyager seems to be the realistic longterm solution.
/Mikael

Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  ecsguy
Date  :  03 Jan, 2007 on 09:12
http://www.virtualizationdaily.com/archives/86_the-os-is-under-attack-continuing-the-conversation.html

The OS is under attack ... continuing the conversation
Published December 6th, 2006

... excerpt..
So it's really hard for me to see replacing Mac OS X with a hypervisor, unless that was an Apple designed hypervisor that's heavily optimized for the sharing of hardware. Today that doesn't exist, BUT just a year ago, neither did the possibility of running Windows at NEAR NATIVE SPEEDS on your Mac.


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  ecsguy
Date  :  03 Jan, 2007 on 09:18
Another eCS user says running eCS in a VM is sometimes better than running native with no driver.

--------------------------------------------------------------------------------------------------

http://tech.groups.yahoo.com/group/eComStation/message/73665

Subject [eCS] Re: APM

On Sat, 02 Dec 2006, "jmurthy" <jmurthy@...> wrote:

Thus it is equally irritating to constantly hear that one should just go to ebay and buy old hardware. I use my laptop for some rather CPU and RAM intensive work and need modern systems.

It's all about choice but it appears to me that there are fewer and
fewer options to keep eCS, particularly on laptops.

Waiting for two years after buying a new laptop to use wireless is not worth it at this stage.

Right now it may be that the best option is to get a Mac with Parallels. That way the hardware would work right out of the box and I would be able to maintain my applications the most important of which would be Wordpro and Mesa 2.

Jayant


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  ecsguy
Date  :  03 Jan, 2007 on 09:39
Must be a lot of servers running "64bits sound cards and multiple monitors"

------------------------------------------------------------------------------------------------

http://www.virtualization.info/2006/11/vmworld-2006-round-up.html

VMware launches Workstation 6.0 public beta
Wednesday, December 27, 2006

... excerpt..
# New physical hardware support
Support for USB 2.0 devices, 64bits sound cards and multiple monitors

# Headless mode
Virtual machines can now run in background, without the VMware interface running

# VNC Remote Control
Virtual machine can now be controlled through VNC instead of using guest OS remote management tools (no need to install VNC server inside the guest OS)

------------------------------------------------------------

http://www.parallels.com/products/desktop/beta_testing/

Parallels Desktop for Mac Update Beta3 (Build 3106)

... excerpt..

* NEW! USB 2.0 support - "Plug and play" popular USB devices like external hard drives, printers, and scanners, and use them at full native speed.

* NEW! Full-feature virtual CD/DVD drive - Burn CDs and DVDs directly in virtual machines, and play any copy-protected CD or DVD just like you would on a real PC

-------------------------------------------------------

QEMU is the only VM that gives the eCS community a chance to stay even with the major OSes.
http://fabrice.bellard.free.fr/qemu/


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  Cris
Date  :  03 Jan, 2007 on 10:05
Hi all,
I have lost any interest in this discussion, as all the pro-virtualization folks seem to ignore the main points every time, focusing on the glitches.

I completely agree with melf: I see no point in running anything as a guest on another OS, unless it is an occasional run just to use something that's not available natively.

dlundh: I run eCS every day under M$ VirtualPC at work, since it's the only way I'm authorized to run another OS (I'm forced to use WinXP Pro at work). I run it to maintain some of my software projects when I have no time at home; I do it because I am FORCED to, and I know this is NOT the way I would like to use OS/2 (or any other OS btw... not as my main OS).
OTOH, I wonder if ecsguy is running eCS virtualized. I'd say no.

ecsguy: you're always taking Parallels as an example, and then you say "QEMU is the only VM that gives the eCS community a chance to stay even with the major OSes"

But this is not the main point. The main point is: WHY OH WHY should I use a guest OS as my main OS when I have a NATIVE OS (HOST OS) that fully exploits all of my hardware, and does it better then any guest OS could possibly do?
This is NOT the way to the future of OS/2: it could be a way of protecting your investment and to keep on using OS/2 while you prepare for a switch. It surely is NOT the way to attract new users or to keep the current users happy.
Again I dare you: try to use OS/2 as a guest OS for a reasonable amount of time: you'll find that you'll be working in the HOST OS more and more until the point where you'll find it more convenient to simply dump the GUEST.

That's my last word on this thread. Period.

Goodbye


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  Sebadoh sys3175@optonline.net
Date  :  11 Jan, 2007 on 22:32
You could just install winXP on your mac and run it AT native speeds.. I agree, this discussion is pointless. If I had to own a mac (which cost way too much) to run OS/2 i would just use MacOS. If I had to have winxp on my machine to run OS/2 I would simply USE winXP.


ecsguy (03 Jan, 2007 09:12):
http://www.virtualizationdaily.com/archives/86_the-os-is-under-attack-continuing-the-conversation.html

The OS is under attack ... continuing the conversation
Published December 6th, 2006

... excerpt..
So it's really hard for me to see replacing Mac OS X with a hypervisor, unless that was an Apple designed hypervisor that's heavily optimized for the sharing of hardware. Today that doesn't exist, BUT just a year ago, neither did the possibility of running Windows at NEAR NATIVE SPEEDS on your Mac.



Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  ecsguy
Date  :  15 Jan, 2007 on 22:42
Virtualization continues to get closer to native speed.

Accelerated video available in the guest. That means that today, virtual video can be faster than running native for some video cards.

----------------------------------

http://www.innotek.de/media/VirtualBox_Brochure.pdf

... excerpt..
Virtualization - unlike ordinary applications - always operates at the lowest system level and therefore comes with risk of being able to compromise your system and data due to implementation errors and backdoors. Only a vendor that offers source codes and maintains an open development strategy can be considered to be fully trustworthy.This is where VirtualBox excels.

....
Better video support. While the virtual graphics card the VirtualBox emulates for any guest operating system provides all the basic features, the custom video drivers that are installed with the Guest Additions provide you with extra high and non-standard video modes as well as accelerated video performance.


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  ecsguy
Date  :  15 Jan, 2007 on 22:45
VirtualBox, Win4Lin, KbVM are all based on QEMU source code

XEN also uses QEMU code to run unmodified guests like Microsoft wwindows.

---------------------------------------------------------

http://www.virtualbox.org/wiki/News

* Jan 15, 2007. InnoTek today released VirtualBox Open Source Edition (OSE), marking an important milestone in the development of PC virtualization software. VirtualBox OSE is the first professional PC virtualization solution released as open source under the GNU General Public License (GPL). With VirtualBox, customers get the most versatile virtualization product on the market, both for enterprise and individual use. VirtualBox' open source license allows everyone to contribute to the development of the product and customize it to suit individual needs. Backed by Europe's largest team of virtualization experts, VirtualBox continues to be developed and supported.


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  Terry tgindy@yahoo.com
Date  :  16 Jan, 2007 on 00:00
http://www.virtualbox.org/wiki/Status%3A%20Guest%20OSes
says this about VirtualBox guest status under OS/2 Warp:

Works: Yes.

Stability: Good.

Performance: Very Good.

Remarks: Requires VT-x hardware virtualization support. No Guest Additions available yet.

QUESTION: What does "VT-x" mean?


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  magog
Date  :  16 Jan, 2007 on 02:02

Terry (16 Jan, 2007 00:00):
Remarks: Requires VT-x hardware virtualization support. No Guest Additions available yet.

QUESTION: What does "VT-x" mean?


The latest processors feature a 'virtualization' enhancement.
Intel calls it Vanderpool. The following processors should support it: Pentium 4 (6x2), Pentium D (9x0), Xeon (3xxx/5xxx/7xxx), Core Duo and Core 2 Duo processors.

AMD calls it Pacifica and it's available with the latest CPUs (all socket AM2, F and S1 processors). The old Athlon 64 CPUs for socket 939 and 754 or the old Athlon XP for socket A don't have this feature.


It think VT-x is only a reference to Intels virtualization and may not include AMDs implementation.


You may also want to read this:
http://en.wikipedia.org/wiki/X86_virtualization

---
Regards,
Juergen Ulbts (Germany)

*** Java Movie Database - http://www.jmdb.de/ ***
*** New version available (2006-11-17) ***
*** Website has been updated ***


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  ecsguy
Date  :  16 Jan, 2007 on 13:54

magog (16 Jan, 2007 02:02):

It think VT-x is only a reference to Intels virtualization and may not include AMDs implementation.


AMD code name was Pacifica. Now it is called SVM

The VirtualBox change log shows that Innotek added AMD SVM support in the latest release.

----------
Chapter 10. ChangeLog
10.1. Version 1.3.2 (2007-01-15)
VMM: experimental support for AMD SVM hardware virtualization extensions


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  ecsguy
Date  :  16 Jan, 2007 on 14:02
Looks like Parallels has fallen behind Innotek. Parallels does not expect to have video acceleration for a couple more months.

-----------------------------------------------------------------

http://arstechnica.com/journals/apple.ars/2007/1/11/6582

Ars at Macworld: Interview with Parallels Ben Rudolph

January 11, 2007 @ 2:26PM - posted by Jacqui Cheng

... excerpt ..
when might we expect the next major beta (key word here being "major") and he said that it should be within the next couple of months. There will be a few incremental releases in between (here was even one released today! Build 3120), but the next major beta should bring us a lot of cool stuff that many users have been asking for.

Such as?

Much to many of your delight, hardware graphics acceleration is expected to come in the next major beta, which as I said in the paragraph above, will be available within the "next couple of months.


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  BigWarpGuy
Date  :  17 Jan, 2007 on 18:07

Ben Dragon (17 Nov, 2006 01:27):
Well, I guess everyone has heard the latest news by now. In case you haven't Scitech has ceased development of their SNAP video drivers and is selling the code... all of it... and moving onto other things.

What does this mean for eCS/OS2 users?

Well, I foresee several possibilities:


    1. Someone buys the code and
    a) drops eCS/OS2 support
    b) or continues eCS/OS2 support and either enhances it or not...

    2. No one buys the code and we're left out in the cold or... the least likely...

    3. No one buys the code and they release it as open source.



I'm hoping for the 1b option myself, or, as a distant 2nd, the last option.

But what will we, as a community, do for upcoming video hardware support if we get left out in the cold WRT video drivers?

I don't think anyone would like to go back to the situation that we had in the past prior to Scitech and SNAP, (or SDD).

There is a major new MB setup coming that, supposedly, won't require drivers as we now know it, but that hasn't happened yet and no one knows what's actually going to be in the final production version. But if it does happen as billed it will be OK for us... if and when that happens, but what will be do in the meantime?

WRT to the short term SNAP will work with what's out there, but for the long haul...

Well... I guess we could pass the hat... :\


Perhaps a 'bounty' to help raise funds? Then one could start a non-profit group to develope it? (just thinking out loud).

It would be nice if Serenity Systems could buy the code; at least the OS/2 part.

Since Mensys has something for all os's, perhaps they could buy it and develope it for all os's?

---
BigWarpGuy
- - - - - - - - - - - - - - - - - -
OS/2-eCS.org
Director of Communications
- - - - - - - - - - - - - - - - -
supporting the past OS/2 user and the future eCS user
http://www.os2ecs.org


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  zircon
Date  :  31 Jan, 2007 on 04:25
Projects like SNAP may be difficult to sustain as open source should manufacturers comply with Microsoft's requirements to enforce copy prevention on Vista.

There's an interesting article on this at
http://www.cs.auckland.ac.nz/~pgut001/pubs/vista_cost.html

Vista will apparently prevent copying of multimedia files by downgrading picture and sound quality if the hardware is not "authorized" by Microsoft. "Authorization" is determined via a database of "fingerprints" generated through a hardware functionality scan. Vista will adulterate the signal stream before it is fed to any "unauthorized" hardware.

This scheme could be circumvented by emulating the hardware in software and recording the unadulterated signal to file instead of just playing it back. To prevent creation of such emulation software the hardware specifications must be kept secret.

I doubt whether this policy will be embraced by all manufacturers in the long run, but it may make it more difficult for open source programmers to get the hardware specifications needed to write new drivers at least in the near future. Reverse engineering is possible, but programming is much easier if the specifications are known...


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  BigWarpGuy
Date  :  31 Jan, 2007 on 19:26
I hope they see this as a monopolistic move on MS part and do something about it. It seems they are using their dominance of the OS market to push every one out.

---
BigWarpGuy
- - - - - - - - - - - - - - - - - -
OS/2-eCS.org
Director of Communications
- - - - - - - - - - - - - - - - -
supporting the past OS/2 user and the future eCS user
http://www.os2ecs.org

Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  obiwan
Date  :  01 Feb, 2007 on 01:30
The more Microsoft tries to tighten its grip, the more users and will slip through its fingers....

Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  abwillis abwillis1@gmail.com
Date  :  01 Feb, 2007 on 04:21
There is one way I can think of to make VM workable as a primary platform (even if not desirable). Take something like freebsd and set it up to only bring up VM. Then you could launch OS/2, maybe windows, and even freebsd or linux within that structure. It would be done in such a way that there is no base OS that would be usable as such... just the guests, similar to VM on the mainframes.

Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  BigWarpGuy
Date  :  01 Feb, 2007 on 14:24

obiwan (01 Feb, 2007 01:30):
The more Microsoft tries to tighten its grip, the more users and will slip through its fingers....

That sounds like a quote from Princes Leia to Darth Vader (or the evil head guy of the Death Star; I don't remember his name).

---
BigWarpGuy
- - - - - - - - - - - - - - - - - -
OS/2-eCS.org
Director of Communications
- - - - - - - - - - - - - - - - -
supporting the past OS/2 user and the future eCS user
http://www.os2ecs.org


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  ecsguy
Date  :  02 Feb, 2007 on 09:11

abwillis (01 Feb, 2007 04:21):
There is one way I can think of to make VM workable as a primary platform (even if not desirable). Take something like freebsd and set it up to only bring up VM. Then you could launch OS/2, maybe windows, and even freebsd or linux within that structure. It would be done in such a way that there is no base OS that would be usable as such... just the guests, similar to VM on the mainframes.

Ding, Ding, Ding. We hava a winner!

Another example is "microcode". It is a software layer underneath the OS (between OS and hardware). It insulated the OS from much of the actual hardware such that you could change hardware architecure and not have to recode the OS.

PCs just repeat mainframe designs 20 years late (but at an affordable price).

VMed eComStation

Make it so.

QEMU - http://fabrice.bellard.free.fr/qemu/
KbVM - http://kvm.qumranet.com/kvmwiki
VirtualBox - http://www.virtualbox.org/


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  Sebadoh sys3175@optonline.net
Date  :  02 Feb, 2007 on 17:47
Again.. the problem with this all is CPU cycles, anytime you put another layer in you slow things down. And again, having an emulation layer which dumbs thins down to a level which an older OS can understand (without updated drivers) is just the same thing as going out to a store and buying an older device. There will be no benefit from it. I can see advantages in certain points if done right:

The Host OS installs printing support and a virtualizaed network, you print through the guest OS using a standard printer interface. (this is pointless is we have cups)

The Host OS connects to the physical networking device and allows you to connect through a virtualized driver. (this is actually GOOD =) however it kills the ability to truely use the guest as a server, since the overhead is unacceptable under such conditions).

The Host OS controls the HD's, and virtualizes hardware access so that an older guest OS can make use of such hardware. However, there is a likely hood of corruption under heavy use. If the file system driver doesn't have a hardware level hook there is a chance (pretty good one with certain journalling file systems.. ::cough: that things can go quite wrong in a quite indetectable fashion.

I'm not going to even get into graphics hardware, due to the many layers involved even with native drivers you are just going to stack up overhead on the graphics hardware and the video hardware.

Hey.. I never claim that virtualization doesn't have it's uses. But again.. the real world useage is hard to determine, along with the fact that many people don't want to pay the price of entry. Could virtualization help people who already own eCS or similar operating system and wish to keep using it on newer hardware. Yes.

Will it help a company like SSI stay in buisness and gain NEW customers. Unlikely.


ecsguy (02 Feb, 2007 09:23):

abwillis (01 Feb, 2007 04:21):
There is one way I can think of to make VM workable as a primary platform (even if not desirable). Take something like freebsd and set it up to only bring up VM. Then you could launch OS/2, maybe windows, and even freebsd or linux within that structure. It would be done in such a way that there is no base OS that would be usable as such... just the guests, similar to VM on the mainframes.

Ding, Ding, Ding. We hava a winner!

Another example is "microcode". It is a software layer underneath the OS (between OS and hardware). It insulated the OS from much of the actual hardware such that you could change hardware architecure and not have to recode the OS.

PCs just repeat mainframe designs 20 years late (but at an affordable price).

VMed eComStation

Make it so.

QEMU - http://fabrice.bellard.free.fr/qemu/
KbVM - http://kvm.qumranet.com/kvmwiki
VirtualBox - http://www.virtualbox.org/[/quote]


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  Sebadoh sys3175@optonline.net
Date  :  02 Feb, 2007 on 17:54
Have you TRIED virtualbox's "accelerate video" I have. The 3 OS's I tried on it (one actually being winxp) actually have faster video in VPC/2 then they did in virtualbox on XP. (and the winXP machine is a AM2 4200+ (x2) and the OS/2 machine is a Athlon64 754 3000+ =)



ecsguy (16 Jan, 2007 14:02):
Looks like Parallels has fallen behind Innotek. Parallels does not expect to have video acceleration for a couple more months.

-----------------------------------------------------------------

http://arstechnica.com/journals/apple.ars/2007/1/11/6582

Ars at Macworld: Interview with Parallels Ben Rudolph

January 11, 2007 @ 2:26PM - posted by Jacqui Cheng

... excerpt ..
when might we expect the next major beta (key word here being "major") and he said that it should be within the next couple of months. There will be a few incremental releases in between (here was even one released today! Build 3120), but the next major beta should bring us a lot of cool stuff that many users have been asking for.

Such as?

Much to many of your delight, hardware graphics acceleration is expected to come in the next major beta, which as I said in the paragraph above, will be available within the "next couple of months.



Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  obiwan
Date  :  18 Apr, 2007 on 00:08
"Told you so" moment for obiwan:


obiwan (28 Nov, 2006 02:43):

Some folks have been working independently on their own OS/2 video drivers. SciTech has worked so well that these efforts have been largely isolated and unnoticed. If SNAP does decline, the good work of independent driver developers will attract recognition, support, and hopefully improved collaboration. This is good because SciTech was never going to expand the feature set of SNAP on OS/2.

We should be past using the "D" word by now.


See Alternative video drivers for eComStation posted last week.

Sometimes I even impress myself.


Subject  :  Re:Scitech, SNAP and OS/2...
Author  :  BigWarpGuy
Date  :  18 Apr, 2007 on 00:22
If it expands to support other chipsets, I could see it as a possible replacement for SDD. The future of OS/2-eCS is looking better.

---
BigWarpGuy
- - - - - - - - - - - - - - - - - -
OS/2-eCS.org
Director of Communications
- - - - - - - - - - - - - - - - -
supporting the past OS/2 user and the future eCS user
http://www.os2ecs.org

Powered by UltraBoard 2000 <www.ub2k.com>