• Welcome to OS2World OLD-STATIC-BACKUP Forum.
 

News:

This is an old OS2World backup forum for reference only. IT IS READ ONLY!!!

If you need help with OS/2 - eComStation visit http://www.os2world.com/forum

Main Menu

Ecomstation 2.0 GA

Started by PAUL555, 2009.04.21, 19:02:06

Previous topic - Next topic

PAUL555


When is Ecomstation 2.0 GA going to be released ? What is the reason for the delay of over 4 years ( Aug 2004 to April 2009 ). Why cant they release additional changes as service pack-1,2,3 e.t.c similar to Windows ? The development cost itself must be going up to a large amount. 

Paul   

warpcafe

Hi Paul,

while I can understand the question and also the way it was asked (we all feel like this... for years now ;) ) I think you have to keep in mind that SPs will be delivered to solve issues with an otherwise "finished" product.

With eCS 2.0, the intended maturity baseline is to make it run on more than only the majority of contemporary PCs. See, wehen you buy Windows or get Linux or MacOS(X), the things that work are Internet, Display, PowerOn and PowerOff, your Mouse... and everyone expects that, right?
Wit eCS it can happen that due to ACPI, one or more of these things will not work on a wide range of systems. So, the possible end-user that has a first-time exposure to eCS will simply kick the CD into the bin and never touch it again. Also, a SP usually is delivered as delta to an installed operating system. It is not the standard way that an OS is re-released with an included/integrated SP on the media.

So, for eCS 2.0 we don't have a stable GA with perhaps "additional changes" on top of it later... we have a release that is not as "generally suited" as it would be required. That *might* be two reasons for it.

HTH, Cheers,
Thomas
"It is not worth an intelligent man's time to be in the majority.
By definition, there are already enough people to do that"
- G.H. Hardy

PAUL555


I understand that the developers have to fix the bugs and add new functionalities. However , is it possible to create a perfect operating system free of all bugs ? Also new hardware is coming out all the time. Playing 'catch up' for new hardware would be a never ending development cycle. 4.5 years and still not released is really too long for an operating system development life cycle.I am sure that once 2.0 GA is released there is going to be a long list of hardware on which it is not compatible or works below expectations.

Paul 

Blonde Guy

As I understand it, we'll see release candidate 7 with an ACPI implementation that is ready for GA. I don't think it's ready at this time, so that is still being prepared. Once RC 7 is out for testing, the GA will follow in a short time unless some pretty bad bug is reported.

If a bad bug is reported, then once it is resolved, we'll see release candidate 8. The process would repeat until the GA is released or the developers run out of resources.

Implementing ACPI is a huge undertaking, and I agree with not releasing the GA until it is good enough.
Expert Consulting for OS/2 and eComStation

Pete

Hi All

Having recently tried ACPI V3.15 build 2 (ACPIPRE2.ZIP) and found it to be a "non-booting build" until USB and Uniaud drivers are removed from the config.sys file I can only state that ACPI is definitely not ready. The mainboard chipset involved is nVidia nForce430/MCP61 - not exactly "leading edge" now.

Hopefully all the logs and screenshots that I prepared will help the ACPI developer(s) to resolve those problem(s) fairly soon - without causing similar problems on mainboards using other chipsets.

Guess we'll see RC7 when enough of the ACPI problems get solved for enough people.

Regards

Pete



DougB

ACPI is definitely a major problem. It is also supposed to be one of the major enhancements to eCS 2.0. Unfortunately, the development team seems to have chosen a bad approach, by trying to do it right, when nobody else even considers doing that. The results seem to be, that they need to tailor the program  for each, and every, possible brand, model, and BIOS version, that somebody manages to create, to make it work. Throw in the fact, that there is probably a new one, every day, and many of them are not created according to specs,  and they have no hope of ever accomplishing what they set out to do. I also very much question if they can even fix some things. It has been over a year, since they were given the challenge of making my Acer Aspire 4720 laptop work. So far, there is no difference, at all. GENMAC still tries to use IRQ0, when I try to use ACPI (GENMAC works when ACPI is not used). Too bad, because the rest of it seems to work fine, except for the stuff that needs a driver, that doesn't exist.

ivan

As I see it, standing on the sideline looking in, there appears to be no co-ordination of the effort to get a GA out!

The base OS is OS/2 - which works.
Daniela has solved most of the drive problems.
Most common network cards have drivers.
Paul has given us a uniaud driver for the most common audio chipsets.

What is left that is important?  ACPI and WiFi drivers.

So there are two areas to be worked on, ACPI and WiFi, to get a full working OS out of the door.  Where is the hold up?  Who is co-ordinating the effort? If no one then why not?

For both ACPI and WiFi there is a very reasonable amount of material available in the linux community, can much of this be applied to OS/2 like the uniaud project and others that Paul and others have ported for us?  Or has the development of these essential parts degenerated to an ego pumping exercise?  If this is the case then new developers should be found that will work in the project and see it to a successful conclusion.

Until the problems with ACPI and WiFi are solved all other tweeks should be relegated to secondary status to allow a working GA out to customers. 

As it stands OS/2, through the medium of eCS and its long delay, is being relegated to an also-ran hobby OS - which I resent.
Since way back when OS/2 v2.1 arrived I have encouraged others to use OS/2.  At the moment I am prevented from doing so because of the very limited availability of copies on places like eBay.

If Daniela, Paul and a few others can produce a wealth of drivers and programs in the time it has taken to get a part working ACPI layer ready - I must ask what is the problem? and will there ever be a GA of eCS?

The project co-ordinator need to come forward and give a truthful statement about what is going on before ALL credibility is lost, as I lost out on getting OS/2 into a small organisation today because I didn't have enough spare copies of OS/2 available.


Blonde Guy

I'm not waiting for the GA. I'm selling eCS-preloaded computers based on the release candidate we have now. I'm selling eCS licenses and support agreements to those who want to build their own computers.

I've posted this before, but here is the url: http://www.blondeguy.com/ComputersForSale.html

I'd like the work to proceed more quickly, but I understand how difficult an undertaking ACPI is. In the mean time, ACPI, uniaud and GenMac do work for some configurations.

I'm sorry about reposting, but I want to make clear that eCS is not in as much trouble as some of these posts make it appear. I'd like eCS 2.0 GA to be out. It would simplify the ordering process greatly.
Expert Consulting for OS/2 and eComStation

cytan

Just out of curiosity, does 2.0 work with Intel I7 quad core? I plan to build a machine around this CPU at the end of the year.

cytan

ivan

Cytan, the answer is maybe.  I imagine the OS/2 part will but the big question is will the addon ACPI etc.

Blond Guy, what you are saying is that eCS works - IF you select the hardware carefully.  That is OK if you are buying new but, and it is a big but, what happens when the client wants to use what they have already?  You don't sell corporate customers pie-in the-sky.

As Doug pointed out the approach to the development of the ACPI layer is wrong - making it work for individual chip sets will work - IF you have all the time in the world!  ACPI in linux works for most machines out in the world - if they can do it why not OS/2, or the case in question eCS?

Much as I would like to see eCS go forward I'm not holding my breath - 4 years plus to get a part working ACPI layer is a bit much.

The Blue Warper

Hi all!

Just a short note.  While it's true that we all would like eCS 2.0 to come GA soon, and that the ACPI code building approach is maybe questionable (i.e. the closed source strategy. But one should ask: how many OS/2 system-level code developers are left who are capable of actual contributing to the project?).  I mostly agree with the view expressed here by many of you.  But I also think we shouldn't forget (what I believe is) the main reason why the OS/2 ACPI-subsystem developers are going so slow (in relation to users' expectations, at least).  And the reason is kernel.  ACPI is a complex subsystem which has to low-level interface with the kernel.  There are thing like IRQs, for example, where an open-source kernel would have both greatly improved developing (and releasing) speed and system stability, and allowed eCS to run on a greater number of hardware configs.
So I understand the complaints, but I think we should look with fairer eyes to this delay.
As to Linux code being available, this is true, but is this stuff useful for OS/2-eCS ACPI subsystem, if we haven't access to kernel source?  Part of it would probably be (especially that part designed to handle ACPI specs), but how would one interface that to the kernel?  That's why (I think) it's taking all this time to release eCS 2.0 GA.  But, having said that, I agree that a different approach (not tuned on an almost per-single-hardware-config basis, so to say), i.e. a more general, hardware-independent approach, would have been better.

We could perhaps suggest to the developers something like this (I'm not a developer, so please don't blame me if I were wrong...).  What about splitting ACPI.PSD into two parts: one which would just contain the ACPI-layer; the other the hardware specific code?  The former would interface directly with the kernel, and would contain the basic ACPI layer, acting as a sort of ring-0 Hardware-Abstraction-Layer (HAL).  The latter would interface ACPI-HAL with the actual hardware.
Perhaps this way the developers could more easily manage and maintain the code.  Whenever ACPI specs are updated, they could accordingly update their build.  The real-hardware counterpart (which I imagine in a device driver or PSD form) would interface with the ACPI-layer by performing some ACPI-calls and passing to the ACPI-subsystem their own hardware-tuned, hardwares-specific parameters.
Thus every single system would execute his own code (basic-ACPI + hardware-specific 'driver') and conflicts caused by execution of code written to accommodate different hardware would hopefully be reduced.
Just some thoughts, though.

Regards

rwklein

Quote from: ivan on 2009.04.23, 23:42:20
As I see it, standing on the sideline looking in, there appears to be no co-ordination of the effort to get a GA out!

The base OS is OS/2 - which works.
Daniela has solved most of the drive problems.
Most common network cards have drivers.
Paul has given us a uniaud driver for the most common audio chipsets.

What is left that is important?  ACPI and WiFi drivers.

So there are two areas to be worked on, ACPI and WiFi, to get a full working OS out of the door.  Where is the hold up?  Who is co-ordinating the effort? If no one then why not?

For both ACPI and WiFi there is a very reasonable amount of material available in the linux community, can much of this be applied to OS/2 like the uniaud project and others that Paul and others have ported for us?  Or has the development of these essential parts degenerated to an ego pumping exercise?  If this is the case then new developers should be found that will work in the project and see it to a successful conclusion.

Until the problems with ACPI and WiFi are solved all other tweeks should be relegated to secondary status to allow a working GA out to customers. 

As it stands OS/2, through the medium of eCS and its long delay, is being relegated to an also-ran hobby OS - which I resent.
Since way back when OS/2 v2.1 arrived I have encouraged others to use OS/2.  At the moment I am prevented from doing so because of the very limited availability of copies on places like eBay.

If Daniela, Paul and a few others can produce a wealth of drivers and programs in the time it has taken to get a part working ACPI layer ready - I must ask what is the problem? and will there ever be a GA of eCS?

The project co-ordinator need to come forward and give a truthful statement about what is going on before ALL credibility is lost, as I lost out on getting OS/2 into a small organisation today because I didn't have enough spare copies of OS/2 available.




I have stepped forward multiple times in forums to update. Our communication (that is Mensys) leaves to desire for.  But we are with a few people doing maybe to much work. Next to all the work that goes into eComStation we stick out our neck to get events rolling. For example I myself organized Warpweekend once. I'm the president of OS/2 VOICE (BTW any volunteers that can help out ?).


On the communication topic ? I from time to time have time to reply in forums here. But I don't have time to do so all the time. There are to many forums where people expect a reply.  We are looking at ways to improve it.


If you think the device driver situation is oke, well it is and its not.

Printer driver support is not oke. But I'm working on that with OS/2 VOICE, Paul Smedley and Neil to get CUPS to state to fix that.

Video card support, don't count much on SNAP updates we (Mensys) is in touch with the company that purchased the SNAP technology. Hence the reason we got Panorama driver developed by eCo software. There are one or 2 bugs. But overal it performs quite well. And the feature set we have SNAP, please note that a lot of the luxury features like dualhead and zoom support etc are all chipset specific. Right now the human resources are not realy around to update SNAP. Enhanced to Panorama are coming. But the amount of video chipsets that don't simply work with SNAP is rising with the day.

Right now we are coordinating a new Flash version (hope to get a public beta out of Flash 10 in about a month). But there are some technical challenges to fix.  I did the legal work and coordinate with somebody else the development.

It was also Mensys that invested close to $20.000 to higher a developer to do some serious debugging work UNIAUD.
It was to got HDA audio to work to a large extendand get out some of other bugs. Not forget there are other developers like Paul Smedley and a group of other developers (I forget there name, look at Netlabs). Who do UNIAUD development.
This was also done for Flash support.

Also Mensys coordinates (with one developer) the release of Open Office for OS/2.

Also the USB stack is running out gas. I will spare you the details but a project is on its way to get this updated as well.
For example enuration of devices on the USB bus does not work properly.

Then on ACPI development. ACPI is not that easy as it sounds. Infact on Linux its far from a nice dream. A friend of mine has written a kernel module for USB webcams. And Suspend resume with ACPI is also far from a dream on Linux.
Also Adrian Gschwend from Netlabs confirmed this. Its not that we should not put every effort in to make eComStation to work. But its clear ACPI does not work as specified on paper according to the specifications. 
Next to that the ACPI is pushing the kernel infrastructure to terrains never explored. Heck we need the APIC support (that is support above IRQ 15). A lot of device drivers are hard coded upto IRQ 15. Drivers like UNIAUD and the Daniela driver have received updates for this. Suspend resume is also proving difficult to make work.

Next to this eCS 2.0 when you support has a lot of patches and updates (like loading less snoopers drivers) and updated testcfg.sys to load on modern hardware.

A lot of R&D takes place to keep eCS working. Take a look at bug 2345(discs bigger then 512 GB) and bug 2354 (how to get eCS to instal next to the Vista).  You can view the bugs on www.ecomstation.com (log in). Its stuff that is all expected to simply work for perspective of the customer.

Mensys hopes to release the eComStation silver release for public testing before the 5th of May.
A large amount of improvement is being worked to make it run better with virtual machines like Virtual PC, Parallels Workstation, Virtual Machine and VMware contacted us to get official support for eComStation.

Its safe to eCS 2.0 is the biggest overhaul on OS/2 since Warp 4. And it is and was needed.
And I wish it would have not taken that much time.

The community is there, like Paul Smedley doing all these GCC ports and tools. And Peter Weilbacher working on Firefox.
It just shows you what you can for an operating system. I'm certain Paul does mind if I tell you that he works as a manager at a large company. Most of the coding he does he does as a hobby in spare time. It just goes to show that it does not have to eat up your time. And as Paul told. (Correct me if I'm wrong Paul) you had a course of C/C++ at the University or something like that. The point is that Paul is as far as I know not somebody specialized in IT. Just goes to show what can be achieved. And we have enough people in the community left to get to work and contribute somehow.

Mensys and Serenity Systems can work on OS/2. But I'm certain more people will support the statement that Mensys in that sense does maybe a bit to much to keep the OS going (Thomas Klein, Paul Smedley). We have a commercial interest and its also advocacy. It was not the first time that Warpstock Europe got started because I coordinated the communication between  the different parties

The whole point I have been trying to make is that we are trying to keep eComStation current and we are pushing with all man power we have.  OS/2 can be made to run on more moder hardware but it needed a lot of catching up to do. IBM has done updates to it. But more engering works on the track are needed.

eComStation is the Mensys its job. But there is plenty of volenteer work that is done by a very very small group of people. And we coule certainly use some more hands to lighten the load.

The message I'm trying to get accros is that its realistic to be hopefull to keep OS/2 running, that is eComStation. As we have plans for plans for beyond eComStation 2.0. But its a lot of extreme hard work to get the job done.


We also released a Japanese test version of working on the two Chinese versions.
And yes there is still company demand (from everywhere all over the world) to keep OS/2 running.
But again all the work added togheter is a lot of work. And this message only scracthes the surface of what is done.


Roderick Klein
Mensys

ivan

Roderick, I thank you for your report.  It does show that work is being done.

I also get a sense that you are frustrated because you are wearing too many hats at this time.  Maybe it is time for you to ask the wider community for help - I also know that would most probably result in a total lack of response, but you might get nuggets of insight like that of The Blue Warper in the post above yours.  How much of that would work only the project co-ordinator could say, here I speak from my days of project co-ordinator on various government construction projects round the world.

Regarding the need to keep people informed of progress, maybe a short monthly newsletter sent off to newsgroups and other places where there is an interest.  It also gives you the chance to 'blow the trumpet' at every advance.

Again from my perspective, I can see a few possible ways to improve productivity but I will not put them here as they need more information than I have at the moment.

If you think a part retired engineer - who is finding himself with more work than when he wasn't part retired - can possibly be of help e-mail me (ivanjt (at) free.fr).

And again I thank you for the report as I now know not to give up hope.

ivan

rwklein

Quote from: ivan on 2009.04.25, 20:05:07
Roderick, I thank you for your report.  It does show that work is being done.

I also get a sense that you are frustrated because you are wearing too many hats at this time.  Maybe it is time for you to ask the wider community for help - I also know that would most probably result in a total lack of response, but you might get nuggets of insight like that of The Blue Warper in the post above yours.  How much of that would work only the project co-ordinator could say, here I speak from my days of project co-ordinator on various government construction projects round the world.

Regarding the need to keep people informed of progress, maybe a short monthly newsletter sent off to newsgroups and other places where there is an interest.  It also gives you the chance to 'blow the trumpet' at every advance.

Again from my perspective, I can see a few possible ways to improve productivity but I will not put them here as they need more information than I have at the moment.

If you think a part retired engineer - who is finding himself with more work than when he wasn't part retired - can possibly be of help e-mail me (ivanjt (at) free.fr).

And again I thank you for the report as I now know not to give up hope.

ivan


The thing of an update per month. I get to hear from people its all done in just 5 minutes just like all jobs. I don't make a fuss about it. Its not really frustration. When it comes to eComStation work I have to complain. Its my job at Mensys to work on it.
However the work done on all kinds of other things eats up time. I have asked over the last couple of years other people to help. But ended up loosing more time asking then getting the job done myself. So after about 8 years of communtiy effort you kind of get to enough to some degree.

Some things for OS2 VOICE are also still on the backburner and we could use help with OS/2 VOICE. I did ask around in a circle of friends. But we did send emails to the current VOICE members asking for members to join the board. Never did a reply come back.

Thats not a problem but some OS/2 users who are vocal in newsgroups and mailing lists are good at complaining about lack of progress on certain things. When it comes to eComStation etc they have a right to complain since its a product Mensys sales. But then to organize something like Warpstock or Warpweekend and to then get complaints and with other things.

Its so simple to complain. To say it could be done better but in few to no cases to see something come from the ground.
People have a right to there opinion but some people complained that Warpstock this year is organized in Stralsund Germany.
Well we (that is Christian Hennecke) send out emails asking for more offers, but nobody came forward.
Then my state is of mind is, talk is cheap.  Somebody said "he Stralsund is not in central position". Yeh from which person his perspective ? People from Scandinavian countries and Poland might come faster now!

BTW the website for Warpstock 2009 will go live soon and will be annouched soon, 1 to 2 weeks.

Roderick


Paul Smedley

Quote from: ivan on 2009.04.23, 23:42:20
Paul has given us a uniaud driver for the most common audio chipsets.

Just to clarify here - my only real involvement in Uniaud has been to update the ALSA level to support new chipsets, in a coordinated manner to make future updates easier.

Uniaud was created by Innotek/IBM and thankfully the source was released, so that Vlad_est and netlabs could continue to maintain it.

Cheers,

Paul.