ecs 2.2

[R.M. Klippstein]

Job well done Sigurd! Thanks

klipp

[jdeb]

That is encouraging. Thanks Sigurd.

[Ben]

I did not say Sigurd, that you have no valid business concerns regarding, or not regarding Mensys.

I was addressing the responses to this thread in general.

I did say, however, that there is no reason to believe that Mensys has abandoned eCS... and if there is such a reason, it has not been posted in this thread. What there has been posted, is merely the manifestations of insecurities that have rippled into rumours; they never go anywhere good.

But, more to the point, with the dozens, (if not hundreds), of wide-spread rumours that there has been about the demise of OS/2 over the years, the very last thing that Mensys needs, is another death rumour to start... given that they have spent years, trying to quell the existing ones; this can only be problematic for us all.

It is good that you received and posted, your responses from Mensys, (someone who is not afraid to make a phone call and initiate real research), for you have ended it well... or rather I should say, it ends well here... assuming no more relevant replies follow.

At any rate, I am glad that your fears have been put to rest.

[warp]

Most users currently using eCS do so because they like it.  And we are talking about version 2.1.  Which means most things people need actually work.  Admittedly, there are problems with some modern hardware, but ...

So what is it current users really need? And how to attract new users and hence build a stronger financial platform?  My guesses include:

• Net access, wired and preferably also wireless.
• Well working key software in an increasingly connected world = well working web browsers and email clients (which we have, but where the age of the current Gecko platform is starting to show).
• Working VPN/remote computer access (which we really do not have)
• With JAVA becoming increasingly important (despite some security issues), an updated JAVA (where work is on the way)

With this perspective it could be that scarce developer resources could be better spent not focusing on the operating system itself (i.e., getting eCS 2.2 up and running in a stable way).  In brief - how often do people use suspend and resume compared to accessing some web page with JAVA or Flash?

Or is my lack of computer knowledge revealed by these questions?  Without an updated (and up to date) platform, the tasks of maintaining and developing what most users really need become too large?  And in that case, the current Mensys focus on the platform is well placed.

Let me change gears and talk about something I know more about than computers: economics and strategy.  For worried  current eCS users and for future users, it would be nice if Mensys could share their current thinking with us.  That would make current and prospective users more informed. Some would decide to end their use of eCS.  Others would gradually start using eCS.  We like open source.  But even more, we like transparent strategies.  So Mensys, please end your silence and invite us on your journey.  Without us, I struggle to find a good future for eCS.

Eirik

We discussed briefly with an Italian developer and asked him which are the possible steps for a complete rewriting of OS/-eCS. I translated that ideas into English and I would like to share them on this newsgroup. I don't intend to start a war, but I would like to stimulate possible constructive ideas. Probably most of you already know such things.


"The problem is, besides the source code, the big number of components, which are not produced by IBM.

Even if we remove/replace such components, there are many problems.

The most important ones:

- recompiling and updating the kernel to 64 bit with readjustment of all data structures. Theoretically, that shouldn't be too complicated, but many kernel’s parts  have been coded in Assembler x86 and that could be a limitation.

The kernel has about 25 years from the 2.1 version and it didn’t get updated since 2006. In spite of that, it is competitive with the most recent kernels.

Besides that, there are components which are really old and didn’t use the features introduced by the CPUs from 2004 until now.

-The same thing for the whole graphic subsystem.

-A complete elimination of the parts written by Micro$oft and a possible replacement with open source components: DOSBox and Virtualization.

-Drivers: The big, complex chapter. In this area we would need a complete change of model. Perhaps we colud use the Linux or BSD driver model as start, but we would need to convert OS/2 for license reasons.

- PM and WPS: SOM is still very powerful and it could be reproduced. The task would be huge and many expert programmers would be required.
PM now has many limits if compared with the past, but its APIs remain a milestone (I am not joking: for completeness, coherence and quality, they beat the Win64 APIs 6-0 6-0 6-0.

- TCP/IP: Even in that area, there would be a need of many programmers for stack's coding. The stack itself derives completely from BSD.
The solution could be taking the BSD's stack, but changing MPTS with something else. In my opinion, the model itself is not bad, but it needs a complete update and an integration with an updated driver support.

- Many other points. We would a lot of programmers. Mensys doesn't have so many resources."

[Ben]

Some good info, Warp, thanks.

But most of this has been discussed in the past and most recently with the Voyager Project that turned into mist well, years ago now. 8(

Really, the only thing that remains alive, (save what Mensys does), is the OS/3 project which has enough life left into it to say that it is still alive... but nothing else.

[Martin Iturbide]

Warp, Ben.

It had been hard to do a full open source clone at one shot, by trying to clone kernel, Loader, GPI, PM, SOM, WPS at the same time. Thats why I had been discussing the idea of an open source OS/2 clone based on the idea of Voyager, but focused as first step only to replace the WPS classes with open source clones.

I really think that if we focus on an open source WPS classes replacement over eComStation and Warp 4.52 we can send a good message to the Internet about we are doing positive steps to have an open source clone of OS/2 and get more help.

I posted the comments on the history of my blog Open Warp.

About OSFree, I like a lot htat project because they have choosen a formal open source license. But sadly there are no resources helping on this project, plus, it is also too big for "one shot", it need to develop littler milestones showing components that can be replaced on the working OS/2.

About OS/4 (russian team). I also wish them good luck, but I what I don't like is that they do not use a formal open source license. It is freeware for individual use, and that blocks anybody that wants to do something deeper with that project. I don't know why the russian team working on that project thinks that open source helps big corporation to make money on they efforts.

I think that they do not understand "CopyLeft" (a concept inside the GNU GPL license) which forces everybody to give out the source code of any improvement they do inside the project. So, it does not matter if a corporation or an individual is doing money with open source as long as they give back the source code of the improvements. In other words it does not matter if you are big or small, the source code is the same for everyone and have the same "technical" conditions.


But what I need is to try to build an open source development team that wants to search only for one goal for the moment "A 100% replacement  of the eCS-Warp 4.5 WPS classes".
- Use an open source license (copyleft) to have the rules clear that the source code will belong to all of us (community).
- It does not matter if they only want to focus on one single class, or if they only have 20 minutes at month to help.
- The idea is to reuse the source code that we have from other projects (Xworkplace, HWman, XShadows, etc)
- The idea is that every little effort sum to have a WPS clone.
- Start with anything, it does not matter if it is WPClock or any other, but have an 100% open source replacement of it.

Regards

[warp]

Warp, Ben.

It had been hard to do a full open source clone at one shot, by trying to clone kernel, Loader, GPI, PM, SOM, WPS at the same time. Thats why I had been discussing the idea of an open source OS/2 clone based on the idea of Voyager, but focused as first step only to replace the WPS classes with open source clones.

I really think that if we focus on an open source WPS classes replacement over eComStation and Warp 4.52 we can send a good message to the Internet about we are doing positive steps to have an open source clone of OS/2 and get more help.

I posted the comments on the history of my blog Open Warp.

About OSFree, I like a lot htat project because they have choosen a formal open source license. But sadly there are no resources helping on this project, plus, it is also too big for "one shot", it need to develop littler milestones showing components that can be replaced on the working OS/2.

About OS/4 (russian team). I also wish them good luck, but I what I don't like is that they do not use a formal open source license. It is freeware for individual use, and that blocks anybody that wants to do something deeper with that project. I don't know why the russian team working on that project thinks that open source helps big corporation to make money on they efforts.

I think that they do not understand "CopyLeft" (a concept inside the GNU GPL license) which forces everybody to give out the source code of any improvement they do inside the project. So, it does not matter if a corporation or an individual is doing money with open source as long as they give back the source code of the improvements. In other words it does not matter if you are big or small, the source code is the same for everyone and have the same "technical" conditions.


But what I need is to try to build an open source development team that wants to search only for one goal for the moment "A 100% replacement  of the eCS-Warp 4.5 WPS classes".
- Use an open source license (copyleft) to have the rules clear that the source code will belong to all of us (community).
- It does not matter if they only want to focus on one single class, or if they only have 20 minutes at month to help.
- The idea is to reuse the source code that we have from other projects (Xworkplace, HWman, XShadows, etc)
- The idea is that every little effort sum to have a WPS clone.
- Start with anything, it does not matter if it is WPClock or any other, but have an 100% open source replacement of it.

Regards

The Voyager Project seemed to be interesting, but perhaps it was another example of ineffective communication.
I agree that clonig the whole system is too much. Replacing the WPS classes could be a good first step of a successful path!

[Andi B.]

The Voyager Project seemed to be interesting, but perhaps it was another example of ineffective communication.

I doubt communication was the main problem but to lees volunteers to do the work.

[Martin Iturbide]

I think that the two things are important:

1) Good communication about the progress, as little or much that is happening, and communication of what is missing and required.
2) Just focus on one thing at the time (like starting with WPS Classes). Sometimes we try to cover too much on the first step and people gets overwhelmed by the huge monster that is cloning OS/2.


...and yes, we are missing developers. But we have to keep trying.

[Doug Bissett]

I am not so sure that the WPS is the place to start. The WPS is still working well. Other parts are beginning to fall apart, and if they don't work, the WPS isn't going to be very useful.

The immediate parts, that come to mind (in no particular order) are:
UEFI. This will be a killer, very soon.
Video driver (SNAP has been obtained, and Panorama is working pretty well, but acceleration is missing).
IPv6 (not yet required, but...).
WiFi drivers. (planned, but not yet done).
WLAN drivers.
Bluetooth support.
Firewire support (this was started years ago, but was never finished).
USB 3.0 support.
Support for disks greater than 2 TB.
Take some of the load off of Paul Smedley. He is overloaded by keeping the development tools up to date, he doesn't need to be the one who updates things like CUPS, HPLIP, HPAIO, and a dozen other things that have been sorted out, but are essential to keep eCS viable.

I am sure there are a few hundred other things that need to be updated/fixed, before the WPS comes to the top of the list. Of course, there are developers who know various parts of the system better than other parts, so there is really no fixed order to doing the updates. It doesn't make much sense to have someone, who is a driver specialist, trying to update the WPS, or the other way around. I would note, that most of my list would be drivers. It also doesn't make much sense for 2 driver specialists to be working on the same driver, independently.

Another thing, is that any updates/fixes, need to be tested. Those who can't do the development, can do testing. They can also contribute to those who do the development, to keep them interested.

[Martin Iturbide]

Hi Doug.

I'm focused on an Open Source OS/2 because we need to grow the community and keep the continuity of the platform. I think that one of the stoppers to make an OS/2 clone is that OS/2 is still working fine today, so nobody wants to touch the things that keep working.

I share that we need all that, but if tomorrow a good wifi, bluetooth, IPV6 drivers/programs shows up, that will not help this community to grow, it will only give us air until the next gen of devices shows up.

That why we need a long term strategy, we need to make OS/2 accessible to everyone and ensure its life beyond IBM,  Mensys and us as individuals. That is the importance of making an OS/2 clone. It is not only to support hardware today, it is to maintain the platform in the future. The one path that I can think about is to make an open source clone.

Thanks Doug for offering on testing, I'm not a developer, so I'm also signing in for testing, icons, documentation and whatever I can do. But I need to try convince developers to look into WPS and try to generate more development in this side.

Regards

[Eugene Gorbunoff]

1) Does Voyager project relate to OS/2? I don't this so. As I remember, it was about copying WPS desktop to other platforms.
That's why OS/2 people didn't supported it.

2) Currently people buy eCS licenses. People don't pay for support service to service providers.
In this model, there is no need to open source OS/2 / eCS, All development goes on eCS money. There is no other interested commercial companies.

It will be automatically open-sourced if several companies create solutions based on OS/2, attract many customers, pay for all bills, pay for all development. In this model, this companies will open source own code to cooperate.

But this doesn't happen. So, eCS will stay semi-commercial, semi-open-sourced.

 

[Martin Iturbide]

Sorry Eugene, but we have different approachs.

You sell eComStation software licenses and that is the only business model you know. Thats why you don't understand that open sourcing (creating a OS/2 clone) will be a better for all as a long term strategy.

Please notice that right now the "business" are big corporate customers that buys eComStation since they can not quit OS/2 for the moment. Once this big customers migrate to newer platforms that will be no more business. sorry.

The rest that you wrote I don't understand it. If more people buys eComStation and develop more software for it it will not get "automatically" open sourced. Any close source software that gets generated today for this plataform is just "future abandonware". It seems that you have some problems on the translation.

Regards

[Eugene Gorbunoff]

I am not against open-source.

eComStation includes many open-sourced components already.

My message: eCS will stay "semi-open sourced" (partly open sourced), and partly closed.

It's impossible create open source OS/2. As I said, it can be created by several corporations but we don't have this businessmen.
To maintain open source OS you need millions dollars every year. To maintain eCS you need XX thousand dollars only.
There is no funds (capital) on eCS market, that's why eCS is partially closed.

[Martin Iturbide]

I disagreed.

While there are people telling us that it can not be done, that it is impossible, that it is better to lay down our arms and wait for the destiny to take over, there are other projects that shows me that it can be done:

- Linux is a Unix clone.
- Haiku OS a BeOS clone.
- ReactOS is a Windows clone.

All those projects are steps ahead of us cloning the platforms that they like. I don't know why people think this is a project to be completed in two months. The reality is that we need to start now if we want to see some progress in some 5 years.  We need to create traction, we need to attract more developers.

Let the people that have other interests to keep saying whatever they want, but I can not hear anymore to the people that says that it is impossible, they are already defeated.