OS/2, eCS & ArcaOS - Technical > Programming

An OS/2 Workplace Shell Clone - A Project?- Suggestions? - Developers?

(1/2) > >>

Martin Iturbide:
You all know that my final dream will be to have a formal open source OS/2 clone. I also guess that the majority of the OS/2-eCS community share the same dream, even that can only be called a dream.  But when I see projects like Haiku OS and ReactOS I see also that it is not impossible.

Since I want to move a little step closer to that dream I started a blog to share some ideas about it. And my first steps was trying to document at a high level the OS/2 architecture and try to find some points of where should we start.

I proposed that it will be interesting to start an OS/2 clone from "Top to down" which means cloning WPS, SOM and PM. And after analyzing this three components I also proposed that will be interesting to start cloning WPS classes on top of IBM's SOM and IBM's PM. why?

* WPS classes can be called as the core of WPS
* WPS classes are an important way on how we interact with OS/2-eCS
* eComStation can benefit from this project and improve it functionality, while also contributing to the open source project
* Cloning WPS will be a strong political statement outside the OS/2 community that we are moving forward on an OS/2 open source clone.
* We will move one step forward on having a long term strategy for the Operating System.
I also wrote some pitfalls on the blog about starting with a WPS Clone:

* Developers... I don't have (yet) no one interested on the project. It is also hard to find developers with WPS and SOM skills.
* Money. It will be very hard to find someone for free.
* Any strange behavior (quirks) on IBM SOM will need to be supported until IBM's SOM can be cloned.
* After a working clone of  WorkplaceShell is created, it will require future re-works/tweaking when the SOM and PM clones gets created.

I want to see the possibility start a project to create an open source WPS classes full clone. I don't have development skills, I can't do it alone and I need developers, suggestions, funding and moral support :)

But I think that first I need to contact some developers with WPS/SOM skills to try to define a plan, costs, time frames and also evaluate the open source software that we currently have that can be re-used.

I'm open for suggestions.

Jan-Erik Lärka:
I like your idea!

How far from finished was SOM 3 (Beta)?
Add it to ...?
Possible to pay them to finish it?

Andi B.:
Martin, you know netlabs Voyager? Have you talked to Adrian and Chris about that?

Martin Iturbide:
Thanks for replying.

Hi Andi B.

Sure, the idea is not new, it is very inspired in Voyager [1]. I remember seeing the presentations long time ago. But trying to focus on the three components (WPS, SOM and PM) at the same time was too complicated. That's why the idea is just to focus on WPS as a start and try to complete at lease one of the three things.

Some time ago I talked to Chris about it. He told me that he prefers to work on SOM first and later in WPS, since SOM has some "twearks" (strange behaviors) that will force you code WPS in one specific way, and later if you work on a SOM clone, you will have to rework WPS. Other thing that was also an issue when cloning SOM was that the temptation to fix what it (IBM Version) has broken was too high, and fixing it may result on an incompatible SOM version. (Note: This is how I understood it) But Chris is no longer active in SOM development.

I will talk to Adrian about it on Netlabs IRC and see his point of view.

Hi Jan-Erik Lärka

I don't know about SOM 3. The only efforts I have seen with an open source SOM was somFree (http://sourceforge.net/p/somfree/) But it was not OS/2 focused, but I think that some stuff can came from it at its time.

NOM, which was netlab's SOM, I don't know it current status.  I can see the source code on the SVN - http://svn.netlabs.org/repos/voyager/nom/

Please do not take my word as final statements, like I told you before I'm not a expert on this subject, so I'm open to any suggestion or constructive criticism.

Jan-Erik Lärka:

I installed it once one machine and tried it out the other year (compiled some example code) and it did was it was supposed to do as far as I can remember.

Wikipedia: IBM System Object Model with links to archived IBM pages and the toolkit on

Hobbes: IBM SOMobjects Developer's Toolkit Version 3.0 for OS/2 (OS/2 only)

SOM is a huge project I've read, so if we could replace SOM 2.1 with a stable SOM 3.0 seem closer at hand then to start from scratch. But to move SOM 3.0 Beta to GA may provide a challenge?


[0] Message Index

[#] Next page

Go to full version