OSNews

You don't need to guess. I worked on this project. IBM is never going to be able to open source OS/2 unless Microsoft agrees. OS/2 contains complete details on SMB/CIFS something that Microsoft does not want public. Probability that Microsoft will agree to open sourcing OS/2 = zero.

Then what did the PFIF just get from Microsoft for 14,000 euros? This year's Betty Crocker cookbook?

Edited 2008-01-22 00:33 UTC

There is a big difference between a spec from Microsoft which will be chock full of bugs (maybe on purpose) and a working, tested implementation. Did they get the spec for SMB as well as CIFS, or CIFS only?

There are a bunch of other secret MS technologies in OS/2 too. Remember, this was a joint venture between Microsoft and IBM, not an IBM only project. I worked on OS/2 full time and was a Microsoft employee.

MS is not going to want to open the code for the HPFS file system. Has the disk format for HPFS ever been made public? MS wrote HPFS.

Well, a working and tested implementation would be Samba. Jeremy has already described how they would treat the spec. They'll be checking everything out with formal regression tests to make sure the details of the spec actually work with real clients, and then make any changes within those real world constraints.

I'm not disagreeing with you that MS would be unlikely to OK the release of a FOSS OS/2. I'm just quibbling over that particular reason.

There are a number of things that I can think of off the top of my head that would excite me more than an open source release of OS/2, anyway.

It was a very advanced OS... for its day.

You'd be surprised if you saw the source. Half of OS2 is written in x86 assembler. Probably around 2M lines of it. The x86 C compilers weren't very good back then and a decent programmer could easily beat the compiler. RAM was a precious commodity and a lot of work went into reducing the memory footprint. It took about two days to do a complete build. MS didn't even have a network connection to IBM, we mailed floppies to exchange code.

IBM doesn't want to Open Source OS/2 for several reasons. Microsoft not even being associated with those reasons. But, with their "commitment" to open source and their inability or desire to even work on this just makes me wonder how deep that commitment runs.

It would be prohibitively legally complex and expensive and the actual benefits to the open source community are debatable.

I didn't know you were a Microsoft employee at the time, but then again I was still in high school. I think this OS/2 thing should be put to rest as a legacy OS. Nice to run on your hardware or in a virtual machine if you can get it or need it, otherwise not really relevant.

I'm actually far more interested in your work on Xorg, particularly concerning Xegl. Will we ever see that completed or in some other way integrated into Xorg?

The group that controls X.org does not want to make X more 3D oriented. So it is pointless to continue working on XEGL when the core X developers won't support the concept. Apparently X is going to stay 2D only until every child in rural India and Uganda has a 3D card. It doesn't seem to matter what users in Europe and America want.

However, I've recently heard that NVidia may be backing a XEGL type system that would be developed through Khronos. No code yet that I am aware of, but the project is being talked about.

I'm currently working on a Linux based embedded audio project at www.digispeaker.com but it's not ready for primetime.

ROFL

AND (at least in its day) a complete, better-than-Microsoft's, implementation of Windows. Now talk about MS opposing the publication of THAT.

Uh, OS2 used a full version of Windows 3.1 for Windows apps. It wasn't a better Windows than Windows, but it was a better DOS than DOS.

Almost correct; OS/2 did (does) use a heavily modified Win3.1 kernel which was (is) a lot more efficient than the original. Other than that, it indeed comes with a full Win3.1 package

And maybe we can call it:

"Open Source/who"?

Get it? os/2...open source/who?

Aw, never mind!

What a terrible joke! I modded it up it was so bad. Here's another...

Why should you never iron a four leaf clover?

Because you shouldn't press your luck.

My wife laughed so hard at your joke that she almost passed her quiche through her nose! :-)

And maybe we can call it:

"Open Source/who"?

Get it? os/2...open source/who?

Aw, never mind! "
I suggest the logo should be a blue police box.

Maybe nobody wants to implement the Presentation Manager again because they're afraid it may hang up over and over. (sarcasm)

I remember the PM/2 used to hang but the machine kept working perfectly fine (happened on Servers too). It was annoying

I worked for IBM back then. We all had OS/2 2.11 then we moved to 3 (warp) and finaly 4 (Merlin). Later, internal machines started coming with NT 4.0 and those who still needed OS/2 (lots of internal APPS inside IBM were written with Borland C++ for OS/2, REXX, etc.), ran it through virtualization or opening a session on a server.

It was nice. I loved being level 3 tech support, I was in direct contact with some of the developers from time to time. Ahhh.. but I was 20 years old back then :0

This is precisely what I've been interested in seeing from IBM for years.

All of these petitions asking IBM to release the source are engaging in pie-in-the-sky wishful thinking, in my view, but IBM was willing to allow SSI to create a Live CD using OS/2 technology. That was more than I expected. Why not release something along very similar lines -- a stripped down binary-only desktop version with absolutely no support?

Vanilla OS/2 by itself isn't that thrilling anymore, except (as others have mentioned) for its WorkPlace Shell, but when the base OS/2 platform is combined with a selection of freely-available third party tools, it becomes a fairly good lightweight alternative to Linux and other UNIX workalikes and/or derivatives out there.

*Any* decently performing kernel can be made usable given a sufficient selection of drivers and end-user applications, and OS/2 is no exception to that.

This would depend on how the royalty agreement with Microsoft was written. At that time percentage based royalties were common. If OS/2 is on a percentage based royalty then it would be possible to distribute it for free. But there may have been a cause requiring a minimum per copy royalty to be paid to Microsoft. I don't remember how it was written. IBM would certainly have had the right to give away free demo copies, which was how they did the magazine cover release.

The magazine cover copy was the full working version. I installed it many times and registered it with IBM.

Putting it on a magazine cover was a one shot promotional activity. Did the license on the CD allow the CD to be copied? I doubt that it did.

Microsoft probably would not would view altering the license to allow free, unlimited redistribution as a promotional activity. They would demand a royalty buyout before allowing it.

I don't think it actually took them 2 years to respond. I think they were just referring to a letter they sent him in 2005.

I'm not sure about that assertion, but what I know is that in spain, BBVA Bank used to use OS/2 less than 3 years ago. And I haven't been in the bank lately (hate it), but last time I checked, they had newer boxes with Windows 2000 and a TN/3270 or similar connected to OS/2 Lan Manager or whatever... because of their "legacy" application ran under OS/2.

Replacing a Bank's IT infrastructure and backend is by no means a simple task.

If there are Airlines and other "big" companies using it, I highly doubt they'll open source it. It would be very irresponsible. There may be bugs and security holes that are best left covered under obscurity and lack of interest. Given that IBM is unlikely to release a new fixpack for those, IMO is best to leave OS/2 where it is, slowly fading into history and replaced by new stuff.

I've seen dozens of ATMs with OS/2, although now most of those have some sort of Windows.

Madrid Barajas' Airport uses Windows NT/2000 and/or Novell on some of the displays. The Check-In terminals are Windows now and so on...

It takes time, but eventually legacy things get replaced. (sarcasm-> tell that to AIX/Cobol)

A pitty that Microsoft won the race there. Those terminals could be perfecly using some variant OS. I guess they went for the support.

Edited 2008-01-22 10:17 UTC

I don't know where you get the idea from that AIX is some kind of legacy platform but for your information, it's nothing of the sort. There is not a single large banking facility here in Luxembourg that does not run AIX for at least it's DB infrastructure. Couple that with the fact that there is not a single large financial institute on this planet that does not have a branch in Luxembourg and you have got to be joking when your saying that AIX is legacy.

Not just the airline industry, there is a vast sway of atm machines that run OS/2.

They are the ones that will deliver money to me while the Win2000 ones are blue-screening !

Where is the corresponding iPhone reference?

A FOSS AIX would probably be of limited usefulness. As I understand it, insanely tight integration with the hardware is one of its greatest strengths.

Rather than going through the intensive effort that opening AIX would require, I would prefer to see IBM invest the resources in adding some of AIX's more useful capabilities to existing FOSS OSes.

And do we really *need* another FOSS Unix? I'm all for 'coopetition' between multiple developer communities. But beyond a certain number of major players, the benefits of the competition become overshadowed by the fragmentation of developer resources.

As I understand it, insanely tight integration with the hardware is one of its greatest strengths.

Really? I just left an IBM shop last year, and while I did not work as the sysadmin, the guy who did lead me to believe that modern IBM machines run AIX in a virtual machine called an LPAR.In fact, he told me that IBM's "best practice" for a machine that is only going to run one instance of AIX, is to run it inside an LPAR, and not on the bare metal.

I was under the impression that virtualization was IBM and AIX's big strength, not tight integration with the hardware...

Well, I'm basically repeating my interpretation of what another member of this forum, for whom I have great respect, has said in the past. And you know how screwed up interpretations can be. So I'll defer to anyone who has more actual experience with AIX than I. ;-)

IBM AIX and Linux do run on bare metal on RS/6000-pSeries-System p and they call the hypervisor ( some kind of enhanced firmware ). AIX nor Linux is not hosted within another OS.
IBM ported AIX to another hardware platform ( Intel IA64 Itanium ) but canceled it.
AIX is tightly integrated with the HW for good reasons. The firmware is Open Firmware ( not PC BIOS nor EFI ).
You can dynamically load device drivers and configure them w/o rebooting. AIX is a POWER (POWER4, POWER5) RISC based OS. It's a vertical integrated platform not unlike the Apple Macintosh platform.

Ha! I used to run that a LONG time ago and it was great. But it wasn't really WPS... It just fudged with the window decorations, fonts, and a few other things (the main useful one being adding right-click context menus). It just made Windows 3.1 _look_ sort of like WPS, but it was mostly just cosmetic.

It was valuable for OS/2 users because it at least gave the illusion that their windows programs were running in WPS windows (i.e. visual consistency). Hell, I ran it on plain old Win3.1 for a while (i.e. no OS/2) - just cause I liked the way it looked.

So it was just lipstick on a pig so to speak. It would be of no use to Voyager or OSFree.