Slashdot Mirror
User Group Urges IBM To Open OS/2
axonis writes "A report on Tom's Hardware tells of one of the last active OS/2 user groups, which has announced an initiative to garner support for IBM to release its long-neglected OS/2 operating system into the open source community. IBM announced earlier this month that it will withdraw its operating system OS/2 officially from sale on December 23 this year and will offer support only through 2006." From the article: "Making OS/2 Open Source will benefit all IBM customers that had invested in this OS...Customers that are willing to continue using OS/2 will get the benefits of an open OS that will be continuously developed by individual developers and/or software companies, their ownership fees will decrease and they will have the enhanced security of an OS that will continue to be relevant due to the open-ended nature of open source (following the BSD and Linux examples)."
Another open source OS would be welcome. At the very least ideas and features can be examined and possibly implemented in the bigger players (Linuxes). But diversity is always good, and what does IBM have to lose?
Unless of course they are making a successor, but that doesn't seem very likely.
The article explains exactly why IBM wouldn't release it as open source. They don't want customers to get their support via the open source community... They want their customers to buy support from IBM, regardless of the operating system that must be used.
Make's sense, it could put more pressure (as if there isn't enough already) on the competing open-source community.
It was originally a colaboration between MS and IBM. So chances are MS owns some of the code.
I still have more fans than freaks. WTF is wrong with you people?
Do non-disclosure agreements expire necessarily? Or is that something that would only happen if it was written into a contract for some reason?
-Valiss
They really have nothing to gain from open sourcing OS/2 and potentially a lot to lose from doing so.
If Solaris is any example, it costs money to open source code. You have to pay someone to scour the code for inappropriate or confidential information.
Lawyers need to work through any licensing agreements with third parties and so forth.
They're potentially exsposing themselves to lawsuits by showing their knickers to the world. I mean for all we know OS/2 could be filled with stolen UNIX source code and the last thing IBM wants is to actually validate SCO's claims!
Bottom line is that IBM has nothing to gain from spending (wasting?) money to open source OS/2. It's a shame, but that's life.
Maybe that's why the article wants IBM to release "as much OS/2 code as possible", so maybe we have an OS/2-Lite version, (sucn as the 4.4BSD-lite Unix version without the AT&T code) so that the community can fill in the blanks later.
I was at an ATM in a convenience store last summer during a thunderstorm. The power went out and when it came back on, I watched the ATM boot. Guess what? OS/2. There is no way that IBM's lawyers are going to let that code loose so that people can pick it apart. Just the suggestion probably gives them visions of a pony-tailed hacker going from ATM to ATM and filling his Volkwagon mini-bus with cash.
IBM has open sourced about as much of OS/2 as it is going to. OS/2's file system (JFS) was opened up as well as IBM's Omniprint driver. So it isn't like we can really claim that IBM is entirely opposed to opening up OS/2. They've already opened large swaths of it to be rewarded by constant complaints that what they've opened isn't enough.
The balance is probably so tainted by third party licensed code (and not only from Microsoft) as to make separating out the IBM code from the third party code an expensive proposition.
One thing that might be interesting is that there was an unsupported IBM WPS clone that could replace PROGMAN.EXE as the shell for Windows. It might be interesting if that particular skunkworks product could be released as well, but I'm not going to hold my breath.
So here goes: After reading the first wave of posts it seems that there are other entity's source code in OS/2. So is it possible for IBM to make available its source code for OS/2 only? If they provide the code with gaps, couldn't those in the Open Source Community fill them in? My gut tells me that to do so would be far too complicated for the benefits, but not being a Software Engineering type I don't know for sure.
If brevity is the soul of wit, then how does one explain Twitter?
Of course given that most of the kernel was written in assembler gives it limited practicality, but it would be an great exercise in kernel design to look at OS/2's SMP engine that was so wickedly fast.
Couple of things: 1. The objection was that BSD requires attribution -- and the claim was the MS was still using BSD code but not giving attribution -- therefore violating the BSD license (which allowed them to use it) therefore violating the copyright. I don't know whether that's true or not -- but that was the claim. 2. You're serously asserting that slashdot posters advocate downloading copies of windows xp? (Legal or not?) I don't think I've ever seen that. I've seen a lot of Microsoft bashing -- and a couple of times I made the suggestion that a very reasonable thing for the Federal Government to do was to refuse to handle ANY Microsoft copyright violation cases while Microsoft failed to comply with antitrust laws - or the consent decree microsoft also completely ignored. The last time I made that suggestion was at least three years ago -- I probably posted as an AC -- I read slashdot, but didn't post much and didn't have an account. So: Criticise slashdot posters for what they really do -- oversimplify the issues and demonize microsoft. Copyright scofflaw'ing has a small amount of support, but it's certainly not the norm. BTW - I think Microsoft deserves much of the abuse it gets -- I just wish more of it was well reasoned, rather than knee-jerk.
Right, and security through obscurity works so well
For what has been a quite secure system (ATMs wouldn't use it otherwise), and that is at EOL, actually it is.
Opening up a code base is the best way to get a stable, bugfree codebase *in the end*, but it certainly has growing pains. Particularly since you don't have an incremental model, suddenly you expect everyone to look at everything. Chances are developers would look at interesting features, crackers would run around looking for holes.
What purpose would it serve to push bughunting and patching back into high gear at this point? It'll do nothing to serve the customers in neither the short nor the long run, since in the long run they won't be using OS/2. Get them migrated over on Linux, and let it die. Ask IBM to port any interesting bits (if they haven't done so already) to Linux.
Kjella
Live today, because you never know what tomorrow brings
Well, IBM could clean it up. The best way would be to rewrite all portions that can't be licensed under an open license; the easiest way would be to simply rip out the offending parts and tell the community "this and that is what the parts we had to remove did, so you'll have to replace them before you get something usable again". Obviously, the former would be better for the users and those still interested in OS/2, but the latter would be a very good starting point, too - remember that in projects like OpenOffice.org or Mozilla, a vast portion of the original code was eventually replaced, too, even though the original code *was* avaiable.
That being said, it's certainly sad to see OS/2 come to an end. I remember buying a copy of OS/2 2.0 when it came out, in 1992 or so - it was *very* impressive indeed. What I liked most was the fact that even very hardware-specific DOS programs like demos or games could still be run without difficulties in most cases; I don't know how IBM managed to pull this off, but they did an admirable job.
I'll always have a soft spot in my heart for OS/2.
quidquid latine dictum sit altum videtur.
They've flooded the market with the damn things. However, there IS hope. Several of the smaller players are already going over to Linux, being fed up with MS charging what they charge on the machines and not having anything more stable than they have to begin with. It's just going to take time (Hell, those self-check-out kiosks in the stores lately use Windows XP, not even embedded XP- I know, the silly thing crashed in the middle of my checking out over at Wal-Mart and they rebooted it in front of me, like it wasn't anything at all (Like it was commonplace...)) and more people being bit hard by the damn stuff MS has been jamming down their throats.
I am not merely a "consumer" or a "taxpayer". I am a Citizen of the State of Texas
OS/2, on the other hand, hasn't even been fixed for over four years.
Maybe you should have gotten a clue when the guys who invented OS/2 lost interest in it. Those who knew it best are mostly using something else nowadays.
Get over it.
OS/2 1.x, which came out in the late eighties, not early nineties, was targetted at the 80286 and probably did have significant portions written in assembler.
By the time OS/2 2.0 came out, a large amount had clearly been rewritten in a higher level language, presumably C, as it had gone from being targetted to the 80286 to the 80386, and there's little reason to believe that IBM had done this retargetting by rewriting significant portions in 80386 assembler, which just about everyone would have seen as bone-headed.
More to the point, by the early nineties, IBM was demonstrating PowerPC versions of OS/2 in public.
I find it improbable that the bulk of OS/2 is written in assembler today. I suspect large portions of the kernel may be, but, well, replacing the kernel isn't exactly difficult, given there are so many portable, optimized, kernels in existance that could form the basis of a replacement.
You are not alone. This is not normal. None of this is normal.