I'm not an expert, and I've never worked from IBM so I'm relying on second and third hand information, but here's what I understand:1. From a "licensing issue", I've read that IBM owns more source code from OS/2 for the PPC then it does the Intel version of OS/2. So from a straight licensing issue, there would be fewer roadblocks to IBM open sourcing code from OS/2 for the PPC, however
2. From what I recall, IBM had developed a compiler and tools which would allow application developers to be able to compile their native OS/2-Intel code for OS/2 PPC with minimal modifications. If portions of code was taken from the Intel OS/2 and just recompiled for OS/2 PPC using such tools, and if such code was owned by 3rd parties originally, then that code would be subject to the same licensing as on the Intel version.
3. For IBM to open source any code, it costs them money - they have to go through the code and make sure they own all the rights, then it all has to go through their lawyers, etc. So for IBM to do this, spend the money, there needs to be a *strong business case* for doing so. IBM is a business, they don't do costly stuff just because it seems like a nice thing to do. They want to see either some way that they'll benefit in the future, or some benefit to their customers.
The big example of this is JFS. JFS came originally from AIX, and is used by many IBM customers (big customers who spend lots of money). There were probably a lot of those customers looking to adopt Linux, but obviously could not sacrifice the data and information currenlty on AIX (using the JFS file system). So IBM by open sourcing JFS, getting it to run under Linux - they were able to keep those customers happy. And those customers will probably continue to look to IBM for help with their AIX and Linux needs. The cost of open sourcing JFS probably was a lot less then the profits they have gained by keeping those customers happy (or perhaps attracting new customers).
That leads to the question: Is there a stong *business case* for open sourcing OS/2 for the Power PC? An operating system which had so few people interested in it that they decided to pull it from regular distribution and make it a "special order only if you called PSP directly and knew the proper part # to order" less then a month after its official release?
4. The final point is that I've read rumors that people at IBM may not know where all the source code for OS/2 for PPC is. As I recall the plug was pulled around January 1996, at which point development of the product ground to a halt. I have heard rumors that there were some bug-fixes and even a few fixpaks which came out for it, and there was code which later turned up in Warp 4, but nothing substantial. And the group which was known as PSP at the time has gone through several reorganizations since then. While I believe the source probably still exists somewhere within IBM, I can also believe that it might be difficult to even find all that code. Certainly the majority of people who originally worked on it have moved on to other things...
Just my thoughts.