From what I understand, there would be fewer licensing issues for IBM to open-source OS/2 for the PPC then there are with the Intel version of OS/2. IBM did a lot of the development of the PPC version internally, so presumably they own most of that code. On the other hand word has it that there were a lot of external developers who contributed code to the Intel version of OS/2, so open sourcing that code would require a lot more effort and money.But the real difficulty that I see is that for IBM to open source *any* code relating to OS/2, there needs to be a *strong business case* for doing so. IBM is a business, they are not going to invest time and money to open source code unless they can see a strong *business reason* for open sourcing that code, and that business case needs to include a return on their investment (the cost of open sourcing the code at a minimum).
To understand what I'm talking about, just look at some of the code they have chosen to open source. Example:
JFS: Remember that JFS originally came from AIX and was ported to OS/2. When IBM did eventually open source JFS for use in Linux, they used the OS/2 code base (which had been improved from the AIX base), however they took the OS/2 specific stuff out of it (at least that's what I understand). But what's the business case and return on investment for open sourcing JFS? Simple. It allows companies who have a big investment in AIX who were asking about Linux to be able to use Linux and keep the same file system. Less worry for those companies. IBM gets to keep them as AIX customers and if they play their cards right they'll be able to gain contracts to maintain those Linux systems. It makes sense, and the return on the investment compared to the cost of open-sourcing JFS is probably quite significant.
Object REXX: Keep in mind that REXX was originally designed to offer a cross-platform scripting language, which could run on IBM mainframes, servers, and other IBM offerings. By open sourcing Object REXX it can help to ensure that Linux (or any other OS for that matter) can have an up to date implementation of REXX, helping to give continued life to those IBM mainframes and servers by helping to ensure that REXX stuff can run on the latest Linux boxes and the older IBM stuff. Again, it keeps the customers happy, and is more likely to keep them with IBM.
The reality is, I havn't yet seen a strong *business case* for open sourcing PPC OS/2. It would cost IBM money to do this, and if I were an IBM manager, I'd ask "what return on our investment will we get by open sourcing this code?". Remember that PPC OS/2 had almost no customer base.
I'm sure the PM, SOM, and WPS code would be useful in trying to get a WPS equivelant running on top of another kernel. But in my opinion before IBM even considers releasing this code, the question that needs to be answered is "how would this benefit IBM?" What return would they get for offering this code? How would it help keep current IBM customers with IBM, how would it bring something to Linux (keeping in mind that IBM's current strategy is based on Linux) that would be a benefit to IBM customers considering Linux? And how could it bring new customers to IBM? So far I havn't seen any of these questions really adequately answered, and as long as they remain unanswered, I think it is unlikely that any of this code will be released. Of course I could be wrong.