Herbie, I really must disagree with your assertion that closed-source = "safe". The opposite is more often true.
A couple of points: 1) Open source doesn't mean "open committers". Take the Hercules license as an example: Anyone is free to "fork" the source code, provide their own "version", so long as they make all of the source code available. They are not forced to allow anyone to submit changes to their "version". 2) I never suggested that the open source z-arch reference implementation would be used to run your safety-conscious workloads. 3) I'm not suggesting anything about z/OS or z/VM code (yet). In the end, closed architectures usually die. The market will simply not tolerate them in the long run. Sun has realized this - look at what they are doing with Sparc and Solaris. (They had the "advantage" of being forced to do so :-) IMO, if IBM locks up the z Architecture with IP, it only devalues it to the point of irrelevance. While they are apparently worried about z architecture emulators cutting into their margins for low-end systems, their real competition is in z/OS cross-compilation and API emulator environments, which are improving all the time. Maybe they think that their future is in z/VM virtualization of Linux, but their competition will probably be pretty tough. Consider this: what if z hardware OEMs once again flourished to the point of say a 50% market share? - Wouldn't the overall mainframe market grow, or at least not shrink as much? - Wouldn't IBM still be able to sell high-margin software and services? - How much software and services revenue will IBM get when z/OS customers re-target their COBOL/CICS/JCL to run on MicroFocus-like platforms? Many will disgree; saying that obviously IBM has done the big-picture analysis and has chose their path accordingly. I think that it is more likely a turf issue or a "prisoner's dilemma". Unlike Sun, their success means that they won't be forced to open the z architecture until it is too late. Kirk Wolf On Dec 5, 2007 6:16 AM, Van Dalsen, Herbie <[EMAIL PROTECTED]> wrote: > Kirk wrote: > >> IMO - > >>- The architecture *interfaces* (POP) should be completely public > >>- An open-source software-based *reference* implementation should be > >>available (Hercules++ ?) > >>- Software should be available through something like PWD at costs > that > >>promote developers of all sizes to support the platform > > In my opinion, what makes IBM code safe in terms Auditing risk, is the > fact that only IBM labs work on it. You need a really P'd-off IBMer to > plant a Trojan in the code, and a few P'd-off testers to miss it during > testing. So I would not be in favor of open source for the mainframe. I > think too many companies depend on the current quality level of the > software. What I would be in favor of is a platform where developers > outside of IBM can present new software designs/ideas to be included > after proper securitization. > > Regards > > Herbie > > ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
