On Fri, 21 Mar 2008 17:20:24 -0500, Kirk Wolf <[EMAIL PROTECTED]> wrote: >I hesitate responding to your post - its mass might otherwise tend to kill >this thread, which IMO would be a good thing :-)
I'm with you on that point... however .... ;-) >I agree with most of what you've written. IBM needs to transition from >worrying about protecting z/OS source code to worrying that nobody will >want to see it. Having done development in the Linux kernel (for System z) and in z/VM (a mix of source and OCO), having source code is of limited use. If you don't know *why* it works the way it does, the source is meaningless. Over time you can certainly teach yourself *how*, but not *why*. One of the most difficult problems I faced writing the original IUCV support for Linux was the complete lack of professional high-level and low-level design materials, not the lack of source code. And that source code is in C, where the "how" is more obvious than in my native language (assembler). It just didn't help. And, sure enough, there were later changes in the internal "interfaces" that I unknowingly misused. Nuances I had missed. It worked, but it wasn't pretty and ultimately to be re-written. Don't read the above as being anti-Source; it isn't, and I'm not. I just I hope it illustrates that source code is not a panacea. For some things it is necessary, but not sufficient. >I have a different view of Peter Relson's posts - I think that he is only >trying to explain which parts of control blocks are meant to be interfaces >and which are not. This distinction has no analog in modern operating >systems, where interfaces are expressed *entirely* by APIs and service >routines, and not by skipping through PSA->ASCB->etc.etc. Control blocks aren't the problem. Understanding the why's and wherefore's of the values in them is the problem. They're in the same control block, so they have a relationship, but what is it? I think Peter's most important point was that of stability. The more you use with things the author didn't intend you to use, the more risk you add to your system. I think that the clear delineation and respect for "application interfaces", "product-specific system interfaces", and "non-interfaces", has in no small part contributed to the longevity of the platform. Compatibility is not an accident. >In any case, I appreciate Mr. Relson's (and other IBMers) contributions to >IBM-MAIN, even if I don't always like what they say. Glad to hear it! "Intelligent minds may differ." >A sore point with me is always "submit a requirement", but that's >another topic. When you feel up to it, please start another thread and let's find out why. It's not the first time I've heard that. Alan Altmark z/VM Development (25+ years) IBM ---------------------------------------------------------------------- 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
