To all responders, Thanks for your contributions.
It appears that the SYST, system task, attribute implies a "property" additional to those I gave in my first post which are documented with the description of the attribute in the MVS Initialization and Tuning Reference manual. According to John McKown this "property" allows the program to select a protect key other than 8. VTAM was mentioned as an example but VTAM's is one of the PPT/SCHED entries from which I removed SYST with no ill effects. Here I am again in very - mixing metaphors - rusty territory for me. Perhaps, again, there is a "compensation" for the lack of this "property" and maybe it is AC=1 for a load module where the original load module resides in an "authorized" library - or something like that. Whatever - it would seem that VTAM, as normally installed, has the necessary "compensation" as maybe do a few other programs from whose entries in PPT/SCHED I removed the normally present SYST attribute. According to Pat O'Keefe, the same lack of "ill effect" - apart from the message - seems to apply to TCP/IP for MVS (that is, CS IP as well as CS SNA). Incidentally a relatively short-term interest I had in VSPC, along with APL, got me a month or so stay in the environs of San Jose, CA, so VSPC is one of my more warmly remembered products. As a matter of chronology, SCHED became an alternative to PPT during the time I took a more detailed interest in the interstices of PARMLIB (1980 onwards) so that puts the date later than the birth of IP and its related protocols. Maybe the concepts behind the PPT are what was considered to be set in stone - to use another metaphor. I wonder if there is some aspect of the "system task" other that being free of time constraints and requiring to be initiated by a START command which at least prefers a single step. Chris Mason ----- Original Message ----- From: "Craddock, Chris" <[EMAIL PROTECTED]> Newsgroups: bit.listserv.ibm-main To: <[email protected]> Sent: Tuesday, 28 February, 2006 6:06 AM Subject: Re: SYST > > >It means no more and no less. There are no other "sinister" effects. > > >>... > > > > Given that, it's too bad they didn't pick a different name A task is > no > > more nor less a system task based on the SYST flag. For example, it > is > > entirely reasonable for a TCP/IP proc to first execute the program > > defining TCP/IP's subsystem interface, then execute TCP/IP itself. > That > > nullifies the SYST attribute (resulting in a nastygram in the log) but > in > > no way effects the way TCP/IP behaves ... as long as TIME=1440 is > > specified. TCP/IP is still a system task (as opposed to SYSTEM TASK) > > either way. > > True enough, but the SCHEDxx stuff was fully baked long long before TCP > was even a glint in some DARPA brainiac's fevered imagination. It's a > pity that IBM didn't update the PPT processing code to accommodate such > modernities as ARM, but like everyone else they have limited resources > and such a change would be far down the priority list. > > CC ---------------------------------------------------------------------- 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
