On Wed, 29 Apr 2009 22:16:55 +0800, Clement Clarke wrote: >The Jol Universal Command language (which can be used as a replacement >for JCL and Clists, and which can run jobs in Batch or TSO) has quite a >few symbolic variables preset. > ... >Jol is a free form scripting language similar to PL/I and Rexx. > >It is available for Z/OS and a Windows version can create JCL to be >submitted to the mainframe. > Does this imply that on z/OS it does not submit JCL?
Does it perform DSN ENQs en masse to preclude the possibility of deadlock? Can it be used to create multi-file tapes? To do this reasonably, one needs (analogues of) VOL=REF and RETAIN, available in JCL but not in TSO. >> You are not the only one. I've always found it frustrating that &SYSUID is >> the only available symbol/variable that is available for use in batch. >> Seems to me that there should be many more. Aside from the temp data set >> names you can use - and this is no news to everyone - we have to hardcode >> EVERYTHING in JCL. IBM should start to look at JCL like more of a scripting >> language, IMO, and provide a lot more of what you originally posted about. >> I'm the third one. Alas, what we want is contrary to the design objectives of JCL, which needs to perform a static assessment of resources required by a job in order to avoid deadlocks and preventable locking of idle resources. The suggestions frequently made here that the ambiguities between Reader values, Converter values, and Execution values could be resolved by providing multiple symbols are naive (or perhaps sarcastically rhetorical): given the need for static assessment, probably only the first, certainly not the last, is technically feasible. But granted that much, the wish most frequently expressed here is for time and date to incorporate in data set names. For this purpose, the Reader time would be widely useful; it corresponds closely to the tailoring, scripting, exit, and periodically updated INCLUDE member circumventions that many of us have used. We're adults. If new system symbols were available in batch JCL such as RDRTIME and RDRDATE with the obvious mnemonic value, we're capable of understanding that if a job lingers in the input queue for 6 months, those variables will have old values, not current ones, even as we understand the similar behavior of our tailoring etc. circumventions. As an analogy, SDSF allows me to sort on a job's submit time, start time, or completion time. In fact, I've chosen the first because it is notionally closest to my concept of chronological order. Users should be given a choice between the hazards of DYNALLOC and the exaggerated (in my view) uncertainties of static time variables. -- gil ---------------------------------------------------------------------- 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
