________________________________ From: Frank Swarbrick <[email protected]> To: [email protected] Sent: Mon, March 1, 2010 7:21:52 PM Subject: Re: Item on TPF
I guess I should note that I am speaking only as an application developer. So I come at it from a different perspective than most of those on this list. But a few things I miss from VSE: - Superior JCL symbolics. - System level symbolics availble for use in batch JCL. - JCL "DATE" card to override system date. - Return code checking that actually makes sense (can anyone give a good reason for how COND works?). - LIBDEF statements for setting/overriding a library search chain. - JCL symbolics available to batch FTP client. - JCL (JECL) FNO parm for setting FORMDEF/PAGEDEF/FORMS/etc. by specifying a single name. Those are what I can think of offhand. I notice that all of my beefs with z/OS have to do with JCL. Interesting; I'd never really thought about it in that context. Frank Frank, The JCL date is rather interesting but the concept is unique to dos as far as I know dos doesn't come close to a shared spool like MVS does. I would suspect that if dos ever allows shared spool the issue would have to disappear or the concept would have to change. As to the cond I would (like you) probably not say it is not one of MVS's best implementations. It truly does make sense in an off handed way, the problem is as always is compatibility. It was not implemented in a user friendly manner and so it was burnt into the OS as far as compatibility forever more. I am not all that familiar with LIBDEF's but if you are talking steplib or joblib there is more enough flexibility as far as I am concerned. I cannot tell you how many times a programmer picked up an incorrect module because of steplib/joblib problems. Since I know nothing about libdefs I would suggest that you look at steplib/joblib and see if it isn't superior to libdef. There are one or two issues that cause issues with the steplib and that is with authorization checking that I am guessing that DOS does not have. But again MVS is vastly more superior in security than DOS is. The JCL symbolics for FTP is a mixed issue as there are both good and bad issues with having them. I suspect (but do not know) if DOS has anything like the MVS convertor/interpreter it is the big green monster from hell (pardon the english) and there is no easy interface to it (that I have seen documented). I would suggest that while it might seem "nice" to have it, I would suggest that this "short coming" is really a blessing. As to the pagedef/formdef item. The 3800 was not invented when the C/I was written and adding support for it (from what I heard) raised a ugly head at IBM. I vaguely remember it was rewritten in the 1990's (early) but there was only so much they could do. I am not excusing it but the implementation was pretty much in line with the rest of JCL. I am not claiming JCL is "best" it works reasonably well with the guarantee of backward compatibility that IBM is famous for. Just remember one thing MVS is IBM's Flagship OS and IBM does a damn (pardon the english) job to make sure there is compatibility between releases. I think its been said on here many times that most of the programs that were compiled and linked 30 years ago still run unchanged today (there are exception of course) but at most sites I have been working at a upgrade in the OS is not any major issue as IBM goes to great pains to maintain compatibility. Its been years since I have worked on DOS (probably 25 and that was dos under VM) it was a PITA then from what little I read its improved a little but noting close to the improvements in MVS in the same time frame. Ed ---------------------------------------------------------------------- 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
