Path: csiph.com!weretis.net!feeder6.news.weretis.net!panix!.POSTED.panix2.panix.com!not-for-mail From: docdwarf@panix.com () Newsgroups: comp.lang.cobol Subject: Re: Need an example CALL USING RETURNING Date: Sun, 7 Apr 2024 01:42:41 -0000 (UTC) Organization: Public Access Networks Corp. Message-ID: References: Injection-Date: Sun, 7 Apr 2024 01:42:41 -0000 (UTC) Injection-Info: reader1.panix.com; posting-host="panix2.panix.com:166.84.1.2"; logging-data="23626"; mail-complaints-to="abuse@panix.com" X-Newsreader: trn 4.0-test77 (Sep 1, 2010) Xref: csiph.com comp.lang.cobol:12864 In article , R Daneel Olivaw wrote: >docdwarf@panix.com wrote: [snip] >> I've sorked in shops where the SEARCH verb was forbidden, spit out by >> pre-compiling tools, because 'people get confused by it' (or Chief Senior >> Programmer didn't understand it). Same with SORT. Same with INSPECT >> REPLACING because 'you Never Know when the subroutine will be wanted in >> the Online region and that's a Bad Thing.' >> >> It's frustrating but... their shop, their rules. Find where they keep >> their skels ('templates'), read the code and learn their songs.' [snip] >I have never worked on a site where SEARCH or INSPECT REPLACING were >forbidden, that's ludicrous. Their checks to me cleared the bank. Early on in my career I figured that if the shop had been kept 'healthy' they wouldn't have need of my skills. One place required an In Progress folder that contained source listings and copies of the File Descriptors, a hex dump of a data record and the fields underlined and bracketed (kind of [__FLD1__]]__FLD2__] 'for documentation'. I found out they had access to FileAid, a utility that would print out a list of 01 FILE01REC. 05 FLD1 (X06): 'VALUE1'. 05 FLD2 (X06): 'VALUE2'. 05 FLD3 (906): 000003. ... and so on. I started putting those into my folders and it garnered notice, the Group Lead said it made life a lot easier and I ought to put together some samples and documentation for the other programmers... ... but since that wasn't part of the Work Order that paid me he wanted me to do it on my own time. I said 'You've got to be joking' and his response was 'I'm as serious as a heart attack, on your own time and within a week.' I laughed, said something about a snowball fight in Hades and he actually made noises indicating that if I didn't give free time for this my contract might be 'reconsidered'. I said 'If I'm not here then you're sure not to get this. If I'm not paid for my hours then you're sure not to get it. If you can get a forty-hour work order cut so I can dedicate a week to this your people will love you and you'll make back that money from time saved by your team in the first week afterwards. I'm going to lunch.' He got the work order, I put together JCL and control statements and documentation, the next week at least three people stopped by my cube and said 'You've made our lives so much easier!'... ... and at the end of that week I gave my two weeks' notice, that place was too sick even for me to tolerate. >SORT is a bit different, some of the programs I wrote ran in Transaction >environments, and there were memory constraints. These programs did not >normally perform any sorting anyway, but in one case - sorting a table >in memory - I actually managed to get a C sorting library to work for me. >"Transaction environments" could be "Online" in your terminology. That hearkens back to 'no INSPECTs because it might go online', yes. I had a situation about twenty-five years ago where the input was a multiple-record-type dataset, similar to a header-rec for customer info and then a bunch of transactions under it; I got some COBOL from... maybe it was Mr Mosely, maybe from Mr Svalgaard's (sp?) ETKPAK, it bubble-sorted the transactions in date sequence so a spiffy set of tallies could be kept... it's still running in Prod, somewhere. They were giants in those days. DD