Gil wrote: >No, you're thinking of ISPF. As infrequently as possible!
<snippage of various ISPF gore> >Item: CMS ISPF is worse. Well, yeah. Without even any details, we know this to be true! >Item: I could leave my office, inadvertently leaving an XEDIT session >running. At home I reconnected with a different terminal geometry. >XEDIT queried the screen size and redrew the screen, losing only >uncommitted changes to the screen. When I wish for similar facility >in ISPF, I'm called absurdly unrealistic. I suspect this is because CP is better about terminals than VTAM ever was. Not an excuse--it's IBM's ecosystem, they should have had similar support in the "premier" OS--but I bet that's the underlying answer. IOW, with VTAM, there wasn't an equivalent of DIAG 8C to return the screen geometry. And perhaps they didn't have the same level of error handling that DIAG 58 provides, so when the write went bad, they couldn't recover gracefully. So this may have seemed like black magic to the ISPF developer (I assume they're down to one part-timer, since it doesn't seem to actually improve over time). >It feels as if ISPF developers accepted Rexx unwittingly with an >attitude approaching sabotage. It does. And that might be part of the reason that so many z/OS folks still say "Oh, I don't use Rexx". >VM20779? Wayback Machine? GIYF? Not. Heh. Circa, um, 1984? 1985? It was a huge APAR that changed RESERVEd lines to be per-screen instead of being global to XEDIT (among other things). In retrospect, pretty clearly done to enable FILELIST et al. >Circa 1985 we had been using, in a small way, an ISV display manager >(XMENUE?) when one user had a need to upgrade from CMS/370 to CMS/XA. >Our minor use didn't justify the cost of upgrading the display manager; >I converted to XEDIT. Function improved. I could rely on XEDIT >facilities to scroll text entry boxes; insert/delete lines (relates to >the OP's requirement), etc. Impressive indeed. I wondered if XMENU would come up! In 1985 it was still owned by Kolinar. In 1988 we (VMSG) bought Kolinar, and the two guys from there joined us. XMENU (aka XMENU/E, for a while, in the period when IBM was adding /E to things) was a very robust and mature CMS display manager, complete with a WYSIWYG screen creator/editor. I did a lot of work on XMENU in the 1990s, clearing up some long-standing problems. My favorite was a hassle with fields using extended attributes (the SA order): If you had an extended attribute field where, say, each byte was a different color, XMENU wanted to optimize the screen rewrite after you hit an AID key. But the MDT wasn't set for the entire field, and so that made it difficult. I finally realized there was an unused bit in the extended attribute byte, and could use that as my own per-character MDT when needed. Just had to make sure to clear it before sending the data stream back to the terminal! (And if you understood the above, congratulations, you, too, have wasted a ton of brainpower and memory on 3270 data streams!)
