<IBM> Mainframe and "UI's" (GUI or not)
in [Cobol]
|
Prev: Lack of graphics downfall of mainframe was Re: Java is becoming the new Cobol
Next: New draft of COBOL revision available
From: William M. Klein on 6 Jan 2008 00:50 (follow-on to another thread) In a "different world - long ago and far away" IBM and Microsoft were working and planning together. They jointly (with MUCH yelling and finger-pointing in the background) came up with the wonderful world of OS/2, Release 1.0 running on PS/2 machines. At the same time, (roughly 1990) IBM introduce the marvelous concept of "SAA" (Systems Applications Architecture - I think, but could be wrong on this acronym) which was going to help the ENTIRE world (well all the world as IBM saw it which was primarily doing business on IBM mainframes and S/3x midrange machines) to have a "single well defined UI (user interface)" and this was called "CUA" (Common User Access). At the same time, there were SAA cross platform (S/370, S/400, and OS/2) definitions of things like "COBOL" (eventually '85 Standard plus some extensions), REXX for interfacing applications with the operating system, soon thereafter "LE" which included both callable services that were portable across OS and available from all languages *and* a common run-time that let many languages, e.g. COBOL, C, and RPG (believe it or not PL/I was very late to this game) play well together. But back to "CUA". At the time it was introduced, EVERYONE (not just IBM) who provided end user interfaces to IBM shops wanted to provide "ISPF look-alike" screen interfaces for their applications. ISPF was even (don't ask me the exact timing) introduced for OS/2 on PS/2's. However, IBM (with a LOT of push from Microsoft) did "user interface efficiency studies" and discovered that users worked in a "object action" rather than "action object" way. This meant that CUA had you select an object (eventually an "icon") and then decide what to do with it (based on what that object could/should do) INSTEAD of the traditional ISPF methodology of deciding what to do first and then deciding which object to do it on/with. Traditional IBM mainframe shops HATED it (if not universally, so much so that I can't remember any major independent vendors that produced CUA conforming applications. There must have been some, but not many.) Meanwhile, while things in Boca Rotan (joint home for both IBM and Microsoft work on OS/2) was turning from snide comments to just short of gunshot battles at down", Microsoft decided that "end users" of workstations (aka PCs) really cared a LOT more about the user interface than whether or not the operating system was "truly multi-tasking". IBM, in its usual - at that time - assumption that it not only was omniscient but omnipotent - KNEW that its mainframe customers would NEVER accept an OS that wasn't a workstation version of MVS as far as processing security and approaches go. Hence, the world of Microsoft and Windows 3.1 and ABOUT the same time as combined IBM/Microsoft OS/2, R1.1 (I may have the release number wrong). *** All of this boils down (IMHO) to the conclusion that it IS fair to blame IBM for their failure to get to GUI's for mainframe interfaces, but it was NOT the "mainframe" OS or language people that caused the problem. Rather it was the OS/2 people who failed in providing a multi-tier user environment with "back end" (maybe still procedural) code on the mainframe and a GOOD "GUI" interface integrally coordinated with it and available to COBOL (and RPG for OS/400 and eventually PL/I) IBM mainframe shops. P.S. None of this has anything to do with the life and death of GDDM which was IBM mainframe software that actually DID provide "graphical" interfaces for mainframe-only applications. This software cost lots, used incredible amounts of resource, was expensive, and simply was NEVER a "main player" in the IBM mainframe (extended 3270) world. There were some (medium-common) applications that used it, but I never knew of any shop that actually EXPECTED the "common" end-user to have access to such applications. Such shops probably DID exist, but I never knew them. -- Bill Klein wmklein <at> ix.netcom.com
From: tlmfru on 6 Jan 2008 14:08 William M. Klein <wmklein(a)nospam.netcom.com> wrote in message news:c_Zfj.251310$ZO3.156544(a)fe02.news.easynews.com... > (follow-on to another thread) > > > that were portable across OS and available from all languages *and* a common > run-time that let many languages, e.g. COBOL, C, and RPG (believe it or not PL/I > was very late to this game) play well together. > I was studying PL/1 in university in the mid-60's. Is PL/"one" different from PL/"eye"? PL (I didn't like it in spite of the name!)
From: Judson McClendon on 6 Jan 2008 14:32 "tlmfru" <lacey(a)mts.net> wrote: > > William M. Klein <wmklein(a)nospam.netcom.com> wrote: >> >> that were portable across OS and available from all languages *and* a common >> run-time that let many languages, e.g. COBOL, C, and RPG (believe it or not PL/I >> was very late to this game) play well together. > > I was studying PL/1 in university in the mid-60's. Is PL/"one" different > from PL/"eye"? I studied PL/I at university in the 1970's. But, as far as I know, PL/I has always been spelled with Roman Numeral one ("I"). I Googled "PL/1" to see if anybody out there spelled it that way, and found a page that did. But on the page itself, they reference the "IBM PL/I Family". If you follow the link, it is to an IBM website, where it is "PL/I". Since IBM invented PL/I, they do have the right to name it, and they did: "PL/I". -- Judson McClendon judmc(a)sunvaley0.com (remove zero) Sun Valley Systems http://sunvaley.com "For God so loved the world that He gave His only begotten Son, that whoever believes in Him should not perish but have everlasting life."
From: tlmfru on 6 Jan 2008 16:44 I was curious about Bill's remark below that PL/I was a late-comer. Did you mean, Bill, that PL/I was invented late in the game or just that it was added to the common run-time late in the day? PL Judson McClendon <judmc(a)sunvaley0.com> wrote in message news:w1agj.59513$rc2.31391(a)bignews1.bellsouth.net... > "tlmfru" <lacey(a)mts.net> wrote: > > > > William M. Klein <wmklein(a)nospam.netcom.com> wrote: > >> > >> that were portable across OS and available from all languages *and* a common > >> run-time that let many languages, e.g. COBOL, C, and RPG (believe it or not PL/I > >> was very late to this game) play well together. > > > > I was studying PL/1 in university in the mid-60's. Is PL/"one" different > > from PL/"eye"? > > > I studied PL/I at university in the 1970's. But, as far as I know, PL/I has > always been spelled with Roman Numeral one ("I"). I Googled "PL/1" > to see if anybody out there spelled it that way, and found a page that did. > But on the page itself, they reference the "IBM PL/I Family". If you follow > the link, it is to an IBM website, where it is "PL/I". Since IBM invented > PL/I, they do have the right to name it, and they did: "PL/I". > -- > Judson McClendon judmc(a)sunvaley0.com (remove zero) > Sun Valley Systems http://sunvaley.com > "For God so loved the world that He gave His only begotten Son, that > whoever believes in Him should not perish but have everlasting life." > >
From: William M. Klein on 6 Jan 2008 18:08
I meant that it was added late to SAA. It was also NOT in the first release of LE (Language Environment - the common run-time environment). I always used to joke about the first release of LE supporting COBOL and C (on IBM mainframes) and said that this was a solution to the "burning needs" of all 3 programmers in the world with a combination of COBOL and C on IBM mainframes (in the early 90's). The reason that PL/I was "late" into SAA was that the OS/400 people weren't very convinced that they needed to provide a "conforming" implementation. The internal IBM rules for SAA were (originally) that something couldn't be added unless ALL 3 environments (OS/390, OS/400, and OS/2) had "in-progress" implementations that the expected to provide within 2 years. This is why the original SAA COBOL (in 1990 or so) was actually either the '74 Standard or the '85 Standard at the "intermediate level". IBM had a high-level compiler on OS/390 and Micro Focus could provide one for OS/2, but OS/400 was slow on this too. On the other hand, the OS/390 people *DID* provide an RPG - which I don't know if anyone ever actually used. Both SAA and CUA were very reasonable concepts but incredibly badly delivered and advertized. -- Bill Klein wmklein <at> ix.netcom.com "tlmfru" <lacey(a)mts.net> wrote in message news:6Zbgj.15221$3b7.10225(a)newsfe23.lga... >I was curious about Bill's remark below that PL/I was a late-comer. Did you > mean, Bill, that PL/I was invented late in the game or just that it was > added to the common run-time late in the day? > > PL > > Judson McClendon <judmc(a)sunvaley0.com> wrote in message > news:w1agj.59513$rc2.31391(a)bignews1.bellsouth.net... >> "tlmfru" <lacey(a)mts.net> wrote: >> > >> > William M. Klein <wmklein(a)nospam.netcom.com> wrote: >> >> >> >> that were portable across OS and available from all languages *and* a > common >> >> run-time that let many languages, e.g. COBOL, C, and RPG (believe it or > not PL/I >> >> was very late to this game) play well together. >> > >> > I was studying PL/1 in university in the mid-60's. Is PL/"one" > different >> > from PL/"eye"? >> >> >> I studied PL/I at university in the 1970's. But, as far as I know, PL/I > has >> always been spelled with Roman Numeral one ("I"). I Googled "PL/1" >> to see if anybody out there spelled it that way, and found a page that > did. >> But on the page itself, they reference the "IBM PL/I Family". If you > follow >> the link, it is to an IBM website, where it is "PL/I". Since IBM invented >> PL/I, they do have the right to name it, and they did: "PL/I". >> -- >> Judson McClendon judmc(a)sunvaley0.com (remove zero) >> Sun Valley Systems http://sunvaley.com >> "For God so loved the world that He gave His only begotten Son, that >> whoever believes in Him should not perish but have everlasting life." >> >> > > |