Prev: Problem with automatic reallocation of allocatable scalar on assignment
Next: a wiki entry for gfortran
From: Terence on 8 Aug 2010 19:39 On Aug 9, 4:01 am, glen herrmannsfeldt <g...(a)ugcs.caltech.edu> wrote: > mecej4 <mecej4.nyets...(a)opferamail.com> wrote: > > Richard Maine wrote: > > (snip) > > >> Probably f66 compiler manuals. IBM was quite late in releasing their > >> f77. I forget the exact date. > > (snip) > > It seems that there is no wikipedia entry for VS Fortran... > > > I have a copy of IBM GC26-3986-1, Release 1.1, the 2nd. edition, January > > 1982, VS FORTRAN, Application Programming: Language Reference. The preface > > says "... includes full ANS FORTRAN plus IBM extensions". > > So, 1982 for release 1.1, maybe 1981 for release 1.0? > > > ENTRY is definitely there, and is covered in pages 81-83. > > ENTRY was in the IBM OS/360 (Fortran G and H) compilers, so it would > have also been in VS Fortran, as an extension, even if not standard. > > -- glen Yes, for Fortran IV on mainframes, but not on IBM or Buroughs compilers for table-top computeers, with the MS-written F77 used (with a slight modification by Burroughs) for both machines (it more or less says so in the Burrroughs manuals, where they added a far more detailed i/o failure code section). I couldn't post a reply to an earlier Glen contibution - wouldn't take the short paragraph.
From: robin on 8 Aug 2010 22:16 "Ron Shepard" <ron-shepard(a)NOSPAM.comcast.net> wrote in message news:ron-shepard-42257F.11315808082010(a)forte.easynews.com... | In article <4c5ed409$0$34573$c30e37c6(a)exi-reader.telstra.net>, | "robin" <robin51(a)dodo.com.au> wrote: | | > Maine has been wrong on both matters that he contributed :- | > 1. He said that integer operations were produced by floating-point op codes, | > and as a consequence, that results were 48 bits. | > (untrue; instructions for 60-bit integer addition and subtraction had | > separate op-codes from the floating-point equivalents); | > 2. that of the 64-character set, CDC used one character value (the double | > colon) | > to signify end-of-line, which created a conflict because there were | > then effectively 65 different purposes obtained from those 6 bits. | | I'm not particularly interested in this discussion of character sets | on ancient machines that I never used, but I have to say that the | above two sentences are not accurate representations of posts to | this thread for the past few days. Really? Here is a direct quote re point No. 1 :-- "From: "Richard Maine" <nospam(a)see.signature> "Newsgroups: comp.lang.fortran "Sent: Saturday, 7 August 2010 1:58 AM " > An interesting thing about integer arithmetic operations was that... > there were none - at least none with their own dedicated opcodes. You > just used the same opcodes as for floating point, and the hardware > special-cased operands with the 12 high order bits all 0 or all 1. It > seemed like the machine was designed to crunch floating point numbers, > with integers being a special case." | You have to wonder why this incorrect summary would be posted when | it is so easy to read back through the threads to see what Richard | Maine actually said. Obviously Ron Shepard, you didn't go back far enough. | $.02 -Ron Shepard
From: robin on 8 Aug 2010 22:23 "Ron Shepard" <ron-shepard(a)NOSPAM.comcast.net> wrote in message news:ron-shepard-42257F.11315808082010(a)forte.easynews.com... | In article <4c5ed409$0$34573$c30e37c6(a)exi-reader.telstra.net>, | "robin" <robin51(a)dodo.com.au> wrote: | | > Maine has been wrong on both matters that he contributed :- | > 1. He said that integer operations were produced by floating-point op codes, | > and as a consequence, that results were 48 bits. | > (untrue; instructions for 60-bit integer addition and subtraction had | > separate op-codes from the floating-point equivalents); | > 2. that of the 64-character set, CDC used one character value (the double | > colon) | > to signify end-of-line, which created a conflict because there were | > then effectively 65 different purposes obtained from those 6 bits. | | I'm not particularly interested in this discussion of character sets | on ancient machines that I never used, but I have to say that the | above two sentences are not accurate representations of posts to | this thread for the past few days. | | You have to wonder why this incorrect summary would be posted My description at point No. 2 (above) is a correct description. Maine attempted to put another spin on it.
From: robin on 8 Aug 2010 22:30 "Richard Maine" <nospam(a)see.signature> wrote in message news:1jmwkxc.apqohnfuffswN%nospam(a)see.signature... | Ron Shepard <ron-shepard(a)NOSPAM.comcast.net> wrote: | [on Robin's paraphrase of me] | | > I have to say that the | > above two sentences are not accurate representations of posts to | > this thread for the past few days. | | I'd agree. Seems like every time he cites someone, including himself, he | has to paraphrase the citation to twist it into whatever he wants to | claim it said instead of what it actually said. (But then when you quote | his exact words, apparently that's a misrepresentation of what he said | because those words really meant something else). This is a familliar | pattern. I see no point in detailing the corrections. I'll now go back | to ignoring Robin's posts as not worth responding to. You can't face up to the fact that you posted erroneous material and that I corrected it. Initially, you posted incorrect material on two points on the CDC machine (integer arithmetic and character set). Lately, you posted a third erroneous statement about character handling on the CDC machines, which I again corrected. Instead of making insulting and derogatory remarks about someone else when you are shown up to be incorrect, why don't you take a good look at yourself !!
From: LR on 9 Aug 2010 00:09
mecej4 wrote: > It seems to me that the last few posts in this thread (about 64 or 65 > characters on the CDC) indicate that the colon may have been used as an > escape character, somewhat like the backslash, \ , in C, with a pair of > backslashes being then needed to signify the literal character. > > This is speculation on my part, but this may throw some light on why two > colons had quite a different effect than the once-repeated effect of a > single colon. Not quite. For one thing, it's not really colons, just two octal zero six bit characters in the last position of a 60 bit word. But you could have the line or record or whatever we might call it end before those last two characters, for example, for a seven or 17 or 27 etc character line, then you'd have three octal zero characters terminate the line. For a six or 16 or 26 etc character line, four octal zero characters terminate the line and so on. A nine or 19 or 29 etc, character line would have had 11 zero octal characters terminate the line. And finally a ten or 20 or 30 etc character line would have ten octal zero characters terminate the line. This implies that if we're to follow Robin's convention, it would really be more accurate to say that there were 64 display characters + 11 termination characters for zero terminated records = 75 characters. But of course, CDC referred to them as the 63 and 64 Display character sets. As has been pointed out this leads to two problems. 1) You might possibly end up with two colons on in the last two positions of a word, resulting in a problem. 2) Ending a line with any number of colons using Display 64 wasn't possible. Appendix A in the PLI reference has some nice tables. The footnotes tell more of the story. http://www.bitsavers.org/pdf/cdc/cyber/lang/pl1/60388100B_PL_I_Version_1_Reference_Manual_Oct79.pdf LR |