[SOLVED] Gfortran statement does not compile

rnturn · 09-06-2014, 06:26 PM

Quote:

Originally Posted by suicidaleggroll

I typically use a "|" though...

I used for the ampersand because it was an otherwise illegal character (though I think you could use in quotes as part of a FORMAT statement) if it ever found its way out of column 6 and it means "and" which made sense to me for a continuation character.

rnturn · 09-06-2014, 07:16 PM

Quote:

Originally Posted by PeteTy

format(2hth,2he ,2hst,2hri,2hng,2hs ,2hca,2hn ,2hon,2hly,2h b,2he,2h 2,2h c,2hha,2hra,2hct,2her,2hs ,2hlo,2hng)

Obfuscated Fortran!

jpollard · 09-06-2014, 08:06 PM

Mine was similar, though it didn't happen to me.

The student (I think it was a business student), got help from one of the computer science students. The two got the program working... then sent it to the card punch so the student could turn it in.

But both of them forgot that the card punch didn't do printing on the card...

rnturn · 09-06-2014, 11:10 PM

Quote:

Originally Posted by jpollard

But both of them forgot that the card punch didn't do printing on the card...

IBM sold a big honkin' machine that did the printing on the top of cards fed into it. Not every shop had one. I knew of a few people (my dad was one, a former boss was another) who were to able to read punch cards that had no printing on them.

jpollard · 09-07-2014, 06:40 AM

Quote:

Originally Posted by rnturn

IBM sold a big honkin' machine that did the printing on the top of cards fed into it. Not every shop had one. I knew of a few people (my dad was one, a former boss was another) who were to able to read punch cards that had no printing on them.

We were on a converted IBM site - from a 360 to a DEC System 1077 at school. The single card punch was only used for a couple of years after conversion, so people could get their data "archived" (the most familiar form). Tapes were not "portable" - we could READ IBM tapes, but conversion was tricky and didn't always work (record terminators didn't seem to exist - so it was always whatever the blocking was, but that didn't always match a multiple of the data record length...)

The only time I could read blank punch cards was when it was a boot deck for a PDP-8... Each instruction could fit in a single column... and of course, being binary, had no printing ever. That was what was used (most often) for the PDP-8 remote job submission (used by the business school) node. Could use paper tape (did a couple of times), but the card reader was faster. The early DECNET was point-to-point only (basically only DDCMP, combined with device source/destination) would fit in 4k (barely), so that would be about 80 cards vs about 100 yards of fragile paper tape (likely this the wrong number, I just remember it took forever to load).

rnturn · 09-07-2014, 11:43 AM

Quote:

Originally Posted by jpollard

... about 100 yards of fragile paper tape (likely this the wrong number, I just remember it took forever to load).

Ouch! I would think that something like that that you might be using frequently would have been on mylar. We "inherited" an 11/70 from NASA Goddard (used it for spares and we grabbed the floating point processor board and an extra power supply for use in our existing /70 so we could run F77) and it came with a ton of paper tape but, oddly, no reader. I recall the DEC HW guy telling me that it was a paper tape version of XXDP. We already had it on an RK06 disk pack so the paper tapes just sat in the cabinet as a curiosity. Last time I saw a paper tape reader was in a documentary about NOAA hurricane chasers. The computers they were using on the plane used paper/mylar tape readers to load software. The intense buffeting would destroy any disk drive. Nowadays I'm sure they're using USB and/or solid state drives.

jpollard · 09-07-2014, 03:13 PM

Mylar tape would have been nice... but it wore out readers fast (not to mention the punch). It also had a tendency to jam (slightly stretched mylar would slip - where the paper tape would just tear. Paper tape was easier to repair too).

Most of the paper tape I used was on the PDP-8. The PDP-11 had the punch (it was where the programs were created), but after a couple of boxes of tape it had to be serviced.

I don't think 11/70s ever came with a paper tape reader by default.

Fortunately, the PDP-8 all had magnetic memory. Rarely did it need to be reloaded, but sometimes just restarted (only a second or two then).

The PDP-11s were even easier, IF you knew the trick - halt, reset, load address 0, start.

All of the PDP-11s used physical memory for booting - and the boot block was always in the first 512 byte of memory. The trick was realizing that the kernel didn't use physical memory, but had a kernel memory map - which excluded the first 512 bytes. When DEC came out with the semiconductor memory, they also came out with a boot board that would carry out memory diagnostics, and the trick wasn't as useful anymore.

foadsf · 12-02-2016, 12:51 PM

this is a very old question but the wikibook page is still taking victims

after a couple of hours struggling with the codes I realised that the length of the lines are indeed the issue but the solution offered here is not only not working but also misleading. in those versions of fortran the length of any line can't be more than 72 letters. so longer commands should be split into lines where the $ sign sit on the 6th column of the following lines. so the correct code should be:

Code:

C---------------------------------------------------------------------
C AREA OF A TRIANGLE - HERON'S FORMULA
C INPUT - CARD READER UNIT 5, INTEGER INPUT
C OUTPUT - LINE PRINTER UNIT 6, REAL OUTPUT
C INPUT ERROR DISPLAY ERROR OUTPUT CODE 1 IN JOB CONTROL LISTING
      INTEGER A,B,C
      READ(5,501) A,B,C
  501 FORMAT(3I5)
      IF(A.EQ.0 .OR. B.EQ.0 .OR. C.EQ.0) STOP 1
      S = (A + B + C) / 2.0
      AREA = SQRT( S * (S - A) * (S - B) * (S - C) )
      WRITE(6,601) A,B,C,AREA
  601 FORMAT(4H A= ,I5,5H  B= ,I5,5H  C= ,I5,8H  AREA= ,F10.2,
     $13H SQUARE UNITS)
      STOP
      END

you can compile this code with gfortran and when running it wait till you enter three integers separated by commas (e.g. 4,6,8)

I also fixed the wikipage so other people shall not fell into this trap!

)

RockDoctor · 12-05-2016, 07:19 PM

This thread inspired me to head down to the basement and dig out some FORTRAN code I wrote almost in the late 1970s. Dialect was MNF (Minnesota FORTRAN), which, I believe was more-or-less FORTRAN 77. Looks like my only use of Hollerith format specifiers was for spacing of the output lines; all strings were quoted.