Translators---compilers and assemblers---are usually most comfortable
with FB source-language card-image inputs.

There is much to be said for supporting VB input, since the chief
determinant of a  translator's performance is the facility and
rapidity with which it eliminates insignificant blanks from
source-language statements.  (A minimal alternative would be to supply
an exit, as does the HLASM, that makes it possible to reformat inputs
of other sorts into FB ones.)

If I were the deity symbol-table management would be more important
than blank suppression; but, as John Cocke showed conclusively, this
is not the case.  It is almost a law of nature that intellectually
interesting problems turn out in the end to be practicallly less
important than humdrum housekeeping issues.

--jg


On 10/6/12, Ed Gould <[email protected]> wrote:
> On Oct 6, 2012, at 12:07 PM, Paul Gilmartin wrote:
>> ----------------SNIP-------------------
>> For the ultimate irony, did the FORTRAN compiler source code
>> input with RECFM=VBS?
>>
>> -- gil
>>
> Gil:
> The G1 compiler needed FB (80) I never got to tinker with the WatIV
> as the project never needed the compiler so it was ordered and
> installed but essentially never used other than a few trials and error .
> I think (if memory serves me) the G1 did not like terminal I/O very
> well (it worked but was found wanting).
> The TESTFORT command was $$ but was needed and I suppose we could
> have dropped it but debugging was a semi ongoing issue.
>
> Ed
>
> ----------------------------------------------------------------------
> For IBM-MAIN subscribe / signoff / archive access instructions,
> send email to [email protected] with the message: INFO IBM-MAIN
>

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to