[J3] (SC22WG5.6273) RE: [ukfortran] Fortran 90 and BOZ literals

[Malcolm Cohen]

BTW, Fortran 90 (1991) predates PDF (1993), so whatever N692 is is someone's attempt at a fixup.

(And it's not groff - groff only appeared in 1990, a bit late for typesetting a document we were working on several years prior. According to Wikipedia, the "first stable version" of groff was November 1991, several months after publication.)

If I recall correctly, the Fortran 90 standard was typeset using Sun-proprietary macros (and possible Sun-proprietary troff technology). Furthermore, due to the use of proprietary technology, (again IIRC) that troff source code was not supplied to other committee members. The Fortran 95 source started out with a Framemaker reincarnation of Fortran 90 (and I recall it had some glitches from the conversion process).

Unfortunately, the one person who really knew what happened, and thus could explain the details properly, has recently passed away; that is, Walt Brainerd.

So w.r.t. the Fortran 90 standard, the "gold standard" is the published document. In particular, there was no PDF. If someone sometime somehow could scan a published copy in, that would be useful as a historical document, but apart from personal or committee use, would doubtless be a violation of ISO copyright.

Cheers,
-- 
..............Malcolm Cohen, NAG Oxford/Tokyo.

-----Original Message-----
From: Robert Corbett <rpcorbett at att.net> 
Sent: Tuesday, August 11, 2020 6:44 PM
To: General J3 interest list <j3 at mailman.j3-fortran.org>
Cc: WG5 List <sc22wg5 at open-std.org>
Subject: [ukfortran] (SC22WG5.6272) [J3] Fortran 90 and BOZ literals



Sent from my iPhone

> On Aug 11, 2020, at 2:14 AM, Shterenlikht, Anton via J3 <j3 at mailman.j3-fortran.org> wrote:
> 
> 
> 
>> On 11 Aug 2020, at 09:59, Robert Corbett via J3 <j3 at mailman.j3-fortran.org> wrote:
>> 
>> The file
>> 
>> https://j3-fortran.org/doc/year/90/S8.115.pdf
>> 
>> contains a late draft of the
>> Fortran 90 standard.  I trust
>> it more than I trust the N692
>> document.
> 
> Interesting document.
> How close is it the published standard?
> There are lots of handwritten notes - did those made it to the 
> published version?
> 
> Just a few pages later, in 4.3.2.1
> the 4 examples of nondefault character literal constants apper blank 
> in N692, while they are present in S8.115.pdf
> 
> Another glitch in N692 is in 4.3.2.1.1, last line before the numbered 
> list, after:
> 
> "For the default character type, the only constraints on the collating sequence are:"
> 
> there is:
> 
>    @.EQ delim $$ @.EN
> 
> which is groff (or troff at that time?) gone bad - that groff command 
> is meant to instruct groff to use the dollar sign as a delimiter for 
> inline equation environment.
> But seems it was interepreted as a literal text.
> 
> Perhaps somebody somewhere still has that g/troff src?
> If so, we can try to fix it and produce the correct Postscript/PDF.
> 
> Anton


_______________________________________________
ukfortran mailing list
https://lists.accu.org/mailman/listinfo/ukfortran