Fortran 2003 standard: which free document gets closest?
Bart Vandewoestyne
It's been a while since I've done Fortran 95 programming (as a
postdoc, I'm forced to work with Matlab code nowadays and
translating it to Fortran is just too much pain in the *ss :-(
but i recently got interested in Fortran again and since F2003
compiler support seems to be quite good nowadays, I would like to
start digging into the F2003 features to keep up with things.
I would like to have aside a PDF-copy of the latest F2003
standard, but since I don't want to spend money on the real
ISO/IEC document, I'm looking for the PDF that gets as close as
possible to the real standard.
On http://gcc.gnu.org/wiki/GFortranStandards one refers to
http://www.j3-fortran.org/doc/year/04/04-007.pdf while on
http://fortranwiki.org/fortran/show/Fortran+2003 we have
http://std.dkuug.dk/jtc1/sc22/open/n3661.pdf which explicitly
mentions 'Final Committee Draft' at the bottom of each page.
Because of this, I would suspect that the last document is the
more final one, but it dates October 2003 while the 04-007.pdf
dates May 2004...
Which document should i trust most as my 'unofficial F2003
standard'?
Regards,
Bart
--
"Share what you know. Learn what you don't."
Aris
I don't know if these are the "best" drafts, but this link is
prominently missing in your list
http://www.nag.co.uk/sc22wg5/links.html
Bart Vandewoestyne
>
>> On http://gcc.gnu.org/wiki/GFortranStandards one refers to
>> http://www.j3-fortran.org/doc/year/04/04-007.pdf while on
>> http://fortranwiki.org/fortran/show/Fortran+2003 we have
>> http://std.dkuug.dk/jtc1/sc22/open/n3661.pdf which explicitly
>> mentions 'Final Committee Draft' at the bottom of each page.
>> Because of this, I would suspect that the last document is the
>> more final one, but it dates October 2003 while the 04-007.pdf
>> dates May 2004...
>>
>> Which document should i trust most as my 'unofficial F2003
>> standard'?
>
> I don't know if these are the "best" drafts, but this link is
> prominently missing in your list
> http://www.nag.co.uk/sc22wg5/links.html
Hmm... so we now have:
1) gfortran website refers to 'WORKING DRAFT J3/04-007' of
May 10, 2004 11:07
2) fortranwiki refers to a working document with reference number
'ISO/IEC JTC1/SC22/WG5 N1578' dated 2003-10-8
3) NAG refers to a working document with reference number
'ISO/IEC JTC1/SC22/WG5 N1601' dated 2004-5-3
Only document 2) has the words 'Final Committee Draft' at the
bottom of each page, but according to the date and the N-number,
does not seem the most recent document to me.
I'm confused...
James Van Buskirk
news:t%llp.1703$Vm5....@newsfe30.ams2...
I ran into the same problem. See the thread:
Short answer: N1601.pdf is the right one if you don't want to
look at f08 and N1830.pdf.
--
write(*,*) transfer((/17.392111325966148d0,6.5794487871554595D-85, &
6.0134700243160014d-154/),(/'x'/)); end
Richard Maine
> "Bart Vandewoestyne" <MyFirstName...@telenet.be> wrote in message
> news:t%llp.1703$Vm5....@newsfe30.ams2...
>
> > On 2011-04-01, Aris <us...@domain.invalid> wrote:
>
> >>> Which document should i trust most as my 'unofficial F2003
> >>> standard'?
...
>
> > 1) gfortran website refers to 'WORKING DRAFT J3/04-007' of
> > May 10, 2004 11:07
>
> > 2) fortranwiki refers to a working document with reference number
> > 'ISO/IEC JTC1/SC22/WG5 N1578' dated 2003-10-8
>
> > 3) NAG refers to a working document with reference number
> > 'ISO/IEC JTC1/SC22/WG5 N1601' dated 2004-5-3
>
> > Only document 2) has the words 'Final Committee Draft' at the
> > bottom of each page, but according to the date and the N-number,
> > does not seem the most recent document to me.
>
> I ran into the same problem. See the thread:
>
>
http://groups.google.com/group/comp.lang.fortran/browse_frm/thread/62144
abf4a84d427/f4d9cd507b8ef6b5?hl=en#f4d9cd507b8ef6b5
>
> Short answer: N1601.pdf is the right one if you don't want to
> look at f08 and N1830.pdf.
ISO terminology can be confusing. The "Final Committee Draft (FCD)" is
"final" only in that the subsequent things are not formally called
committee drafts. You don't want the FCD. There were changes of actual
substance after the FCD - not just typographical stuff. I forget the
details - just don't use the FCD unless you are reasearching the
history.
N1601 or 04-007 are what you want. They should have the same content -
just different form. One is the J3 version (the US committee) and one is
the WG5 version (the International committee). I'm the one who generated
them both, so I'd have a pretty good idea. From my own file of notes to
myself (Even though the docs came from my keyboard, I'd forget these
things if I didn't write them down).
"The f2k (well, f2003) cd is 02-007r3, aka N1497, Sep 2002.
The f2003 fcd is 03-007r2, aka N1578, Oct 2003.
The f2003 dis as submitted to ISO is N1601.
Line-numbered J3 equivalent is 04-007, May 2004.
ISO edited the n1601 pdf into the official DIS and later the
standard. ISO redid page headings, ttle page, and foreward.
They also did items 2-6 mentioned in post-dis-edits.txt."
None of the post-dis-edits stuff is of consequence. It is
Deleted "Draft International Standard" and date from page headers.
Changed the cover page to reference N1611 as the corresponding
working document and to delete "DIS" from the reference number.
Deleted the last para of the foreward, per ISO DIS changes. Did
not change the fifth para; the ISO change there appears incorrect.
Made the following purely typographical corrections.
1. Fixed the formatting of Table 15.2 in subclause 15.2.1 to
avoid extending into the right margin and to eliminate
spurious lines in the continuation table heading.
2. Deleted a stray '"' at the end of subclause 14.4.
3. Changed comma to period at end of item (8) in the second
numbered list in subclause 7.1.6.
4. Italicized "index-name" in the last sentence of the third
paragraph of subclause 16.3.
5. Added a blank after the ")" in constraints C302 (in subclause
3.2.2), C524 (in subclause 5.1), and C541 (in subclause 5.1.2.5).
6. Fixed the font so that the inverted "!" correctly shows as "<"
in example case (ii) in subclause 13.7.73.
No, you can't freely get the actual standard document. Heck, even the
committee members can't get it. I happened to get one freebe copy for
being editor. (Crappy job of binding, particularly for something so
expensive; looks like it was bound in someone's garage - someone not
very good at such things).
The committee themselves worked with the 04-007 document in doing things
like interpretations. It is handy because it has line numbers, which
make it easier to cite things.
--
Richard Maine | Good judgment comes from experience;
email: last name at domain . net | experience comes from bad judgment.
domain: summertriangle | -- Mark Twain
Bart Vandewoestyne
>
> [...]
> N1601 or 04-007 are what you want. They should have the same content -
> just different form. One is the J3 version (the US committee) and one is
> the WG5 version (the International committee). I'm the one who generated
> them both, so I'd have a pretty good idea. From my own file of notes to
> myself (Even though the docs came from my keyboard, I'd forget these
> things if I didn't write them down).
>
> "The f2k (well, f2003) cd is 02-007r3, aka N1497, Sep 2002.
>
> The f2003 fcd is 03-007r2, aka N1578, Oct 2003.
>
> The f2003 dis as submitted to ISO is N1601.
> Line-numbered J3 equivalent is 04-007, May 2004.
>
> ISO edited the n1601 pdf into the official DIS and later the
> standard. ISO redid page headings, ttle page, and foreward.
> They also did items 2-6 mentioned in post-dis-edits.txt."
Thanks Richard. Your answer cleared things out for me.
Nice to see that after me being away for a few years, I can still
rely on your very precise answers! But I also notice that the
answers are still as long and elaborated as back then ;-)
Regards,
Bart (currently forced into Matlabbing, but with a keen interest in f2003/2008)