ISO/IEC JTC1/SC22/WG5 N1385

France's (AFNOR) contribution to SC22/WG5 2000 Summer Meeting


The French GT5 held its last national meeting before ISO summer meeting
in Paris. As unfortunately no member of our group can attend the Oulu
meeting, due to lack of funding, we intended to affirm our national
body position on points to be discussed at Oulu.

(We adopt here lettering of N1381, Oulu meeting announcement.)

A. Interpretation of F95 standard.

French proposal for Interpretation 001 (visibility of a data object):

Date: 4 May 2000
To  : J3
From: French GT5
Subject: Interpretation 001

ANSWER:
To avoid byzantine discussion and endless patches to the Standard, we suggest
to adopt the ADA point of view on statement (FOR loops in Ada) indices: such an
index is a MUTE (non declared) scalar integer variable, strictly LOCAL to the
statement, and contextually KIND-typed by the statement indications (interval
described). It has no correlation (except for its name) with any accessible
entity bearing the same name.

EDITS [from F95 standard]:

Page 280, Clause 14.1.3, first paragraph:
replace second sentence by:

	It is a scalar INTEGER variable that has the type parameter
	corresponding to the highest range of the scalar-int-exprs in the
	implied-DO; it has no other attributes.
	
and add the paragraph proposed by J. Reid in SC22WG5.1753 of 30 March 2000.

Page 281, first paragraph: corresponding replacement of the second sentence
(except for "forall-triplet-spec" insted of "implied-DO");
also add the paragraph proposed by Zongaro in SC22WG5.1758 of 30 March 2000.


B. Review of the content of Fortran 2000.

	Should a vote occur on the MTE items within the 3 proposed
categories, we should be grateful to have the present classification be
considered as the french National Body (AFNOR) vote, in spite of our not
being physically attending the meeting.

(B) 1. Items to be included:
	B3 (useful for Object Programming!)
	B4 (important)
	B5 (a "must" for a modern language; cf. Java, C, even PL/1!)
	M2, M6
	M7 (no more technical reasons to stick on upper-case letters)
	M15
	
    2. Items to be excluded:
    	B7 
	M1 (no much work, but where to stop?...)
	M5 (just because we feel it too heavy a burden, even if it can be
	    favourably seen from a theoretical point of view).
	M10
	M11 (we feel that procedure pointers should solve the basic problem).
	R4a, R4c (this "little pieces" approach seems hasardous; it might be
		  better to develop interval arithmetic later as whole).
		  
    3. Items to the discretion of J3:
    	B6 and M4	(for which we nevertheless are in favour)
	B1, B2 and M3	(for which we really feel indifferent)
	
     We have no idea on M16 and M17...
     
     
C. Review of F2k WD:

	We suggest to... add a (very)M.T.E. on binary formats for files, to
	adopt the portable XDR format (of SUN origin, but in the public
	domain).

----------------------------------------------------------------------------
	
Title: XDR binary format for files.
Submitted by: France (AFNOR).
Status: For consideration.

Basic Functionality: Allows to use the XDR portable binary format for files.
Rationale: Should ease file exchange between processors and languages.
Estimated impact: low (AFNOR proposes to address this new functionality,
	should WG5 and J3 accept the present proposal.)
	
Detailed Specification: Accept "XDR" (or any other keyword, such as
   "XDR_UNFORMATTED") every time that "UNFORMATTED" is used in the standard.
   
   Page 133, Clause 9.1.2 (unformatted records):
   Add a new paragraph at the end of the page:
	All the Unformatted records of a file can be specified to have the
	portable XDR format (non processor-dependant).
	
   Page 141, Clause 9.3.4.4 (FORM= in OPEN):
   (1st line of first paragraph):...FORMATTED, UNFORMATTED or XDR. The...

   Page 156, Clause 9.6.1 (Inquiry specifiers):
   after UNFORMATTED=, add:
	or XDR= scalar-default-char-variable
	
   Page 158, Clause 9.6.1.10 (FORM= in INQUIRE):
   ligne 2: suppress "and"
   ligne 3: at the end of first sentence, add:
   	, and is assigned the value XDR if the file is connected for XDR
	portable unformatted input/output.
	
   Page 158, add a new clauxe (9.6.1.12 "bis"):
      XDR = specifier in the INQUIRE statement
	The scalar-default-char-variable in the XDR= specifier is assigned
	the value YES if XDR is included in the set of allowed forms for the
	file, NO if XDR is not included in the set of allowed forms for the
	file, and UNKNOWN if the processor is unable to determine whether or
	not XDR is included in the set of allowed forms for the file.

HISTORY: None.
	
-----------------------------------------------------------------------------
(End of proposal, and of Oulu meeting AFNOR document).
------------------------------------------------------------------------------