From martin@ocfmail.ocf.llnl.gov Wed Apr  7 03:13:08 1993
Received: from ocfmail.ocf.llnl.gov ([134.9.48.4]) by dkuug.dk with SMTP id AA02336
  (5.65c8/IDA-1.4.4j for <SC22WG5@dkuug.dk>); Wed, 7 Apr 1993 19:12:44 +0200
Received: by ocfmail.ocf.llnl.gov (4.1/SMI-4.0)
	id AA01381; Wed, 7 Apr 93 10:13:08 PDT
Date: Wed, 7 Apr 93 10:13:08 PDT
From: martin@ocfmail.ocf.llnl.gov (Jeanne Martin)
Message-Id: <9304071713.AA01381@ocfmail.ocf.llnl.gov>
To: SC22WG5@dkuug.dk
Subject: Electronic Varying String Ballot
X-Charset: ASCII
X-Char-Esc: 29


To:      WG5
From:    Jeanne Martin, Convenor
Subject: Electronic Varying String Ballot

With the latest WG5 distribution, you will receive a letter ballot on
varying string issues.  These issues were raised by the comments returned
with the SC22 ballot that closed on January 29, 1993.  Relevant information
and arguments (pro and con) appear in documents N876, N877, and N8778 (also
in that distribution).  In case you wish to reply by email, here is the ballot
in electronic form.  [I apologize if you receive this twice; I sent it out
Monday but I haven't gotten a return copy as yet, so I'm repeating it.]

                         WG5 Informal Letter Ballot
                                     on
                          ISO Varying String Issues
	           (resulting from the recent SC22 CD Ballot)


References: N876 Ballot results from the SC22 Ballot to move to DIS
	    N877 Email correspondence re document form
	    N878 Project Editor's response to ballot comments on CD

 1. There is strong opposition to publishing the Varying String standard as
    an amendment to ISO/IEC 1539:1991 (See N876).  There are two other choices 
    (See N877):
     
        - publish as a separate standard with a different number
	- publish as the second part of a multi-part standard with the number
	  1539-2.  When 1539 is revised, it will become 1539-1.

    Which do you prefer?

    ___  separate standard
    ___  second part of a multi-part standard


 2. Should the pointer inefficiency in the example module caused in part by the
    pointer intialization defect of Fortran 90, be alleviated by adding INIT
    and FREE procedures (as described in N876) to the standard?

    ___  yes (add INIT and FREE)
    ___  no


 3. Do you favor an early extension to Fortran to provide for pointer
    initialization in some form?

    ___ yes
    ___ no


 4. Should generic names be chosen for the I/O procedures?

    ___ yes
    ___ no

    Which do you prefer?

    ___ GET, PUT, PUT_LINE
    ___ GET, PUT, PUT_EOR
    ___ INPUT, OUTPUT, OUTPUT_EOR
    ___ other (specify) __________________________________


 5. Should the basic input procedure be simplified such that it always reads
    to the end of record; that is, the only terminator is EOR?

    ___ yes
    ___ no

 6. Should a SPLIT procedure as outlined in N878 be added to the string
    manipulation facilities?

    ___ yes
    ___ no


 7. Should IOSTAT in the input procedure return a positive value (indicating an
    error) when an error is detected other than a data transfer error, e.g.
    during allocation of storage for the string?

    ___ yes
    ___ no


 8. Should a variant of VAR_STR with a length argument be included?  (See
    arguments in N876 and N878.)

    ___ yes
    ___ no


 9. What name should the type conversion procedures have?  (See arguments in
    N876 and N878.)

    ___ CHAR  
    ___ FIX_STR


10. ISO/IEC 1539:1991 currently does not permit a single generic name to apply
    to both functions and subroutines.  As described in N876 and N878, it would
    be useful if REMOVE, INSERT, and REPLACE could be invoked either way. 
    Do you favor an early lifting of this restriction?

    ___ yes
    ___ no


11. Should the document be revised such that descriptions are modeled after
    those in section 13 of ISO/IEC 1539:1991?  See N878 for an example
    description of INDEX.  Note that this would be a major editorial change.

    ___ yes
    ___ no