errata/282: signing

From: Steven Sharp (sharp@cadence.com)
Date: Mon May 03 2004 - 15:40:00 PDT

  • Next message: Alec Stanculescu: "Re: errata/282: signing"

    The following reply was made to PR errata/282; it has been noted by GNATS.

    From: Steven Sharp <sharp@cadence.com>
    To: Shalom.Bresticker@motorola.com, alec@fintronic.com
    Cc: etf-bugs@boyd.com, btf-dtype@boyd.com
    Subject: errata/282: signing
    Date: Mon, 3 May 2004 18:52:36 -0400 (EDT)

     Alec,
     
     You cannot tell someone what they meant when they wrote something;
     you can only ask them.
     
     Have you asked the people responsible for the signed arithmetic proposal
     what they meant? I have. You are trying to tell me that they meant
     something different from what they say they meant. I have to conclude
     that you are wrong, and that they meant what they say.
     
     You are including the datatypes task force in the distribution of this
     discussion. As a result, I wonder if your real concern is the problems
     that would occur if the rules for signedness were applied to the new
     type system. This concern could have arisen because the 2001 LRM uses
     the term "type" when talking about signedness. It tries to roll reals
     in as a type also, and claim the same rules apply, but this is wrong
     since different rules apply to reals, and always have.
     
     I agree with you that these rules probably won't work if applied to
     datatypes in general. They don't even work for reals. The rules are
     applied to the signedness of vector expressions, which is why they need
     to work similarly to the rules for sizes of vectors. Just because the
     LRM text uses the term "type" for signedness does not mean that new types
     will have to follow the same rules. It just means that if they don't,
     the signedness rules will have to stop using the term "type".
     
     Note that the correct rules for reals do not convert non-real operands
     to real until they are about to be combined with a real at an operator.
     Subexpressions that will be implicitly converted to real this way are
     treated as self-determined until that point. This behavior is similar
     to most programming languages, and is what apparently seems natural to
     you. When it is documented correctly, this could form a starting point
     for rules for other datatypes.
     
     Steven Sharp
     sharp@cadence.com
     

    This archive was generated by hypermail 2.1.4 : Mon May 03 2004 - 15:40:12 PDT and
    sponsored by Boyd Technology, Inc.