[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: "types" or "records"?



   Date: Tue, 13 Sep 2005 11:17:05 -0500 (CDT)
   From: Donovan Kolbly <donovan@xxxxxxxxxxx>

   Some of the terminology in this SRFI seems inconsistent.  The procedural 
   layer talks about constructing "record type descriptors" (i.e., 
   make-record-type-descriptor), but the syntactic layers talk about defining 
   types (i.e., define-type and type-descriptor).

   The very title of the SRFI suggests that it's about records.  Shouldn't 
   the syntactic layer be consistent with that?

I too dislike the term 'type' as used here and would rather that it be
restricted in standards documents only as a term for an abstract
concept in the language, not anything specific as it is used in this
SRFI.  I believe DEFINE-TYPE is an artifact from Gambit, and I'd
rather see the old name DEFINE-RECORD-TYPE or some variation thereof.
Indeed, unless there is a good reason otherwise, I think it would be
better to support the old DEFINE-RECORD-TYPE syntax as the base,
extended for inheritance and automatic field initializers, and then to
build anything more sophisticated later, possibly on top of that.