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

Re: record-mutator vs record-modifier



   Date: Tue, 03 Jan 2006 12:48:25 -0800
   From: Per Bothner <per@xxxxxxxxxxx>

   However, SRFI-76 uses record-mutator where SLIB uses record-modifier.
   Is there any reason for the for this difference?  I think that the
   old name is better than the new name, so why change it?  I gues the
   word "modifier" is somewhat overloaded (it could also be something
   like an adjective or adverb that modifies some other action), but
   at least it's a "real word", unlike "mutator".  The arguments either
   way don't seem strong either way, which argues for using the old name,
   especially given that other record-XXX functions keep the old names.

Aside from the reason Mike gave, there's also the point that 'modify'
often refers to taking the old value of a field, passing it to some
function, and setting the field's value to be whatever the function
returns; e.g., ((RECORD-MODIFIER <rtd> <field>) <record> (LAMBDA (X)
(+ X 1))) might increment <field> in <record>.  This use of the term
'modify' dates at least back to T in the eighties, where there was a
MODIFY macro that worked on its generalized SET.  I'm not fond of the
term 'mutator' either, but I think 'modifier' is no better; I think I
might prefer RECORD-SETTER, but I don't have a really good suggestion.

And I'd rather not have the reflection interface specified at all, or
at least relegated to a different specification.