(j3.2006) (SC22WG5.4564) F08/0038 comments

[Malcolm Cohen]

Bill Long writes:
>The text in 13.2.4, on which the subsequent edits depend

No they don't.

>relies on a poorly defined concept of "reduction".

I see no problem with this text.

All the edit in F08/0038 does is to remove a potential contradiction.  If you 
don't think something is a reduction, then there wasn't a contradiction there in 
the first place!

>It would also cover the cases that currently would not be considered
>"reductions": FINDLOC, MAXLOC, MINLOC, ALL, ANY, PARITY, and THIS_IMAGE.

Huh?  ALL, ANY and PARITY are all rather obviously reductions, and seeing as how 
MAXLOC is specifically mentioned in the text in 13.2.4 already clearly FINDLOC, 
MAXLOC and MINLOC are too.

Anyway, 13.2.4 is merely there to explain the reasoning behind DIM and its 
restrictions, those restrictions actually being stated in every function to 
which it applies, so there is no need for a fundamental rigorous definition 
here - we aren't proving any subsequent theorems from it.

>COUNT seems to be missing ... an oversight.

Not an oversight.

Cheers,
-- 
................................Malcolm Cohen, Nihon NAG, Tokyo.