Re: [boost] C++ Standard Library proposal - Math functions forStatistics

From: Gabriel Dos Reis (gdr_at_[hidden])
Date: 2003-04-15 18:11:11

"Powell, Gary" <powellg_at_[hidden]> writes:

| While I haven't looked at this latest proposal, Walter's special
| math fns proposal was well received because it has a C
| implementation/interface, i.e. no templates. 

I made serious comments against the proposal as it stood.  I regret
our "abstain" vote was  interpreted as "don't care" (*sigh*) -- is it
the case that one should vote "no, over-my-dead-body" in order to have
people listen?   Did I vote "no", that the result would  have been "4 - 3".

The C-compatibility argument is a bogus one.  It creats endless
confusion: that is the lesson learnt from overloading things like cos()
without taking care of integer argument.  And if you seriously
consider implementing those functions, you'll realize that the
implementation are the same for all functions involved.

-- Gaby