[boost] Re: Nuemric Conversion review

From: Andy Little (andy_at_[hidden])
Date: 2003-12-20 08:23:17

"Thorsten Ottosen" <nesotto_at_[hidden]> wrote in message
news:bs0hn4$rct$1_at_sea.gmane.org...
> "Andy Little" <andy_at_[hidden]> wrote in message
> news:bs039d$2d2$1_at_sea.gmane.org...

> Andy, your input is just as important as anybody's---newbie or not.

I'd just like to add...
 I hope the problems (which I understand are Not Fernando Cacciolas fault in
any way and must be pretty frustrating) are sorted out soon, and
I can try the converter out.

I would like more examples of useage, in the docs. I just like the:
[Example
    LetsDoSomethingReallyStupid() ;  // !!!
end example]
 approach
(Yep.. I am a simple guy :-) )There is a lot of math and grammar style
things which would be eased/lightened for  me if translated into some short
example code. It just sorts out 'a' reality of an abstract idea for me. (But
I may need this less when I can run it)

The most interesting part for me is the UDT  (User Defined
Type)classification stuff. IF there were some mechanism(s) for
classification of UDTs then I hope that
a lot of designers of UDTs would attempt to get their classes to work with
it(them). Or at least complain vociferously that the classification is way
off for their project.
As an example .In my physical_quantities type I take the easy way out.
Basically you are stuck with inbuilt types as the value_type. That is not
satisfactory for most on boost... but the simple alternative is no generic
conversion between physical-quantities with different value_types.... unless
anyone has other suggestions. And
Fernando Caccciola does appear to.

regards
Andy Little