Re: [boost] LWG reaction: Header <cstdint>

From: Howard Hinnant (hinnant_at_[hidden])
Date: 2001-11-07 13:01:31

On Wednesday, November 7, 2001, at 12:03  PM, Kevin Lynch wrote:

> Of course, since I'm not actually on the committee, in the end, it
> really won't matter what I think if people really are screaming for
> double sin(int); if that's the case, I'll just have to go back to using
> Fortran (blech!) :-)

Now we wouldn't want THAT to happen! :-)

The complaints that I've seen are usually of a C compatibility nature 
and/or migration nature, for example:

> I've got ten zillion lines of code that is littered with statements 
> like:
>
> 	root2 = sqrt(2);
> and
> 	pi = 4 * atan(1);
> and even
> 	void foo(int i) {double x = sin(i); ...}
>
> It runs fine with Vendor X, but not with your library! (turns out that 
> Vendor X hasn't added the overloads prescribed in 26.5/6 to <cmath>, 
> for example)  I'm going with Vendor X because I can't afford the extra 
> maintenance to port to your lib.

So I've got economic pressure to make their code work (if there are 
"enough" of "them").  As a physicist, I'm sure you will find my 
non-technical argument distasteful.  However it is the reality I live in.

And to make things even worse, please note that simply "int" is not 
enough.  C99 says (and C90 accepts) "of integer type".  When you get to 
actually implementing this in C++ with multi-argument methods like pow, 
it can quickly turn into a nightmare (although compiler magic can help 
out a lot here).  But alas, I seem to be arguing against myself! :-)

I suggest that you do not include this issue in your post.  It is not 
your job to represent Gaby and me.  I /am/ on the committee (and so is 
Gaby).  I'm capable of representing this issue myself.  Although it is a 
completely different matter whether I'm capable of /successfully/ 
representing this issue! ;-)

Represent those issues you are comfortable with.  I want the inverse 
hyperbolic functions too.  I appreciate your voice on /that/ issue.  
Focus where your strength is.  If you are distracted by the int-arg 
issue, it takes away from your ability to represent important issues 
that we both care about.

-Howard