Re: Threads docs updated in CVS

From: williamkempf_at_[hidden]
Date: 2001-08-08 17:27:15

--- In boost_at_y..., Ross Smith <ross.s_at_i...> wrote:
> williamkempf_at_h... wrote:
> > 
> > OK, last time it degenerated into a flame war, so I bowed out.
> 
> Me too.
> 
> > Let's
> > keep the flames out of this and maybe we can get somewhere.  You
> > think that it's to pthread specific, but I totally disagree with
> > you.  Regardless, however, even if it were 100% exactly like
> > pthreads, in what way does this make the library unusable to you?
> 
> It doesn't make it unusable, any more than plain unwrapped pthreads 
or
> winthreads are unusable. It makes it, just like pthreads or 
winthreads,
> sufficiently less than optimal that I'd rather use my own library.

In what specific ways?  So far the complaints (unless I miss the 
intent of what you said) are along the lines of atexit() should be 
called onexit() so it's less than optimal.  Or in a few cases you 
seem to complain that the interfaces provide more functionality than 
*you* need, making them less than optimal.  The first form of 
complaint is nothing more than a religious war, and there's no way 
any library will make everyone happy in this regard.  The second form 
simply makes no sense to me, since it's always better that a library 
provide more functionality than less (provided it doesn't complicate 
usage).  I assume you have specific and detail reasons with 
legitimate reasoning behind them.  That's what we need to hear.

The one exception to these two categories is your dislike of 
condition variables, but they are technically superior to Win32 
events and I won't budge on their inclusion.  If you want to argue 
for the inclusion of an event type in addition then we can talk, 
because there are valid uses for this simplified synchronization type 
as well, but it won't be added in the initial submission.

> I did say that I wasn't interested in starting the debate all over
> again. I answered the question because it would have been impolite 
to
> ignore it, but I'm beginning to regret it. Bill, I'm willing to 
discuss
> it with you in private email if you like. If you'd rather not, I'm
> perfectly happy with that too. But I'm not willing to start it up 
on the
> list all over again, where I'll just get drowned out all over again.

I agree that we do not want to start the pointless debate from last 
time over again, either here or in e-mail.  However, if you can 
provide specific examples of what you thing is wrong with supporting 
reasoning then the discussion would be helpful for everyone and 
should remain here on the list.  I truly do want to know specifically 
what you find wrong with the interface.

Bill Kempf