From: Howard Hinnant (hinnant_at_[hidden])
Date: 2001-08-15 12:11:00
On Wednesday, August 15, 2001, at 11:36 AM, David Abrahams wrote:
> That said, it is only used as an /implementation detail/ of other boost
> libraries. There is no reason a vendor couldn't implement any of the
> library interfaces without using boost::type_traits (except that it
> would be
> a lot of work). From a standardization POV, I don't see how library
> implementation dependencies are relevant to which libraries are
I think what you say is true. It's just that I think ignoring lib
dependencies will make an already difficult task even more so.
As a hypothetical example, consider a proposal for compressed_pair. If
it is going to take compiler magic to find out if we can optimize away
the space of one of the members, we need to consult with the core
working group which can significantly complicate and slow things down.
On the other hand, if we already know how to do this from a previous lib
submission that was looked on favorably, then we can concentrate more on
the question of whether the functionality and interface provided by
compressed_pair is suitable for standardization.
On a related subject, I'm wondering if it would be a good idea to brief
the lwg on the status of boost::threads. I personally have not been
keeping up with this one. But I do think it is a strategically
important piece of work, even if it has not reached maturity.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk