From: Kevlin Henney (kevlin_at_[hidden])
Date: 2000-11-07 11:56:04
In message <188.8.131.52.2.20001106133711.00b8dcf0_at_[hidden]>, Beman
Dawes <beman_at_[hidden]> writes
>I believe "any" should be accepted by boost.
>The docs would benefit from additional guidance to users:
>* What's a simple, compelling example of a problem which class any solves
>in better way than alternative solutions?
Yes, I think this would be a useful addition. I will add a worked
example (a source file linked from the docs), and also structure the
examples section more explicitly.
>* When is class any the technique of choice for representing a generic
>type? IOW, what problem characteristics indicate class any is the right
>* When is class any NOT the best technique for representing a generic
>type? IOW, what problem characteristics indicate class any is the wrong
>tool? An inheritance hierarchy with a common base class (or whatever) is
>sometimes a better way, but how do you decide?
I will add a usage section.
>* Is class any itself CopyConstructible, Assignable, EqualityComparable,
It's currently documented that any satisfies ValueType requirements, but
this should go into more detail.
>* What Standard Library containers can class any be used in? Is there a
>way to have a std::set of any?
This can go into the usage section.
>* What's the typical sizeof() of an object of class any?
Interesting: Do we wish to be documenting at this level for Boost
classes in general?
Kevlin Henney phone: +44 117 942 2990
Curbralan Limited mobile: +44 7801 073 508
mailto:kevlin_at_[hidden] fax: +44 870 052 2289
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk