"Robert Ramey"
We are working on a library for parallel computing that uses the serialization library to ship data via MPI. If you reject the proposed integration of fast array serialization into the library, we will need to encourage 3rd-party authors of serialization for types containing arrays to use some non-standard fast mechanism ...
of course this is what I haven't seen.
What haven't you seen?
I would expect that a "fast...archive" would include the specific functionality for certain types.
Like what types? double for example? What interface would you propose for serializing an array of doubles?
Then users of those types with those archives would automatically benefit from the specialized functionality
Not unless the authors of serialize functions all use the specialized functionality when it's available.
while preserving compatibility with other archive types. this has been done before - for example strings are serialized differently in some archives as opposed to others. I havn't seen anythiing differentabout this situation.
Strings are a closed set of types. In this situation we have a large _category_ of otherwise-unrelated types that can benefit from one particular optimization. But that benefit can only accrue if the authors of serialization functions use a specialized interface for array serialization whenever possible.
Every way we can think of to accomplish that eventually leads to something that feels like "hijacking" the library. We _really_ don't want to do that.
OK - I'll take another look at mattias code and see what can be done.
Thanks, I really appreciate it. -- Dave Abrahams Boost Consulting www.boost-consulting.com