>Basically we need to write our own, more effecient, NamespaceSupport
class.
>Joe was going to work on this, but I'm not sure he's found the time.
I haven't really gotten started on it; I desperately had to spend some time
catching up on the DOM... My initial thought was to do our own
lighter-weight and self-contained hash, similar to the DTMStringPool,
though that's arguably overkill given how short the namespace context lists
are. Binary search thru a sorted list would be more compact but not quite
as fast. Even linear may be worth considering given how short these lists
are. Standard storage-space-versus-speed tradeoffs; simple code, so the
right thing to do is probably to run a few experiments and see where the
sweet spot is.
