String interning /servers

[Tyler Baker]

james anderson wrote:

> Perhaps this is yet another argument for a less primitive intern-basis
> than that of String.intern().
> a. One could well intern QNames rather than Strings, with the
> concomitant ability to specify the strength of the reference. This
> entails the "bothersome" static bindings, but i'm actually fond of them.
> b. One could well structure the intern-map to reflect the namespace
> partitions and flush qname-packages on an lru basis.
>
> In all, the server life-span shouldn't stop one from interning.

For applications that use XML Namespaces, this would make dealing with Namespaces much
more efficient as well as a lot more straightforward if what you are saying is to create a
specialized QName object (or interface).

Tyler


xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/ or CD-ROM/ISBN 981-02-3594-1
Please note: New list subscriptions now closed in preparation for transfer to OASIS.