I know how to eliminate the web server of piggy-bank :-) Just write a
protocol handler, like this one:
http://darwin.zoology.gla.ac.uk/~rpage/lsid/
This should make the piggy-bank xpi smaller and firefox load faster.
The question is, is this just a hack or a hack that leaves room for the
bank architecture to grow in its own way, unconstrained by http?
What's the name of the protocol scheme? "bank:"? "semantic:"? "semweb:"?
(I think Mozilla already has an rdf: scheme) Like in these:
bank://local/...
bank://simile.mit.edu/...
What should go after ...?
What about peer-to-peer bank data sharing? Can I point to your piggy-bank?
directly: bank://your-machine-ip/...
indirectly: bank://simile.mit.edu!your-user-id/...
What semantics do we need to integrate into these URIs? Federation
sources, UI format (XUL vs. HTML), etc.?
David
Received on Wed Feb 02 2005 - 02:43:17 EST