On Tue, 01 Feb 2005 21:43:13 -0500, David Huynh <dfhuynh_at_csail.mit.edu> wrote:
> I know how to eliminate the web server of piggy-bank :-) Just write a
> protocol handler,
Hi David,
My kneejerk reaction to the suggestion of a new URI scheme is
'urrgghh' after recently seeing the feed: approach to RSS feed
subscription [1] - it works, but definitely falls into the category of
ugly hack. But the problem there is that it isn't really a protocol
like http, ftp etc, more like a behavioural thing like mailto:.
So what I'd like to ask is what is the purpose of the protocol you
want the scheme for, and what are the endpoints which will use it?
(Bear in mind I haven't looked inside Piggybank yet ;-).
If the purpose is to access the triplestore, then maybe a possible
approach would be to use something like the http extensions of URIQA
[2]. It should be possible to implement only those parts of the
service that are needed to access the local store (i.e. not a full Web
server), whilst still remaining generally compatible with the http
protocol at large.
Assuming I'm in the right ballpark here, I'm personally interested
because I'm looking for a good pluggable, interoperable way of
accessing triplestores. There doesn't yet appear to be any common
approach that would allow me to easily swap between say Jena and
Redland-based backends. (Although the SPARQL protocol looks
promising).
While I'm posting I might as well ask another question - is there an
easier way of getting at the source of Piggybank than the web
interface? (Sorry, another quick question that the source would answer
- is the HSQL DB store now part of Jena, or something you've built
custom for Piggybank?)
Cheers,
Danny.
[1]
http://www.25hoursaday.com/draft-obasanjo-feed-URI-scheme-02.html
[2]
http://swdev.nokia.com/uriqa/URIQA.html
--
http://dannyayers.com
Received on Wed Feb 02 2005 - 08:38:30 EST