Re: [ox] what I meant to say...
- From: <graham belegost.mit.edu>
- Date: Fri, 4 May 2001 15:03:58 -0400 (EDT)
On Thu, 3 May 2001, Flo Ledermann wrote:
about building networks of people (with skills, resources, ideas, projects
etc.) with rdf, and the necessary technology to "query" this network, like
gnutella, for things you need or other people interested in the same stuff.
people could publish "themselves", their skills and resources and also the
other people they are "connected" to (whatever this means - it is not clear
to me if we need different types of links like "people i know", "people i
trust", "people i love" etc.) on their website (well, i think thats basically
what rdf is all about ;), and have a simple means for sending requests out to
their "neighbours", who in turn forward the request to their neighbours and
so on. with this structure it should be possible to build quite large sharing
and exchange networks without a central server or mechanisms of power.
so it may be that we are actually thinking about very similar things?
Yes and no! I hadn't thought of creating such a distributed system;
that is, I was assuming that queries would be direct to one server;
if that fails, query another server, etc, rather than servers passing
on queries. I think that what I wanted would probably be much easier
in terms of programming (ie less work ;-)
You asked for more details about what I want, so here it is:
I run a site with a mix of links to programs (like freshmeat), links to
hardware designs, and news stories.
I would like people to be able to query the site and get back the entire
contents of a story, or a program description, or a review, in an XML
format. The hope would be to build up a network of such sites which could
exchange information in this way. Sites would be free to use copies
of whole stories (but would probably have to do their own transformations
on the XML to get viewable html). Or a site could fetch a story, add a
translation to another language, and again make the 2-language version
available. This is unlike current use of rdf
(as far as I know) where usually all you can fetch is a list of headlines
linked to the original site, or discover whether a site is X-rated - in
my limited understanding, rdf is supposed to handle metadata, not data? I
was thinking of something midway between
soap and rdf, but really not dependent on either: just using generic xml.
I have not yet written my own dtd; I wanted to see if others
had ideas about what should be included.
So far I think the format should have two main parts - an 'administrative'
section, not normally displayed as html, and containing information about
things like creation date, originating site, etc, and a 'contents'
section, which may be wholly transformed to html for display, or may
simply be a wrapper for data of other types (images, circuit diagrams,
etc).
This is all very vague, as I have no practical experience of using xml
(apart from netscapes rdf channels). I am assuming it would require a
small custom server running on sites taking part, which could understand
simple queries like 'give me all files with content type x since date
y', or 'give me all files with italian text'.
I imagine this could be made generic enough to include your 'people
finding' application - but I don't have the time or the skill to program
your fully distributed system.
What about you? are you still in the planning stage? Do you already
have the code ready to go? :-)
Graham
________________________________
Web-Site: http://www.oekonux.de/
Organisation: projekt oekonux.de