- From: Dan Brickley <danbri@w3.org>
- Date: Wed, 16 Aug 2000 20:32:38 -0400 (EDT)
- To: Seth Russell <seth@robustai.net>
- cc: RDF SIG <www-rdf-interest@w3.org>, eMouth <emouth@egroups.com>
On Wed, 16 Aug 2000, Seth Russell wrote: > Dan Brickley (on another group) wrote: > > > ... we've > > been running a shared-bookmarks experiment. Bookmark files were > > collected from a bunch of us and aggregated into a common XML/RDF > > database. > > I've been running a little experiment of my own with > shared-bookmarks. Any chance of getting hold of a sample of this > XML/RDF ? (context: this was from the new rss-dev@egroups.com group where we're discussing a proposal for an RDF/namespaces friendly v1.0 of RSS. See http://purl.org/rss/1.0/ and the egroups.com for full'n'frank debate... I was discussing some bookmark-sharing work happening at the institute in which I work in Bristol.) Seth: yes and no! This was initially at least an in-house experiment and we don't have privacy clearance to flow the data more widely than within the group of volunteers (who all know each other etc etc). Ultimately I'm convinced that the future of bookmarks is in this kind of data flow, and that it'll become a part of good webbery to allow non-private comments/annotations/ratings to be exposed to the web at large. But that's my personal hobbyhorse... (or one of them...) The actual model for representing the bookmarks as RDF models is something we're discussing. Libby Miller is doing the real work (and might be encouragable into summarising here) - I just scribble things on whiteboards and handwave about it being in-essence the same kind of data as sitemaps, channels and rating server data. I know the issue of representing sequences and taxonomies came up, but we have a number of possible ways of projecting the browser-exported data into a server-side RDFdb. Once it's there, idea is to encourage (through collaboration tools, autoclassification, location-independence etc) people to manage bookmarks/ratings/annotations serverside. The exact RDF model is something that'll settle down over time. My inclination is to stash it all as date-stamped events, eg 'on this data person foo recommended resouce bar with comment blah etc etc', but other models are of course possible. hope this helps, --danbri ps. in a similar vein I started an installation of the http://crit.org/ toolsuite at W3C. Dysfunctional installation at http://crit.w3.org/ (my fault -- I don't think the server has write access to disk so ratings get lost). I want to testbed that as a public rating/comments server, eg. to see if we can integrate it with the Amaya annotations work. Anyone interested in helping with that? or with scoping where ratings/annotations stop and shared bookmarks start... (all looks the same to me!)
Received on Wednesday, 16 August 2000 20:32:41 UTC