libby: based on jena 2 rules, sparql syntax, comments on jena-dev among other things
libby: rulesets can have names, inclusions, prefixes, list of rules, more rules to run
libby: rules themseves are similar, names, comments, prefixes (affecting scoping), terms=> terms and also the reverse - sparql contruct with terms
libby: term can have subjects, objects, variables, termlists, numbers, strings, functions
libby: and the triple patterns are like sparql
libby: and pseudo-terms: filters - which are like sparql filters (acceptance predicates); 'term negation' (unless given, derived, already [both]); aggregation (all)
libby: that's the plan. rules processing will be a flow of variable bindings; aiming for scalability
libby: stratification and parallelisation and specialisation
libby: plan to spray the processing components widely and wisely over a porcessor network and sitributed rdf dbs; group the joins in big triple patterns carefully; running filters early; save state where expensive
libby: have a preminiary implentation; currently doing perf testing against jena 2 rules
libby: rulesets can have names, inclusions, prefixes, list of rules, more rules to run
libby: rules themseves are similar, names, comments, prefixes (affecting scoping), terms=> terms and also the reverse - sparql contruct with terms
libby: term can have subjects, objects, variables, termlists, numbers, strings, functions
libby: and the triple patterns are like sparql
libby: and pseudo-terms: filters - which are like sparql filters (acceptance predicates); 'term negation' (unless given, derived, already [both]); aggregation (all)
libby: that's the plan. rules processing will be a flow of variable bindings; aiming for scalability
libby: stratification and parallelisation and specialisation
libby: plan to spray the processing components widely and wisely over a porcessor network and sitributed rdf dbs; group the joins in big triple patterns carefully; running filters early; save state where expensive
libby: have a preminiary implentation; currently doing perf testing against jena 2 rules
libby: restful RDF user annotation system for three projects; public and private annotations possible; restful service for submitting and receiving annotations; work in progress
libby: examplea nnotation at the end of this page
libby: uses JAX-RS (Jersey reference implementation) a java api for restful web services, Jena (naturally!), and spring app framework (dependency injector)
libby: people have public and private buckets (named graphs some with access control); readable uris for annotations; uses openid; users annotea schema
libby: POST->authentication->authorisation->validation->return url; GET->authentication->CONTRUCT query (blanking out the unauthorised bits)-> retern json etc etc formats
libby: could use a sparql filter but would have annoying caching effects; post filtering is quite nice and easy and understandable
libby: uses profiles of allowed annotation types as config file (bit like applicaton profiles) and while you can throw anything at it, you get a 400 bad request if it doesn't validate
libby: issues: group restrictions? could have group buckets perhaps, making gatekeeper more complex. audit trails? cross domain restrictions. caboto is quite small but dependencies are large
libby: code is here
libby: examplea nnotation at the end of this page
libby: uses JAX-RS (Jersey reference implementation) a java api for restful web services, Jena (naturally!), and spring app framework (dependency injector)
libby: people have public and private buckets (named graphs some with access control); readable uris for annotations; uses openid; users annotea schema
libby: POST->authentication->authorisation->validation->return url; GET->authentication->CONTRUCT query (blanking out the unauthorised bits)-> retern json etc etc formats
libby: could use a sparql filter but would have annoying caching effects; post filtering is quite nice and easy and understandable
libby: uses profiles of allowed annotation types as config file (bit like applicaton profiles) and while you can throw anything at it, you get a 400 bad request if it doesn't validate
libby: issues: group restrictions? could have group buckets perhaps, making gatekeeper more complex. audit trails? cross domain restrictions. caboto is quite small but dependencies are large
libby: code is here
libby: "RDF Plugins for jQuery"
libby: what's the benefit of creating rdf? costs can be quite high
libby: rdfa: low cost, data is there anyway
libby: demo of js that's pulling out the info from the rdfa and making a list of the data in a sidepanel (also a bit of reasoning) - benefit - different views of the data to the user
libby: using plugin to jquery
libby: parsing rdfa, storing (multiple stores), querying sparql-esqely over multiple sources
libby: re-adding triples created from a query back into databanks so can do simple reasoning that way; autoinserting rdfa into the page (blimey!)
libby: code to play with
AndyS: MIT License for code
libby: what's the benefit of creating rdf? costs can be quite high
libby: rdfa: low cost, data is there anyway
libby: demo of js that's pulling out the info from the rdfa and making a list of the data in a sidepanel (also a bit of reasoning) - benefit - different views of the data to the user
libby: using plugin to jquery
libby: parsing rdfa, storing (multiple stores), querying sparql-esqely over multiple sources
libby: re-adding triples created from a query back into databanks so can do simple reasoning that way; autoinserting rdfa into the page (blimey!)
libby: code to play with
AndyS: MIT License for code
libby: "We?ve built a really simple, rather addictive system that lets anyone with a few spare minutes match up a randomly-selected speech from Hansard against the correct snippet of video. You just listen out for a certain speech, and when you hear it you hit the big red ?now? button. Your clip will then immediately go live on TheyWorkForYou next to the relevent speech, improving the site for everyone. Yay!"
libby: sorta vaguely related to the crew work, thoug this is really about transcription
libby: sorta vaguely related to the crew work, thoug this is really about transcription
libby: Leigh Dodds, Talis
libby: high level talk about where we are now and where Talis platform fits into the landscape
libby: talis platform
libby: leigh announces free hosting on the talis platform for open datasets! (has to have an open data license)
libby: high level talk about where we are now and where Talis platform fits into the landscape
libby: talis platform
libby: leigh announces free hosting on the talis platform for open datasets! (has to have an open data license)
libby: didn't catch names of presenters sorry :(
libby: "This talk covers the alleviation of challenges associated with conceptual level data access, across heterogeneous data sources; using Virtuoso's RDF Views, Data Virtualization, Native RDF storage, and Linked Data deployment technology."
AndyS: Orri Erling, Yrjänä Rankka
libby: demo site
libby: OpenLink Software's Virtuoso Submission to the Billion Triples Challenge
libby: TPC-H Virtuoso demo
libby: OpenLink Billion Triple Demo queries
libby: "This talk covers the alleviation of challenges associated with conceptual level data access, across heterogeneous data sources; using Virtuoso's RDF Views, Data Virtualization, Native RDF storage, and Linked Data deployment technology."
AndyS: Orri Erling, Yrjänä Rankka
libby: demo site
libby: OpenLink Software's Virtuoso Submission to the Billion Triples Challenge
libby: TPC-H Virtuoso demo
libby: OpenLink Billion Triple Demo queries
libby: Graham Klyne is standing in for Alistair
libby: "Experiences from the FlyWeb Project"
libby: demo application - find info about a gene
libby: 3 databases, pure ajax app, all-sparql everywhere; want to provide data sources which other apps can build on over time
libby: The Image Bioinformatics Research Group
libby: "Experiences from the FlyWeb Project"
libby: demo application - find info about a gene
libby: 3 databases, pure ajax app, all-sparql everywhere; want to provide data sources which other apps can build on over time
libby: The Image Bioinformatics Research Group
libby: first talk - by Laurian Gridinoc, Knowledge Media Institute, The Open University
libby: it's a firefox plugin that looks up related ontologies from a given webpage paragraph, I think; but he's saying the proble is the lack of related instance data for the ontologies (I think! lemmee know if I misunderstood)
libby: it's a firefox plugin that looks up related ontologies from a given webpage paragraph, I think; but he's saying the proble is the lack of related instance data for the ontologies (I think! lemmee know if I misunderstood)
libby: it's all being videoed for a semweb video annotation project
libby: <MikeJ1971>libby - the website the recording will appear http://crew.rcs.manchester.ac.uk
libby: the blog website is http://crew-vre.net
libby: annotations will appear here (network issues mean they are not immediate today)
libby: andyS would like feedback on what liked / what didn;t etc in the day - please send him your slides
libby: <MikeJ1971>libby - the website the recording will appear http://crew.rcs.manchester.ac.uk
libby: the blog website is http://crew-vre.net
libby: annotations will appear here (network issues mean they are not immediate today)
libby: andyS would like feedback on what liked / what didn;t etc in the day - please send him your slides
danja: Proposed discussion point for the SWIG F2F, if a slot is available.
danja: Assume a person has a resource about themselves - a regular HTML page plus RDF via conneg. How close should these two representations match (given each format's expressive advantages)? In an extreme case consider having a photo as one representation of the resource.
danja: What of the quality factor in the HTTP messages?
danja: What if the HTML version has embedded RDFa/microformats/GRDDLable data - should the derived RDF be identical to that of the RDF/XML representation?
danja: Does (has) this point need to go past TAG? Do we have enough consensus to justify a best practices Note on the issue?
danja: My personal feeling is we need to be flexible, that the representations may be very different, as long as appropriate linkage is in place between them (I'm currently making just such a setup at http://danny.ayers.name editing manually over webdav).
danja: Supplementary question: how best to express rdfs:seeAlsos as HTML <link rel="..." />?
tobyink: rdfs:seeAlso can be expressed in HTML using RDFa as <link rel="rdfs:seeAlso" ...> given appropriate namespace declarations.
tobyink: Also <link rel="meta" ...> has similar semantics.
danja: Assume a person has a resource about themselves - a regular HTML page plus RDF via conneg. How close should these two representations match (given each format's expressive advantages)? In an extreme case consider having a photo as one representation of the resource.
danja: What of the quality factor in the HTTP messages?
danja: What if the HTML version has embedded RDFa/microformats/GRDDLable data - should the derived RDF be identical to that of the RDF/XML representation?
danja: Does (has) this point need to go past TAG? Do we have enough consensus to justify a best practices Note on the issue?
danja: My personal feeling is we need to be flexible, that the representations may be very different, as long as appropriate linkage is in place between them (I'm currently making just such a setup at http://danny.ayers.name editing manually over webdav).
danja: Supplementary question: how best to express rdfs:seeAlsos as HTML <link rel="..." />?
tobyink: rdfs:seeAlso can be expressed in HTML using RDFa as <link rel="rdfs:seeAlso" ...> given appropriate namespace declarations.
tobyink: Also <link rel="meta" ...> has similar semantics.
libby: Graham Higgins on why foaf firstname / lastname won't work for putting mysociety parliamentary data in rdf
libby: Damian Steer on oopsies and gotyas in rdf: alt, bag etc are wrong; so is reification almost always; id too; being able to serislise all graphs would be nice...and many more
libby: (Zach Beauvais?) talking about semweb and 'the cloud' as distinct from cloud computing - data view of the cloud - data as a long term carrier of value
danbri: Ah, I was just wondering what Graham was up to, as I mailed mysociety list a few minutes ago. I should talk to him!
libby: Ben O'Steen(?) Oxford University Research Archive (ORA) - rdf under the hood - see http://oxfordrepo.blogspot.com advice welcomed
libby: Tom Morris: danbri asked for a flawless rdf toolkit for ruby soon. we're getting close to it - Reddy - would like help! ruby rdf mailing list
danbri: I didn't say flawless, just better and evolving, ... :)
libby: Tom Heath: what the heck is vocamp vocamp wiki - want to grow the cloud of linked data - we need vocabs - people sometimes stuffe evrything into rdfs. vocab is hackfest informal, spend time together creating vocabs and getting published and associated meta issues about how to create them well. Oxford sept, 2 weeks next is in galway, more in future - come along or get in touch and run your own
libby: Andy Seaborne - Paths in sparql detail - examples: recursive class finding; xpath-type thing for rdf (including backwards links); this stuff can be quite hard
libby: Daniel Lewis - various updates + autocomparison of religions using data mining and machine learning
libby: Tony Hammond - science publishing - XMP from adobe (pdf) - rdf - can be embedded in any arbitrary binary file - we want to use this for descriptive metadata - basically same data as in rss feeds - DOI / author / title / desc
libby: Rob Styles - work with Jeni over lunch on a readlist app :-)
libby: Chris Wallace - xquery based picture browser over dbpedia - this link I think
libby: please improve the descriptions of those last I was having network crapola
Shepard: Damian's point: just like in Sun's Java APIs things never get removed but deprecated we should deprecate various RDF features so newcomers are aware of what to avoid