- From: Melvin Carvalho <melvincarvalho@gmail.com>
- Date: Thu, 19 Dec 2013 01:03:26 +0100
- To: Henry Story <henry.story@bblfish.net>
- Cc: Wes Turner <wes.turner@gmail.com>, "public-rdfjs@w3.org" <public-rdfjs@w3.org>, Luggen Michael <michael.luggen@bfh.ch>
- Message-ID: <CAKaEYhLhSUPnF7BMqkoGKfrkYwn_J3k1ZaZ-jyqp91Rg5ndMGA@mail.gmail.com>
On 19 December 2013 00:38, Henry Story <henry.story@bblfish.net> wrote: > > On 18 Dec 2013, at 23:59, Wes Turner <wes.turner@gmail.com> wrote: > > Again, I believe https://github.com/antoniogarrote/backbone does almost > exactly what you are describing in regards to creating rdf changesets from > user-supplied input. > > > Thanks Wes, I'll look at that. We were thinking of using Antonio's > backbone lib anyway. In fact he came > over to the "Weave the Web We Want" workshop we had at Mozilla in Paris > late last month, [1] > and we worked with it for a couple of days, but it's still very early > stages and there were a few bugs. > So we continued with rdflib for a bit which I had studied in more detail > over the past year. > > We probably don't yet need much SPARQL support. Pointed Graphs seem like a > very simple abstraction that can > do most of what we need. It also works very well with the notion of > Observables monad from the Rx JS library > as shown here: > > > https://github.com/stample/rww-play/blob/withPointedGraph/public/ldp/lib/pointedGraph.js#L68 > > This allows us to follow links and skip into different graphs, in an > asynchronous ways which can be very useful. > > I'll look at Antonio's code more in detail. If I can get to the point of > mastering it enough > so that I can fix the bugs myself then that will be definitevely something > I'll use. > > TimBL's rdflib is interesting initially because he has all the pieces for > fetching graphs, and it keeps > the HTTP metadata in the default store, using an ontology he developed for > this purpose. This > allows the the framework to know if it needs to update a remote graph when > representations are outdated, > and it would make it easy to do conditional GETs, PUTs etc... This could > easily be ported to > RDFStore I suppose, but I thought I'd first make sure I'd understand > TimBLs thinking in detail first. > Just FYI the core features of rdflib.js - Reads and writes RDF/xml, turtle and N3. - Read/Write Linked Data client, using WebDav or SPARQL/Update - Parses RDF/a. (rdflib-rdfa.js version only) - Local API for querying store - SPARQL queries (not full SPARQL) - Smushing of nodes from owl:sameAs, and owl:{f,inverseF}unctionProperty - Tracks provence of triples keeps metadata (in RDF) from HTTP accesses As well as a few other things like owl sameAs FP IFP, smushing, following links one deep, callbacks, tracking redirects, tracking the status codes of remote links, and indexing the store rdflib is quite modular but doesnt (yet) handle sparql parsing or is fully 1.1 compliant ... I wonder if modular common pieces could be reused ... > > Thanks again for the tip, > > Henry > > > [1] https://github.com/stample/wiki/wiki/Weave-the-web-we-want > > Wes Turner > > > Social Web Architect > http://bblfish.net/ > >
Received on Thursday, 19 December 2013 00:03:57 UTC