On May 16, 2014, at 3:04 PM, Lewis John Mcgibbney <[email protected]> wrote:
> Hi Greg, > > Thanks for your email. > Please see my replies inline. > > On Fri, May 16, 2014 at 2:51 PM, Gregg Kellogg <[email protected]> wrote: > Do you ever expect to fully support RDFa 1.1? > > Yeah we do. Eventually someone will get around to it. If this is an itch you > have, we would very much encourage you to itch it and push it back into the > Any23 codebase ;) Not really a Java guy, so I don't think I can help here. > Look like it's just RDFa 1.0 with prefix support, even that fails many tests. > > Based on which version of Any23? Have you tried this new release or are you > just stating this based on your previous experience/assumptions? As I referenced, I used the http://any23.org/turtle endpoint. Perhaps you haven't updated that yet, the version at the bottom of the http://any23.org page says 0.9.1-SNAPSHOT from 2014-04-13, so it's reasonably recent. > Please refer to thhe following ticket for a new RDFa parser proposal which > can be considered as the flagship issue of the 1.0 release. > https://issues.apache.org/jira/browse/ANY23-137 Looks like there are several comments in there relating to this already. If the snapshot version doesn't include this, then that would explain it. > I think in the past, we've suggested using alternative Java RDFa > implementations; both Semargl [1] and clj-rdfa [2] are quite performant > (substantially faster than the any23 endpoint [3]) and fully conforms to the > spec. > > The end point you refer to has recently been overhauled and will be updated > with nightly SNAPSHOT's of stable Any23 code from now on. I've just jumped on > to leading this initiative so please bare with us until our scripts pull the > next build some time tonight. Excellent; once it's updated, you can either try pointing the RDFa test harness yourself, or feel free to contact me for support. > > Running the basic test suite[4] shows you passing only 58 out of 170 tests. > > Again, based on what version of Any23? > We are currently at 1.0. I would however really like to work with whoever > runs this service to improve the stats as much as possible... once I know > what the stats are actually quantifying and representing. The rdfa.info service is run by the RDFa community; the test suite there is used to determine conformance of processors claiming to support RDFa 1.0 or 1.1, which is where the EARL reports used to generate the overall conformance report come from. The any23 endpoint is configured in for informational purposes, along with that of most conforming processors. The overall implementation report for RDFa 1.1 can be found here: http://rdfa.info/earl-reports/ You can either run the service by cloning the rdfa.info website <https://github.com/rdfa/rdfa-website> and running it using the run-suite script, or most easily, simply enter the endpoint of your RDFa processor (such as http://any23.org/turtle) and it will be run automatically from the web page. This gives detailed conformance reports for each test across all host languages and RDFa variations. I'd be happy to work with you on this. Let me know if I can help further; it would be great if any23 becomes 100% conformant for RDFa. Gregg P.S., there are also some Microdata to RDF tests at <https://dvcs.w3.org/hg/htmldata/raw-file/default/microdata-rdf/tests/index.html> if you're interested. > Thanks > Lewis
