- From: Dan Brickley <danbri@danbri.org>
- Date: Tue, 17 Jun 2014 12:41:30 +0100
- To: Stian Soiland-Reyes <soiland-reyes@cs.manchester.ac.uk>
- Cc: Markus Lanthaler <markus.lanthaler@gmx.net>, Linked JSON <public-linked-json@w3.org>
On 17 June 2014 11:21, Dan Brickley <danbri@danbri.org> wrote: > On 17 June 2014 11:13, Stian Soiland-Reyes > <soiland-reyes@cs.manchester.ac.uk> wrote: >> Brilliant news! >> >> >> However.. >> >> >> Cache-Control: no-cache >> >> means the context will be re-requested for every parsing. (e.g. >> jsonld-java will normally cache contexts - >> https://github.com/jsonld-java/jsonld-java#controlling-network-traffic) >> I guess Google can handle the load, but it would mean slow-down for >> clients. >> >> >> Also the header >> >> Vary: Accept >> >> is missing (indicating that you could get a different representation >> if you have a different Accept header) - but given the above that is >> not a big issue.. > > Thanks. This release was bordering on the 'easter egg' level of > maturity, but I think we can fix it up pretty quickly. There's basic > caching in the (appengine) app so the context file is only computed > once, but you're right these details need attending to. OK, thanks Markus, Stian. I've just updated the live site, curl --verbose -H "Accept: application/ld+json" http://schema.org/ | head ...should show better headers now. > GET / HTTP/1.1 > User-Agent: curl/7.30.0 > Host: schema.org > Accept: application/ld+json > < HTTP/1.1 200 OK < Access-Control-Allow-Origin: * < Content-Type: application/ld+json < Cache-Control: public, max-age=43200 < Vary: Accept, Accept-Encoding < Date: Tue, 17 Jun 2014 11:24:30 GMT * Server Google Frontend is not blacklisted < Server: Google Frontend < Content-Length: 15957 < Alternate-Protocol: 80:quic Current code is here: https://github.com/rvguha/schemaorg/blob/master/api.py#L329 Notes on improving it: https://github.com/rvguha/schemaorg/wiki/JsonLd This thread is a fine place to continue discussions too. I'm tempted to look at a format like http://danbri.org/2013/SchemaD3/examples/4063550/hackathon-schema2.js which can be directly understood by D3.js - e.g. http://danbri.org/2013/SchemaD3/examples/4063550/hack3c.html ... but maybe the context file isn't the ideal place to put all those schema details? Is it considered better practice to have an informative, full-featured context file, or something small and compact for speedy access? cheers, Dan
Received on Tuesday, 17 June 2014 11:41:58 UTC