- From: Thomas Baker <tbaker@tbaker.de>
- Date: Thu, 8 Dec 2005 07:26:32 +0100
- To: "Miles, AJ (Alistair)" <A.J.Miles@rl.ac.uk>
- Cc: Dan Brickley <danbri@w3.org>, public-swbp-wg@w3.org
Al, On Wed, Dec 07, 2005 at 05:02:57PM -0000, Alistair Miles wrote: > > > http://www.w3.org/2001/sw/BestPractices/VM/http-examples/2005-11-18/ > > > > > > - Default content type? The 'requirements' section > > currently says that serving RDF/XML content should be the > > default, where content-negotiation is possible. Danbri would > > like a config that allows the administrator to decide which > > is the default. > > > > I plan to serve HTML by default until someone persuades me otherwise. As I understand it, is FOAF currently served at that URI in a form that is both machine-friendly and human-friendly (HTML with RDF in-line). [1] Are you simply saying that you will continue to serve RDF-enhanced HTML by default, or are you suggesting that administrators be allowed to decide to set HTML as the default even if the HTML is not enhanced with RDF? > > Hmm I guess all browsers send 'accept: text/html' don't they? > > > > I want to ensure that http://xmlns.com/foaf/0.1/ remains reliably > > browsable... > > Afaik all browsers send an 'Accept:' header in their HTTP GET requests. > > Afaik no RDF clients or client libraries (e.g. Jena) send an 'Accept:' header in their HTTP GET requests. I suggest that this important point be clarified, perhaps even in the requirements section itself. If the default behavior of RDF clients were to change, then the requirements with regard to default behavior could also change (i.e. be relaxed). Tom [1] http://danbri.org/words/2005/10/25/142 > Setting 'application/rdf+xml' as the 'default' means > returning RDF/XML content in the response, when no 'Accept:' > header has been specified in the request. This configuration > means we are backwards-compatible with current RDF toolkits, > although we should probably still recommend that RDF toolkits > include an 'Accept: application/rdf+xml' header with GET > requests. > > Because all (?) browsers already include an appropriate > 'Accept:' header in HTTP GET requests, the URIs should still > remain 'reliably browsable'. -- Dr. Thomas Baker baker@sub.uni-goettingen.de SUB - Goettingen State +49-551-39-3883 and University Library +49-30-8109-9027 Papendiek 14, 37073 Göttingen
Received on Thursday, 8 December 2005 06:29:29 UTC