W3C home > Mailing lists > Public > public-lod@w3.org > October 2011

Re: Address Bar URI

From: Richard Cyganiak <richard@cyganiak.de>
Date: Mon, 17 Oct 2011 19:58:43 +0100
Cc: "Kingsley Idehen" <kidehen@openlinksw.com>, <public-lod@w3.org>
Message-Id: <96B786F8-3770-4178-9711-DA0F35682667@cyganiak.de>
References: <E338F3E7-C131-471A-AD5D-5B3882361AE9@ecs.soton.ac.uk> <EMEW3|36eea0fa457b931f9e7cd84ba7ed0101n9DD8902hg|ecs.soton.ac.uk|E338F3E7-C131-> <DDD289A9-060D-498A-9F83-68541B8DDF52@astro.gla.ac.uk> <3EC98AA0-13D5-4F49-B953-E00E945C4F10@ecs.soton.ac.uk> <EMEW3|3def92ab1d8be3210126f0dc285bc7c0n9DGMw02hg|ecs.soton.ac.uk|3EC98AA0-13D5-> <7A44633A0AA27A4A98B94B10BDF0AC3554C420@bbcxues27.national.core.bbc.co.uk> <D86820FF-B024-4CB5-9C19-43C36F973966@ecs.soton.ac.uk> <EMEW3|84acc2133619e6a53989a00825adf640n9EEiB02hg|ecs.soton.ac.uk|D86820FF-B024-> <7A44633A0AA27A4A98B94B10BDF0AC3554C426@bbcxues27.national.core.bbc.co.uk> <4E9ADEF3.8000603@openlinksw.com> <7A44633A0AA27A4A98B94B10BDF0AC3554C42A@bbcxues27.national.core.bbc.co.uk> <4E9C16A3.70409@openlinksw.com> <7A44633A0AA27A4A98B94B10BDF0AC3554C43D@bbcxues27.national.core.bbc.co.uk> <4E9C575F.7080000@openlinksw.co! m> <7A44633A0AA27A4A98B94B10BDFAC3554C43F@bbcxues27.national.core.bbc.co.uk>
To: Michael Smethurst <Michael.Smethurst@bbc.co.uk>
X-Mailer: Apple Mail (2.1084)
Return-Path: richard@cyganiak.de
X-OriginalArrivalTime: 17 Oct 2011 18:58:43.0232 (UTC) FILETIME=[CA685A00:01CC8CFE]

Hi Michael,

I take it your proposed design is:

- .html representations only link to other .html representations
- entity identifiers 303-redirect to .html/.ttl based on conneg

You say this is better because it avoids the extra 303 roundtrip that you'd get by pointing the HTML links straight at the entity identifier. Right?

What would you do in the .rdf representations? If you have a triple that links Alice to Bob using some property, would that triple connect the entity identifiers? (I take it that the answer has to be yes, since the triple connects Alice and Bob, and not the two RDF documents that describe them.)

I think a reason why so much LOD data uses the entity URIs (rather than redirect-avoiding .html URIs) in their .html representations is for symmetry. The RDF triples in the .rdf representations connect entity URIs. So it's kinda natural to have the HTML links in the .html representations do the same.

Now whether symmetry is a good reason for taking the 303 redirect hit, is another question…


On 17 Oct 2011, at 18:51, Michael Smethurst wrote:
> for the record the correct answers were:
> - because the conneg is broke
> - no
> - yes
> - because the links are broke
> - you wouldn't, the links are broke
> - yes
> - good lord, no, the links are broke
> > - why do the URblahs that end up in the address bar for dbpedia
> > contain /page/ or /data/?
> > - do you think the address bar should ever show .html when browsing
> > myexperiment.org?
> > - isn't the question of whether i want / get html or data dependent on
> > what I accept and not on the URblah I request?
> > - why do dbpedia links target /resource/?
> > - why would you ever use a href to point to something you can't GET?
> > - isn't that what 'about' is for?
> > - do you expect any sane publisher to take a 303 hit for every request?
Received on Monday, 17 October 2011 18:59:32 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 16:21:17 UTC