W3C home > Mailing lists > Public > public-lod@w3.org > February 2017

Re: Browsers breaking content negotiation

From: Martynas Jusevičius <martynas@graphity.org>
Date: Fri, 10 Feb 2017 00:13:02 +0100
Message-ID: <CAE35Vmy5A2DBb5KWnZxv2hpvrmSjz7mvcfq9eAwSzU+4bdoZiQ@mail.gmail.com>
To: Jerven Bolleman <me@jerven.eu>
Cc: Ruben Verborgh <Ruben.Verborgh@ugent.be>, public-lod <public-lod@w3.org>, "public-rww@w3.org" <public-rww@w3.org>
So that separates the address space into URIs that serve Linked Data
(multiple representations using conneg) and URIs that serve "files" (a
single representation). Right?

In fact our implementation was doing that earlier (disabling RDF
conneg for files, only serving the original format), so now I went
ahead and reverted back to that.

But the browser behavior still does not make sense to me.

On Thu, Feb 9, 2017 at 9:32 PM, Jerven Bolleman <me@jerven.eu> wrote:
> I agree with Ruben.
> Representation specific IRIs are a requirement for a linked data/semweb server.
> 1 IRI for a concept can content negotiate to multiple representations
> that have their own IRI.
> Server side it makes sense to prefer sending the most logical
> representation of a concept if the client has no preference.
> Which for images is the highest quality image format, not their metadata.
> So practically for me I actually prefer if browsers send accept: */*
> it is better than what they did before e.g. chrome used to prefer xml
> over html.
> e.g. for http://purl.uniprot.org
> if the client has no preference we send HTML.
> else if the client wants RDF/XML we send that.
> else we send HTML because that's what we got ;)
> Regards,
> Jerven
> On Thu, Feb 9, 2017 at 9:04 PM, Ruben Verborgh <Ruben.Verborgh@ugent.be> wrote:
>>> image content negotiation breaks with Accept: */*, and we
>>> cannot control its default value in browsers.
>> That's indeed what https://wiki.whatwg.org/wiki/Why_not_conneg means.
>> Now the question is for what applications this poses a problem.
>> https://bugzilla.mozilla.org/show_bug.cgi?id=1249474 is an example,
>> but this is not a Semantic Web application.
>> Your original mail seemed to indicate a possibility
>> that it breaks Linked Data platforms, but I don't see that yet.
>> Best,
>> Ruben
> --
> Jerven Bolleman
> me@jerven.eu
Received on Thursday, 9 February 2017 23:13:36 UTC

This archive was generated by hypermail 2.3.1 : Thursday, 23 February 2017 16:49:10 UTC