W3C home > Mailing lists > Public > public-rdf-wg@w3.org > January 2012

Re: mitigating cost of 303

From: Pierre-Antoine Champin <pierre-antoine.champin@liris.cnrs.fr>
Date: Wed, 04 Jan 2012 18:46:09 +0100
Message-ID: <4F049061.7080007@liris.cnrs.fr>
To: Sandro Hawke <sandro@w3.org>
CC: David Wood <david@3roundstones.com>, Steve Harris <steve.harris@garlik.com>, "public-rdf-wg@w3.org" <public-rdf-wg@w3.org>
On 01/04/2012 03:38 PM, Sandro Hawke wrote:
> I was thinking something more like json:
> 
>         [ { verb: "GET", from: "/r/(.*)", to: "/page/$1", code: 303 },
>           { verb: "GET", from: "(/data/.*/[^.])", to: "$1.rdf", code:
>         303 } ]
>         
> The second rule assumes everything under /data has a form with no suffix
> (no dots in the last segment) that redirects to the same name with .rdf
> appended.

Your example does not address the (quite common) case where the
redirection is content negociated -- e.g. dbpedia redirecting to from
resource/Lyon to page/Lyon if you want HTML, or data/Lyon if you want RDF.

I guess this could be adressed by something like

  [ { verb: "GET", from: "/r/(.*)", to: "/page/$1", code: 303 },
    { verb: "GET", from: "(/data/.*/[^.])", code: 303,
      accept: "application/rdf+xml", to: "$1.rdf" }
    { verb: "GET", from: "(/data/.*/[^.])", code: 303,
      accept: "text/html", to: "$1.html" }
  ]

but there is a delicate trade-off to find between simplicity and
usefulness...

  pa
Received on Wednesday, 4 January 2012 17:46:46 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 16:25:47 GMT