- From: Yves Lafon <ylafon@w3.org>
- Date: Tue, 1 Feb 2011 09:50:02 -0500 (EST)
- To: "Eric J. Bowman" <eric@bisonsystems.net>
- cc: nathan@webr3.org, ashok.malhotra@oracle.com, Jonathan Rees <jar@creativecommons.org>, Larry Masinter <masinter@adobe.com>, Noah Mendelsohn <nrm@arcanedomain.com>, "www-tag@w3.org" <www-tag@w3.org>
On Mon, 31 Jan 2011, Eric J. Bowman wrote:
> Nathan wrote:
>>
>> Aye, and I guess classing some URIs as "media types" based on the
>> first x chars of the lexical form of the URI would not be a good idea
>> (Opacity and all).
>>
>
> It's a fine debate to have on rest-discuss, where we can talk about the
> various philosophies of how self-descriptive messaging might function,
> without limiting ourselves to the constraints imposed by HTTP. The
> HTTP WG would be the right forum to suggest, in the absence of a
> Content-Type header, falling back to some other header which allows
> URIs as tokens -- without having anything to do with media types. Here,
> though, we should treat decisions like the registry being targeted at
> humans or not being URI-extensible (or existing at all), as having
> already been made, IMO.
The current fallback is sniffing, not another header. Adding a new header
won't solve the issues outlined by Larry's document.
That said, minting URIs to query a registry might be helpful.
--
Baroula que barouleras, au tiéu toujou t'entourneras.
~~Yves
Received on Tuesday, 1 February 2011 14:50:10 UTC