W3C home > Mailing lists > Public > ietf-http-wg@w3.org > July to September 2009

RE: [pubsubhubbub] Re: Input on request for link relation

From: Eran Hammer-Lahav <eran@hueniverse.com>
Date: Fri, 18 Sep 2009 11:07:46 -0700
To: "ietf-http-wg@w3.org" <ietf-http-wg@w3.org>
CC: Atom-Syntax Syntax <atom-syntax@imc.org>, "pubsubhubbub@googlegroups.com" <pubsubhubbub@googlegroups.com>
Message-ID: <90C41DD21FB7C64BB94121FBBC2E72343784D5812B@P3PW5EX1MB01.EX1.SECURESERVER.NET>
(This is not aimed at Martin personally)

> -----Original Message-----
> From: ietf-http-wg-request@w3.org [mailto:ietf-http-wg-request@w3.org]
> On Behalf Of Martin Atkins
> Sent: Friday, September 18, 2009 10:34 AM

> I think the main beef that folks have with the URI-shaped rels is that
> they're long and awkward. Why is all that "http://" and "/" noise in
> there? That's only useful if you want to dereference the URI.
> pubsubhubbub.org is enough information for an identifier.

Using URIs as a syntax for namespaces is well established practice. This argument is silly because these values, *especially* the one proposed by PSHB are meant for machine use only (and I am sure those machines don't share this feeling or feel like second class citizens). URIs provide a simple mechanism for namespacing the relation type value, avoiding the obvious problem of name 'land grab'. If you don't like using an http:// URI, you are free to use URNs (which I am sure others will find more awkward).

If this was a value we were presenting to people or asking them to input manually I would agree. This is why we are working on projects like WebFinger to improve the usability of web identifiers for people. But for engineers to make this argument about machine-only patterns, is really just plain silly.

If a short relation type is what makes someone feel better about their protocol, well, I don't think we can help.

EHL
Received on Friday, 18 September 2009 18:17:01 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 27 April 2012 06:51:10 GMT