W3C home > Mailing lists > Public > www-tag@w3.org > December 2009

Re: On reading material for f2f - Web Linking

From: Jonathan Rees <jar@creativecommons.org>
Date: Sun, 6 Dec 2009 19:19:50 -0500
Message-ID: <760bcb2a0912061619h375f8007m6b9f9797e183c2bc@mail.gmail.com>
To: Eran Hammer-Lahav <eran@hueniverse.com>
Cc: Julian Reschke <julian.reschke@gmx.de>, Dan Brickley <danbri@danbri.org>, Tim Berners-Lee <timbl@w3.org>, TAG List <www-tag@w3.org>
I just noticed 'anchor' (sorry! it was there all along, since 2068).
The loss of rev at least doesn't mean loss of expressive power;
anything you could have written with 'rev' can be written using
'anchor'.  So I guess I won't miss it much.

Between 'anchor' and extension relations, Link: provides another
serialization for RDF, which is probably not a bad thing. Since Link:
is an entity-header, you could in principle say everything you want to
say about the resource in link headers (e.g. what its content is), and
omit the content entirely. This is kind of fun...

Link: <data:text/plain,mnot_is_winning>;rel="http://www.w3.org/2002/07/owl#sameAs"

Jonathan

On Sun, Dec 6, 2009 at 2:01 PM, Eran Hammer-Lahav <eran@hueniverse.com> wrote:
> It is being dropped/deprecated for the same reason the Link header was original dropped from HTTP: lack of implementation experience. As Julian noted, (and I think the number is around 90%,) the most common current use of rev is a typo (rev="stylesheet"). The -06 language should be fixed, but 'rev' has not proved to be worth the confusion it create.
>
> EHL
>
>> -----Original Message-----
>> From: www-tag-request@w3.org [mailto:www-tag-request@w3.org] On
>> Behalf Of Julian Reschke
>> Sent: Sunday, December 06, 2009 8:32 AM
>> To: Jonathan Rees
>> Cc: Dan Brickley; Tim Berners-Lee; TAG List
>> Subject: Re: On reading material for f2f - Web Linking
>>
>> Jonathan Rees wrote:
>> > Let's see... we've seen three variants of "rev"...
>> >
>> > 1. HTML 4 and RFC 2068 say the A link rev=B means B link rel=A.
>> >
>> > 2. The -06 and -07 drafts say that "rev" is deprecated but when used
>> > means the same as "rel". This is not just unenthusiastic; it's
>> > antisocial.
>>  > ...
>>
>> I think that's unintended effect of editing this part again and again.
>> This is a bug that needs to be fixed.
>>
>> > 3. The HTML5 draft (I consulted "Editor's Draft 6 December 2009")
>> > incompatibly prohibits use of "rev".
>> >
>> > I agree with you and Tim that the original HTML 4 / RFC 2068 version
>> > is the best of the three, since it eliminates the need to register the
>> > inverse of relationship foo, if foo is already registered. Obviously
>> > some other people have come to conclusion that it's better to double
>> > the number of relationships. I wonder why.
>> > ...
>>
>> I think the reasons were
>>
>> 1) evidence that link/@rev use in HTML documents frequently is
>> unintentionally (as a typo),
>>
>> 2) little correct use of @rev, and the assumption that defining more precise
>> relations is superior to have @rev (see mail thread around
>> <http://lists.w3.org/Archives/Public/ietf-http-wg/2008OctDec/0321.html>).
>>
>> Best regards, Julian
>
>
Received on Monday, 7 December 2009 00:20:24 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 12:48:18 GMT