Re: Change Proposal for ISSUE-127

On 30.11.2010 01:04, Ian Hickson wrote:
> On Tue, 30 Nov 2010, Julian Reschke wrote:
>> On 30.11.2010 00:53, Ian Hickson wrote:
>>> On Sat, 20 Nov 2010, Julian Reschke wrote:
>>>>
>>>> Removal of an unused degree of freedom in defining link relations;
>>>> consistency with link relations in other contexts.
>>>
>>> It just removes one unused degree of freedom in favour of another (instead
>>> of<a>   and<area>   always being the same, it separates them out so in
>>> theory they could be different), while simultaneously making the spec
>>
>> Then let's just state that with respect to this,<a>  and<area>  always are
>> treated the same (thus<area>  wouldn't need to be mentioned).
>
> Why can't we just say that new link relations allowed on both<link>  and
> <a>/<area>  must not be defined to mean different things? That would be a
> far less invasive change than your CP suggests and yet would be equivalent
> to what you're suggesting would be sufficient for the equivalent problem
> with your proposal.

I don't think that the proposed change is "invasive"; it's mainly 
mechanical.

However, you don't consider the impact on other registries of link 
relations; may I remind you of 
<http://www.ietf.org/mail-archive/web/link-relations/current/msg00000.html>? 
That doesn't scale well for additional elements in HTML, or constraints 
in other formats.

>>> more complicated (it increases the verbosity of the table in the link
>>> types section).
>>
>> I'm not concerned with the complexity of the table, but with the
>> complexity of the feature it defines.
>
> Your proposal doesn't change any normative implementation or author
> requirements, so it obviously has zero effect on the feature's complexity.

First of all, it *does* change the complexity of HTML, until you add 
prose disallowing different meaning on <link> vs <a>. You haven't done 
that yet.

Second, I'm thinking of the complexity of link relation definitions in 
general; not only in HTML. For that, it's important to be clear that the 
semantics of a link relation should not depend on where it appears.

Best regards, Julian

Received on Tuesday, 30 November 2010 11:10:30 UTC