W3C home > Mailing lists > Public > public-html@w3.org > August 2013

Re: updated cite definition - please review

From: Charles McCathie Nevile <chaals@yandex-team.ru>
Date: Thu, 29 Aug 2013 11:28:37 +0200
To: public-html@w3.org, "Jukka K. Korpela" <jukka.k.korpela@kolumbus.fi>
Message-ID: <op.w2km9za7y3oazb@chaals.local>
On Tue, 27 Aug 2013 22:42:32 +0200, Jukka K. Korpela  
<jukka.k.korpela@kolumbus.fi> wrote:

> 2013-08-27 17:19, Steve Faulkner wrote:
>>
>> On 23 August 2013 12:52, Jukka K. Korpela <jukka.k.korpela@kolumbus.fi  
>> <mailto:jukka.k.korpela@kolumbus.fi>> wrote:
>>
>>
> [...]
>
>>     Changing things in a manner that has no effect on implementations
>>     seems rather futile. The only real effect would be added confusion
>>     among people who try to take semantic definitions seriously and
>>     continued debate over "semantics".
>>
>>
>> Not futile for authors who think that use of cite should be allowed for  
>> marking up in-text attributions, which the (revised proposal allows) as  
>> well as the title of works
>
> My previous question still has no answer: what impact would this have?

On browsers, probably none - the most likely is an extension that takes  
the cite value, does a search lookup, and tries to give a pointer to the  
source.

On search systems, or on well-curated content (documentation of closed or  
high-quality systems, etc) the same thing would happen in the back end.

And in real workflows, people would use cite as appropriate, rather than  
<i class="attrib"> or something.

> If it’s a matter of allowing the use of markup, as an end, it is best to  
> specify that <cite> may be used for anything. What could be more  
> permissive?
>
> It can hardly be useful to have an element that has such vague semantics  
> that it may mean a title of a work or a person or an organization.

Actually, it can be useful. Obviously it isn't *as* useful as a specific  
identifier, in finding some information. On the other hand, it turns out  
to match a lot of the way that real people think, and identify things.

Does "the Da Vinci picture of a naked man in a circle and square" identify  
anything for you? In my own life, I have found that such vague identifiers  
are often VERY useful - and dropping it into Yandex let me easily identify  
the title from the results page without going further. The first result  
was "Leonardo Da Vinci - Vitruvian Man", and almost every result  
explicitly mentioned the title in the snippet shown.

(I did the same with a bunch of random half-quotes plus author citations,  
and got useful results - even for fictional characters).

Search engines, and human brains, both allow us to make use of partially  
constrained (and partially misleading) data. Semantic elements are often  
useful, even when the semantics are not as precise as a URL (and realistic  
semantic web nuts can explain to you how the semantic web shows that the  
precision of a URL isn't actually all it's cracked up to be either).

Inventing the cite element today may or may not be a good idea (we'd love  
to see more use of the "article" element, but Steve's purely semantic  
"main", which has no mathematically precise definition, is actually  
incredibly useful). But overall it seems to work well enough to justify  
retention.

cheers

Chaals

-- 
Charles McCathie Nevile - Consultant (web standards) CTO Office, Yandex
       chaals@yandex-team.ru         Find more at http://yandex.com
Received on Thursday, 29 August 2013 09:29:12 UTC

This archive was generated by hypermail 2.3.1 : Thursday, 29 October 2015 10:16:34 UTC