Re: Discussion on ISSUE-79: meta-keywords

Julian Reschke, Thu, 25 Feb 2010 15:04:35 +0100:
> On 24.02.2010 17:13, Leif Halvard Silli wrote:

> Yes, the fact that HTML5 makes unregistered values invalid is a 
> problem that probably should be tracked separately (see related issue 
> <http://www.w3.org/html/wg/tracker/issues/102>).

Agreed.
 
> I'm not convinced that insisting on what HTML4 said is constructive 
> though. It's pretty clear that there is a de facto global namespace 
> of name values, and HTML5 needs to say something about it.

When HTML4 speaks about name="ROBOTS", then it makes no mention of 
@profile, but instead talks about how much support name="ROBOTS" has. 
So, it seems like convention/agreement is also seen as a valid way to 
specify name values, in addition to pointing to a specification via 
@profile. 

(In that regard: name="DC.identifier" is just as well known to HTML4 as 
name="Keywords" is. Name="DC.identifier" differs only in the way that 
it is owned/governed/versioned by its own specification organisation - 
which forbids HTML4 from going into details about how it works etc.)

>> But aren't there more meta@name values mentioned in HTML4 that ought to
>> be formally valid in HTML5? I found 3 values that are not present in
>> HTML5, but which HTML4 mentions:
>> 
>>  name="ROBOTS" content="ALL, INDEX, NOFOLLOW, NOINDEX"
>>                                       (section B.4.1 Search robots)
>>  name="copyright"  (section 7.4.4 Meta data
>>                     "hypothetical profile … for document indexing")
>>  name="date"  (section 7.4.4 Meta data)
>>                     "hypothetical profile … for document indexing")
>> 
>> All of which are mentioned in relationship to search engines and
>> indexing. I believe all of these are in use?
> 
> "Robots" is mentioned as "proposed" -- not "approved" -- in 
> <http://wiki.whatwg.org/wiki/MetaExtensions>.
> 
> Note the current requirements listed over there:
> 
> "For the "Status" section to be changed to "Accepted", the proposed 
> keyword must be defined by a W3C specification in the Candidate 
> Recommendation or Recommendation state. If it fails to go through 
> this process, it is "Unendorsed"."
> 
> So I would consider this as proof that the registry doesn't really 
> work, or that the registration requirements are too high (is anybody 
> going to write a W3C Rec defining the "robots"?) -- again, this is 
> related to <http://www.w3.org/html/wg/tracker/issues/102>.

According to HTML4, search engines was starting to support ROBOTS back 
then. And a search engine vendor later on introduced rel="nofollow". 
Clearly there is is a link between rel="nofollow" and <meta 
name="ROBOTS" content="NOFOLLOW">.  HTML4 also makes a point about 
saying (using name="DC.identifier" as an example) that <meta 
name="value" content="URL"> can also be written as <link rel="value" 
href="URL">. So the micro format rel="nofollow" is almost present in 
HTML4. 

Hence, if if rel="nofollow" was picked because <meta name="robots"  
content="nofollow"> was already known from HTML4 etc ... then it does 
indeed look as if name="ROBOTS" should have enough spec backing ...
-- 
leif halvard silli

Received on Thursday, 25 February 2010 16:48:29 UTC