Re: VisibleMetadata

Le 28 mars 2007 à 09:34, Maciej Stachowiak a écrit :
> Evidence for this principle:
>
> - Search engines originally gave weight to "meta" keywords, but  
> have abandoned them because they are so often wrong and dishonest.

Untrue. It depends on the context.
* Some people are using the meta name keywords AND description in  
some application domains ON the Web, like Libraries and Museum. A  
long [list of projects][dc] using dublin core under different forms

dc: http://www.dublincore.org/projects/


Another example, you should know better ;)

* [SpotLight on Mac OS X][1]

Let's take a practical example [The Matrix of W3C Specifications][1]

In the Markup:

     <meta name="Keywords" content="qa, quality assurance,  
conformance, validity, test suite, qa, quality assurance,  
conformance, validity, test suite, matrix, qa, conformance,  
specification, recommendation, standard, test suite, validator,  
validation" />
     <meta name="Description" content="W3C QA - Matrix detailing the  
QA activity of the different specifications issued by W3C." />

In Spotlight:
I have a local copy of all HTML files, I edit on W3C Web site. Let's  
use the command line to list what has been indexed and then is  
accessible from the Spotlight GUI.

% mdls TheMatrix.html

kMDItemComment                 = "W3C QA - Matrix detailing the QA  
activity of the different specifications issued by W3C."
kMDItemKeywords                = (
     qa, "quality assurance",  conformance,
     validity, "test suite", qa, "quality assurance",
     conformance, validity, "test suite", matrix,
     qa, conformance, specification, recommendation,
     standard,  "test suite", validator, validation)
kMDItemKind                    = "HTML Document"

It is *actually* used and it is very practical when searching for  
information.

[1]: http://developer.apple.com/documentation/Carbon/Reference/ 
MetadataAttributesRef/Reference/CommonAttrs.html
[2]: http://www.w3.org/QA/TheMatrix.html


The problem you are explaining about metadata is about usability and  
management of information not that much about the language. I think  
the approach should be to not exclude the two ways of managing  
metadata. There are metadata that you want visible and some you do  
not want visible.

> Note that this principle is not an absolute, it says "more  
> effective". So it doesn't mean you can't have things like <link>,  
> but that it's better to combine the user-visible and machine- 
> readable information. Do you disagree that it is at all better? Do  
> you think it's worse? Should I throw the word "prefer" in there  
> somewhere?

Another counter example for "link" usage. :) link to CSS files.


It's not a design principles. It's more a content manager, author  
recommendation in some cases. The design principles should be  
something around:

The language must have the ability to express visible AND hidden  
metadata, (depending on the requirements of the authors.). For  
example [RDFa][3] can help authors to express visible metadata

Practical example:

<p class="cal:Vevent" about="#xtech_conference_talk">
   I'm giving
         <span property="cal:summary">
             a talk at the XTech Conference about web widgets
         </span>,
         on
         <span property="cal:dtstart" content="20070508T1000+0200">
             May 8th at 10am
         </span>.
     </p>



[3]: http://www.w3.org/TR/xhtml-rdfa-primer/

-- 
Karl Dubost - http://www.w3.org/People/karl/
W3C Conformance Manager, QA Activity Lead
   QA Weblog - http://www.w3.org/QA/
      *** Be Strict To Be Cool ***

Received on Wednesday, 28 March 2007 10:50:46 UTC