Re: Number, Date, Time, Quantity

At 10:48 AM +0300 9/8/06, Henri Sivonen wrote:
>On Sep 7, 2006, at 13:09, Steven Pemberton wrote:
>
>>On Thu, 07 Sep 2006 11:15:30 +0200, Lachlan Hunt 
>><lachlan.hunt@lachy.id.au> wrote:
>>
>>>Let's see:
>>>
>>>    "The Prime Minister, Jan Peter Balkenende, will today travel to..."
>>>
>>>I think it's quite clear to a human reader that Jan is a Prime Minister.
>>
>>You're missing the point. The text we are talking about doesn't 
>>contain the name of the prime minister, and we are not authorised 
>>to change the text. However we are authorised to add metadata.
>
>Who is doing the unauthorization?
>
>>The advantage of machine-readable semantics are many. Searches 
>>become better for instance.
>
>But are the advantages more valuable than the cost of the human 
>effort to put the metadata there? Are the advantages reaped by the 
>same parties who bear the cost? What incentives do authors have to 
>meticulously mark up their documents?

Absolutely true.

Unless we take a life-cycle approach to getting better knowledge
about web content, it won't happen.  The syntax will be populated
with untruths, and nobody will pay attention to it.  So the authoring
tools that include preformed templates and mostly bother the author
for unclear cases will be essential to deploying any sort of metadata
or knowledge about the content widely.

On the other hand, the rationalized model of widget behavior
that we are looking for in PF is what you need to survive as a usable
interface in the mobile space.

And the content side of "Is this page structured well" is what will
get your page preferred in search reports.  So there are positive
paybacks for those who will go the extra furlong for "mindful markup."

>Also, it seems to me that the core competence and point of 
>differentiation of successful search engines is yielding useful 
>results even in the absence of such explicit metadata.

That was five years ago when page rank first hit the scene.

The algorithm has had to get more sophisticated, and page rank now
favors pages with a tight focus and well-marked-up headers for
internal structure. Or so an SEO consultant told me.

But you are right. Mass-market (i.e. money-making) search will use
the most endemic forms of metadata. That is to say markup. Wrapping a
header in an <hn> element as opposed to <span class="style239"> is
the kind of metadata that will win you a better place in the page-rank line.

>--
>Henri Sivonen
>hsivonen@iki.fi
>http://hsivonen.iki.fi/

Received on Friday, 8 September 2006 15:04:29 UTC