[uf-discuss] FYI: two posting about the Semantic Web, the "SynWeb",
scraping and microformats
David Janes -- BlogMatrix
davidjanes at blogmatrix.com
Wed Oct 26 03:26:23 PDT 2005
Tantek Çelik wrote:
>>and the web2.0  theoretically is about
>>sharing and remixing.
> Here is the irony: what you described as a desirable perspective, that
> is"namable objects" with "locations" are *exactly* what the so-called
> "half-assed" hCards are, e.g.:
> name: Ryan King
> location: http://theryanking.com/
>>we can -- equally or more
>>powerfully to anything we're doing now -- start remixing and pointing
>>around and avoid the copying.
> We're already there. Hence why the hCards in use there are as minimal as
> they are.
>>We can say "my train is 227 at leaves at
>>8:45 AM"  or "here's my vcard" .
== Part 1: finding data ==
Yes, users use other things than URIs to make the Internet work, and in
fact their day-to-day "mental" interaction with URIs is quite small
(such as when they see an ad for "dell.com", for example). I recognize
this: but behind the scenes, when the user types in "microformats" on
google or MSN or whatever, it's URIs that make that all work between the
I know you do not disagree with this point. The reason I'm stressing it
is URIs are one of the inventions that really make the Internet work, in
a way that, say, "ethernet cards" do not (we could have used token
rings, or something else). It's a computer/Internet technology the users
are comfortable having in the background.
In fact, I'm going to make a stronger statement for posterity: *the
"microformats view" of semantic data on the web will succeed because the
human part of web interaction is mostly solved*. We don't need a SPARQL
or whatever to query the web; we know how to get to the data we are using.
Beautifully, this is done with URIs but without the user really worrying
== Part 2: doing something with it ==
But once we're there, once we get to the page that has Ryan's train
schedule, how do we mash and combine it with other pages with
The human has done all the heavy lifting, the cognitive part up to this
part -- the stuff we're good at.
But now we want to do the automation part -- what trains could I take to
_Wallice and Gromit_ playing at the Cineplex? A pure mash up type of
question that might practically be answerable in a world of microformats
(as you and I believe). And by practical, I mean "people will use it".
I believe naming -- assigning URIs -- to individual microformat objects
will be required for that, so we can say "U1 is Ryan's vcard" rather
than "U2 contains Ryan's vcard". I don't believe this is asking for much
more work than what is already being done out there; I do understand
that this is not seen much in the wild yet, though in fact I have seen
vcards with individual 'id's assigned.
I'm done with this argument for now -- perhaps the next step for me is
to take the concept to the demo level, which of course may be a few
months away yet!
More information about the microformats-discuss