[uf-discuss] Microformat Base

[Craig Ogg]

On 12/1/05, Scott Reynen <scott at randomchaos.com> wrote:
> I thought I'd go ahead and play around with a microformat-based
> alternative to Google Base.  So far, I have a basic spider that I set
> loose from microformats.org to slowly wander the web.  When it finds
> any known microformat-associated class names, it records the data
> which can then be searched here:
>

This is very cool, but I don't think it is really an alternative to
Google Base.  As has been pointed out in some of the proposals for a
discovery format here, to have to spider a web site to discover its
data is not very efficient or accurate.  From some of the public
statements that Adam Bosworth of Google has made [1], I think Google
is trying to define a single universal schema for all data.  If you
take Google's upload formats (RSS, Atom, etc.) and combine it with
A9's Open Search you end up with a way to query any web site using
REST for structured data about what it contains. I talk about this on
my blog in more detail [2][3])

While some elements are predefined in the schema, it looks like Google
Base is depending on user-defined attributes converging over time for
specific domains (similar to the tags vs categories benefit).  It
appears to offer to refine your search results on attributes it
discovers are shared by a significant number of items in the initial
search.  This allows new attributes to bubble up as they become
popular.

I think microformats offer much more potential to aid adhoc discovery
and use of information while you are browsing:  drag this event to my
calendar, add this person as a contact in my address book, give me
driving directions to this location, give this blog post proper via
credit, etc.  Having this built-in to Firefox or Flock I think would
be very cool.

Craig

P.S. I realize that rel-tag is being used to aid search already -- but
I think it is being almost exclusively consumed from RSS feeds. 
Probably for the efficiency reasons stated above.

[1] http://www.itconversations.com/shows/detail571.html
[2] http://www.softwarevoices.com/archives/20-Democratizing-Information-Speculation-on-the-Future-of-Google-Base.html
(or http://googlebase2.notlong.com)
http://www.softwarevoices.com/archives/17-Did-Adam-Bosworth-reveal-the-real-Google-Base-at-the-MySQL-Users-Conference.html
(or http://googlebase1.notlong.com)