Re: [xml-dev] XML for Video, Pizza Shops & TakeOut

[team refind <netography@refind.org>]

>on 10/19/01 1:30 PM, Fuchs, Matthew at matthew.fuchs@commerceone.com wrote:

>As Bob Glushko likes to say, the perfect is the enemy of the good.  Let's
>_first_ get the tagged metadata up and make that profitable, because then
>someone will determine if it's worth the additional effort to do the
>transl(iter)ation.  Cool doesn't cut it unless there's eventually some ROI.
>The higher you raise the bar for an initial implementation, the less likely you
>are to see it happen - and also the more likely that the result was
>undesireable. Start small and evolve.

we did exactly what you say and then some---albeit off radar for the last 18
months. The framework (ie simplistic DTD for an all encompassing
classification infrastruture) has gone through 4 complete rewrites to the
point where over the last six months, we have not encountered anything that
breaks the framework.

95% of the battle is dividing the web into just 5 broad contexts so that
query agents have a faster time to know where to "grep" and present you
better answers on a tiny page.

such a framework has to know what's it's ultimate goal is at the onset.
the xml itself is anti-climatic. the dtd to make this happen will be
published by the first week of november.

due to budgetary and time constraints, no time to write the
 web interface to our running language independant search engine prototype
that makes use of /refind.xml till about february.  Anything you see
discussed (or left intentionally vague) on our websites has gone through a
few generations through natural evolution with prototypes.

sorry if this is off topic to the list, but even in a perfect world where a
few billion web pages get replaced with custom xml dtd's, rdf, dam+oil,etc..
there is no mechanism enabling standardized queries. the curse of having to
map thousands of tags from individual dtd's into contexts.
the refind framework hopes to build an open source dictionary where for each
word/phrase people use as a query term, corresponds to a branch (ie Xpath)
to the tree of life. That way, agents can execute more quickly in the proper
places (instead of brute force AI magic/grep) with more and more information
out on the web uses simple xml mechanisms.

whenever i ask a question, i want the power of all available information
around me to present. it would be nice for pin point precise answers but
that won't happen for a long time, so i'd rather at least have the list of
places that i have to call or visit to get closer to an answer compared to
not knowing where to start.

the existing web can barely take baby steps.
semantic web expects us all to be olympic sprinters.
standardized queries for resource discovery is the missing link which allows
us to stand upright and start walking so that all of society benefits from
information technology.

What a tacky analogy on a friday afternoon. time to go outside, enjoy the
sunshine.


--dz 
team refind