>Over on comp.ai.nat-lang (a newsgroup about trying to use Artificial
>Intelligence to understand 'natural' languages, like English) I've been
>trying to stir up interest in an "outline of history" project, using
>something like XML...
>The general idea would be to treat all history-webpages as timelines--
>series of EVENT items, with each event having a DATE attribute, and
>normally a PLACE and various PERSONs...
>So something like XML (extensible markup language) could be used to
>'tag' timeline entries (or even _sentences_ within long prose
>descriptions of historical events).
> <EVENT
> DATE="1492.10.12"
> PERSON="Columbus.Christopher"
> PLACE="Caribbean.Bahamas"
> RELATIONSHIP="visited">
> 1492: 12Oct: Columbus 'discovers' America
> </EVENT>
>This will allow search-engine queries on dates (or range of dates) or
>persons or places, etc, with the search-engine easily returning all
>pages that fit the query...
>But the hardest thing will be those RELATIONSHIPs-- that attribute has
>to capture as much historical context as possible, but still use a fixed
>and unambiguous vocabulary.
>So a big priority is to sketch an _outline of history_ using limited
>vocabulary: starting with migrations, and why groups migrate, and what
>happens when they arrive-- relationships between neighboring groups,
>like conquest or assimilation or whatever...
>I have a very rough start at:
> http://www.robotwisdom.com/science/history.html
>...but the research required to take it the next step is overwhelming
>me!
