xml (was Re: [Perlfect-search] data dump script)
Mon, 7 May 2001 13:45:11 +0100
>I believe what you're aiming for is a search engine which could be
>considered to a be part of "The Semantic Web". Have a look at...
>No need to reinvent this particular wheel. Just look up Dublin Core and
>RDF in the many sites that cover this field. You could start at the site
>I maintain, below. (It uses perlfect search of course.)
actually i'm being a lot lazier than that, but i take your point that
things are either done properly or by hacking about and one shouldn't
confuse the two.
i'm reasonably familiar with dublin core and newsml and various other
possibilities, but i'm looking for something much simpler here: i
want to be able to designate a small number of arbitrary fields and
get perlfect to use some as bodies to index and some as descriptors
to return. i don't want to do that using simple regexes, if possible,
because it negates too many of the advantages of using xml in the
first place, but neither do i want to prescribe a great deal of
infrastructure for what are really quite simple tasks.
one day i'll probably have to transpose a lot of data to an accepted
dtd (or schema, now?) for the sake of interoperability, but that day
seems rather far off and the people i work for - activist and
charity types - tend to have low-rent servers that dislike the
overhead of doing things properly. in which case an
adequately-well-behaved use of perlfect as a free-text adjunct to the
main delivery system would be ideal.
but if any of this is wrong-headed, i'd be very glad to hear it.