[topicmapmail] About getting topic from documents

[alex@shelter.nu]

Hi Morillo,

> I'm a little confused about creating topic maps, for example i have 20
> documents i want to extract topics from them but i donīt understand how
> to do this, what is the real method or process for getting topics,

As in any reply; it depends. What sort of documents have you got? What
format are they in?

> also i do not have a lot of time for reading this documents isnīt
> there a tool for making topic maps automatically?

There are a few scattered tools that can help you harvest topics from
various sources. I know Kal Ahmed has an open-source framework (Java)
that looks real good, although I haven't tried it. Ontopia has some
harvesting stuff with their OKS framework, but it is not free. Nikita
has some stuff on XSLT harvesting (http://www.cogx.com/kt2001/) There is
also quite some documents that explain the subject. Do this Google;
http://www.google.com/search?q=topic+maps+harvesting

I'd also try my own pet project xSiteable which can read delimited text
files (i.e. "save as text" from Excel), XML files (a simple format
most people can create from their own applications easy, i.e. a data-
base, an application, even Word, etc), text notation (for quick handling
of textual documents. No autodetection of topics, though). You'll find
more at http://xsiteable.org/ although unless it is urgent I'd wait a
few weeks until version 0.95 is out which actually documents a lot of the
things I talked about here. :) Or you could ask me, and I'll send you
the docs if you decide to go that way.

> well once i have obtained the topics, do i have to make topic maps
> manually?

Depends on the tools. Most often you guide the process and become the
editor of the data, but sometimes it is purely automatic. As always; it
depends.

> and finally what do i do withall the topic maps i have made with all
> this process?

Not sure what you mean, as most harvesters create XTM files which you
can use for whatever Topic Maps purpose you've got.

> How do i implement the topic map document (xtm) in a software aplication.
> thank for listening

Use a Topic Maps engine or create a serialiser that can read XTM to
whatever internal formats you're using. It would be much more helpful
for us if you specified what you want to do, what you've got, where
you want to go, etc.


Alexander