Re: still interested in indexing man/info/howto info?

[Aaron Turner <aturner@linuxkb.org>]

Hi Wade, sorry for the delay.

On Fri, 7 Jan 2000, Wade Hampton wrote:

> I have been a real deadbeat over the holidays -- kids, a wife,
> and several new computers do that to you....
> 
> I will start on trying to build an indexer for man/info/howto and
> other documents this weekend.  I have a couple of ideas:
> 
>     a) setup man2html, info2html, and links to /usr/doc and
> /usr/doc/howto/html
>        -- my normal setup for a development box
>     b) mirror the entire set into an html directory (so all man, info,
> etc.
>        are pre-converted)  
>        -- any idea on a www mirror tool I could use?

Mirror it?  I don't understand.

>     c) use glimpse or similar to index the mirrored tree
>        -- I've used glimpse before, but any other suggestions?

Htdig.

>     d) build a subject list (flat file, noun or noun and verb?)
>        -- ppp & setup, ppp & software, editor, IP masquerade & setup,
> etc.

Ouch.  Painful to get useful ones.

>     e) run glimpse or similar, reading the flat file, getting indexes
> from the
>        mirrored tree (if noun/verb, do an AND function?)

Uh huh.

>     d) manually selecting the relevant pages from the output of step 3
>        and outputing an index to another page (php would be great for
> this
>        as this could be WWW-based) (frame -- top are Y/N buttons, bottom
>        is the actual WWW page, select Y and it gets kept?)
>     e) another script to take the output from step d and populate 
>        the database (the real tie-in to the KB)
> 
> Any ideas?

Got me.  Realize that most of these man pages don't need to be so
completely categorized as the breadth of each man section is very limited.
I really think a UI for an engine that allows you easily to put content in
the database easily is far easier to develop than trying to build an
automated categorizor.  My .02.  I'm not convinced that you can give
Glimpse or whatever enough intelligence to do it without a great deal of
manual intervention anyways.

--
Aaron Turner, Core Developer       http://vodka.linuxkb.org/~aturner/
Linux Knowledge Base Organization  http://linuxkb.org/
Because world domination requires quality open documentation.
aka: aturner@vicinity.com, aturner@pobox.com, ion_beam_head@ashtech.net
The difference between `Unstable' and `Usable' is only two characters: NT