Re: GNU Prolog FAQ

[Daniel Diaz]

Hi Vic,

Vic Bancroft wrote:
> 
> On Mon, 3 Dec 2001, Daniel Diaz wrote:
> 
> > Some other suggestions ???
> 
> Perhaps some automated techniques can be used to extract some FAQ . . .
> it might be possible to do some feature extraction and clustering.  Yet,
> after looking at users-prolog.mbox, any automated FAQ generation will
> first require stripping away some parts of the multi-part messages and
> many of the mail headers.
>
> Of course, it is more sci-fi than AI, but what if a gprolog-FAQ agent
> could actually identify a frequently asked question and could generate
> an appropriate automated reply.

Do you think to some Natural Language Processing and well indexed archive + FAQ 
to detect already
answered questions ? and then to automatically send a reply ? Maybe a simple 
"search engine" is
enough :-).

> 
> > From a practical point of view, I think it would be a good thing to
> > maintain this FAQ under the CVS of GNU-Prolog at sourceforge (either
> > under the module 'documentation' or inside a new module).
> 
> Here here, never underestimate the power of plain text!  It is so "diff"
> friendly as well.
 
plain text is in any case unavoidable! But it should be generated from the web 
site (or conversely
but only one source should be maintained obviously). I proposed in a previous 
mail the combo LaTeX +
HeVeA but it is not restrictive.

> > This would allow other person to help Christina to work on the FAQ.
> 
> I will be glad to help, particularly in areas where there is some
> difference between the ISO definitions and some pre-existing practices.
> I also have access to a vintage quintus, recent mercury under solaris
> current sicstus under a variety of platforms and a disposition to use
> c/c++/jni call interfaces.

Thank you Vic, I add you to our sourceforge site, please cooperate with 
Christian <address@hidden> on
this topic.

Do you think the FAQ should be a part of the documentation module ? In that 
case the combo
LaTeX+HeVeA would be better than other tools and we could add entries in the 
doc/Makefile (to
generate the plain text FAQ, the HTML versions,...). On the other hand if the 
FAQ becomes a *big*
part (with the agent above mentionned) it could be cumbersome in the 
documentation module and
justify an another module. What do you think ?

Also about the documentation, I really need help to improve it: proofreading 
(since english is not
(at all :-) my native language), examples for built-ins, SEE ALSO,... 
Volunteers are welcome :-)) 
NB: I have done a big work recently to be able to generate a Microsoft .chm 
usable under Win32, so
please wait for version 1.6 coming with 1.2.9 before starting something (I will 
update sourceforge
with 1.2.9 in few days).


>  -------------------------------------------------------------------
>  Victor Bancroft
>  Fellow, Artificial Intelligence Center               (706) 542-0358
>  Athens, Georgia  30602, U.S.A           http://ai.uga.edu/~bancroft
>  Principal Engineer, Zvolve Systems                  [v]770.551.4505
>  1050 Crown Pointe Pkwy, Suite 300, Atlanta GA 30338 [f]770.551.4509
> 


-- 
Daniel Diaz   address@hidden
http://pauillac.inria.fr/~diaz
gprolog --version 2>&1 | sed -n -e 's/By //p'