Re: Cluehunting: Expanding The Future

["Dan \"Effugas\" Kaminsky" <effugas best com>]

-----Original Message-----
From: Federico Mena Quintero <federico@nuclecu.unam.mx>
To: effugas@best.com <effugas@best.com>
Cc: gnome-gui-list@gnome.org <gnome-gui-list@gnome.org>
Date: Monday, August 17, 1998 2:59 PM
Subject: Re: Cluehunting: Expanding The Future

>>  First design, then program.
>
>I would advise you to do these simultaneously; it is the way the rest
>of GNOME has been written, and it works well.  It also helps avoid
>having over-engineered specifications and long documents that never
>get implemented.  It will also give you a good knowledge of the
>implementation issues of your proposal.


Except for the fact that it's already sucking up massive amounts of my time
just working through the design phases.  Over-engineered specs aren't good,
but underengineered specs are common.  One of the fascinating things about
the emails I've been recieving are the excellent references and observations
that would totally screw a "well I've already coded this so I'm not going to
change it even though it's better" perspective.

Cluehunting is not ready to be coded, but it's definitely ready to be
analyzed, deconstructed, and rebuilt.

>>  * isn't part of regex?  Interesting point was brought up, though...this
>>  isn't a great idea if it makes cluehunting much harder to implement.
>
>Shell globbing is trivial to implement in terms of regular
>expressions.  I suggest you read a book on finite automata.  I like
>this one:


Uh...:-)

Maybe I'll just replace RegEx with Shell Globbing...

>You may want to look at the Emacs Info documentation.  It explains all
>of these features.


Documentation sucks, that's why I'm hot on screenplays ;-)