[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-usability
Subject:    Re: Easier Searching in KDE
From:       Jonathan Gardner <jgardner () jonathangardner ! net>
Date:       2004-06-04 20:10:35
Message-ID: 200406041310.35485.jgardner () jonathangardner ! net
[Download RAW message or body]

On Friday 04 June 2004 09:32 am, Jamethiel Knorth wrote:
>
> We don't need to use something like ALICE. The system would probably
> benefit more from a grammar parsing system. This is what GNOME Storage
> [1] uses, depending on a Head-Phrase Structure Grammar (HPSG) parser [2]
> and a set of Grammar Rules [3]. I would say that GNOME Storage already
> qualifies as a proof of concept, presuming it works as well as the page
> and its screenshots imply.
>

I think you're missing the core functionality of how ALICE works. It is a 
grammar parser. It takes what you write, does a whole lot of pattern 
matching and parsing, and then generates a response. ALICE is also 
stateful. It will remember what "it" is, for instance.

> Also, this isn't as useful in most search areas. It is great when
> everything on the system is well tracked in a single database, but
> doesn't work so well if the database is unreliable. This makes it great
> for querying the help system, the configuration system, the menu, or
> anything else that is entirely controlled. However, it is far less useful
> when just searching for files according to their names.
>

Yes. It would be unsuitable for searching anything that changes. It would 
probably only be limited to the help system.

In the future, perhaps we can discover a way to make it interact more with 
the system and be able to render more relevant responses on data that does 
change, however. I understand that they have ALICE bots that actually 
modify their database to match the person they converse with, and to learn 
new things, so we would probably want a feature like that.

Someone has tied ALICE into a speech recognition and speech synthesis system 
so that you can talk to it. They are also engineering a way to identify 
facial expressions and to generate appropriate expressions as well. Again, 
this is all future stuff, but the text-to-text static database is already 
well-done.

-- 
Jonathan Gardner
jgardner@jonathangardner.net
_______________________________________________
kde-usability mailing list
kde-usability@kde.org
https://mail.kde.org/mailman/listinfo/kde-usability
[prev in list] [next in list] [prev in thread] [next in thread]