Re: [Evolution] Ideas for Evolution

["tomm" <tomm pentawall com>]

I had a plan for a synonym search alogorithm to enhance the ability of naive users to find what they are 
looking for that might be useful for this...  The idea is that first all words of length shorter than x (four 
or five characters generally) would be ignored.  Then for each of the remaining words, they would be mapped 
to a common synonym- Thus doberman and rotweiler would map to dog.  THere would also be a map for unique 
words.

This would quickly elicit major categories and organizations of data, and additionally form their 
subcategories.  Plus, it would allow naive searchers to more frequently end up near their target category and 
sub category.  (As well as reduce the impact of poor indexing terms by content creators.)

I think that this idea could handily be adapted to the project at hand (I think it would also be nice for a 
distributed web crawler/indexer to be used with DMOZ.org...)

Tom M.
TomM pentstar com


---------- Original Message ----------------------------------
From: Giao Nguyen <grail cafebabe org>
Date: Thu, 1 Jun 2000 12:53:08 -0700

> On Thu, Jun 01, 2000 at 12:33:28PM -0700, Ray Lee wrote:
> > On the subject of rules (which isn't what Miguel started out on, but hey), I
> > think we could cover a lot of ground with an automatic rules generator. When
> > someone drags a message to a folder from their inbox, the rules wizard could
> > look through the message header, pull out the likely sort-upon fields
> > (sender, return-path, list-id, etc.) and pop up a dialog asking the user if
> > he wishes to create a recurrence rule for the action.
>
> While we're talking about automatic rules generator, I wonder if using
> something like term frequency index would work for rules suggestions. It
> would be neat where some chunk of code could determine the fact that
> there's an awful lot of mail from this particular list sitting in one
> particular folder and suggest that I use a rule for the folder.
>
> For TFI to work, we'd need a list of words that we can safely ignore. Like
> "the", "a", etc.
>
> Hey maybe we can use a little animated paper clip ...
>
> G.