Re: [aspell-devel] Word lists that use Affix

[Kevin Atkinson]

On Mon, 6 Dec 2004, Gary Setter wrote:

> <snip>
> Hi Kevin, Jose,
> 
> Again, thanks for pointing out the compression utilities I
> already had. I ended up using munch.
> Munch expects the number of words on the first line. It also put
> the number of items after compression on the first line. Since
> aspell does not expect the word count, it got in the way. I
> modified munch to take a word count as a command line argument.
> Do you think such as change is worthy of a patch?

No.  As that program is going to disappear since the Aspell utility 
will have a munch-list command starting with Aspell 0.60.2.

> Also, making the compressed list was just one step in making the
> minimum dictionary that fit my vocabulary.
> The steps included,
>  making a representative sample,
>  searching the word lists,
>  affix compression,
>  combining lists,
>  compiling into rws,
>  making the multi file.
> 
> I'm not sure about the combining lists step. Looking at the code,
> it seemed to me that making one rws file would be more efficent.
> But I'm not so sure. Any ideas on this?

Yes the fewer compiled dictionaries the better.

> Should there be an application for managing the process for the
> user?

Maybe.

-- 
http://kevin.atkinson.dhs.org