Re: [aspell-devel] soundslike lookup with affix compression

[Jose Da Silva]

Hi Suman,
...seeing that there was no replies yet....

If you track Aspell 0.60.{0..x}, you may note that most of the features are 
somewhat already decided and the trailing suffixes 0,1,2,3... is adding 
stability or bug fixes to the existing 0.60.x version.

As a starting point, my suggestion would be to probably begin looking at 
adding new features in Aspell 0.61. There is no harm in using perl to try 
out your ideas, but eventually, it hopefully gets converted into C or C++ 
and be able to be compiled into code.

In terms of your ideas, you may want to keep in mind some modularity since 
not all languages follow the same rules.

You mentioned using "sound-ex" and you may find some interesting information 
here http://www.sound-ex.com/alternative_zu_soundex.htm which mentions 
metaphone may be a better choice. Following other links on that website has 
several links to other websites, including this one here:
http://aspell.net/metaphone/
Please check out the links since it may answer some questions and bring up 
new ideas as well.

Other ideas and suggestions if you think this is interesting....
Many eyes make bugs more shallow, so if you have the interest, you may also 
want to look at the buglist to see if there is anything within your 
capability which you can identify and possibly whittle down: 
http://sourceforge.net:80/tracker/?group_id=245
If you identify something and can show a solution, great!

Here are also some additional resources too:
http://bugs.debian.org/src:aspell
https://launchpad.net/ubuntu/+source/aspell/+bugs
https://bugzilla.redhat.com/buglist.cgi?quicksearch=aspell
https://admin.fedoraproject.org/pkgdb/packages/name/aspell
http://aspell.darwinports.com/

I hope some of this is useful....




On July 17, 2009 09:46:50 am address@hidden wrote:
> Hi,
>
> This is in reference to the following excerpt from the  <<Things I
> would like to get done>> list at your page  http://aspell.net/
>
>> Support soundslike lookup with affix compression. I think it is
>> possible, although I don't know how effective it will be. The
>> basic idea is to affix compress the soundslike codes and then
>> match the codes up with affix compressed words. If you are
>> interested, email address@hidden, and I will explain it
>> in more detail.
>
> I need not flutter myself to say that I had kind of a similar idea in
> mind, and some bits more, in case what you have in mind above
> is the "Soundex algorithm". I shall be very much interested in
> these developments, and at least a perl version can be built
> pretty fast given that the ideas are fixed.
>
> Just one line about myself. I am a mathematician, well I used to be.
> I have worked as a statistician, and algorithm designer in the position
> of a scientist in the industry, and have a good reputation as  a
> programmer.
>
> Shall be looking forward to your response.
>
> Sincerely,
> Suman Bandyopadhyay.