Ray, That sounds interesting. I can see why you would need a new approach to Maori since the 1st step in Soundex is to collapse or remove the vowels altogether. In a vowel heavy language this would be self defeating. Is there a degree of intonation required to speak Maori? something else soundex doesn't consider. How do you handle the extra vowels during segmentation/syllabification. What is the application for your algorithm?
I'm trying to do some reading about Standard Arabic (SA). Thankfully there's been some recent scholarly work done on this language. They discuss algorithms and the additional/different rules for SA. I'm not sure yet how easily I'll be able to convert their work into code. The approach I'm leaning toward is to add 1 or 2 newer algorithms to the code base, in addition to soundex (Hopefully one of the 3 will handle the name well) then Build a separate index using each of these. Do a selectindex on each, make a Union list, then present a match list to the User for human analysis. Since U2 is now international, and with the advent of NLS, maybe IBM could provide something more than the Basic SOUNDEX function. This would give them a selling edge for getting U2 embedded in public safety apps. Thanks, -Baker -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Ray Wurlod Sent: Monday, January 14, 2008 11:11 PM To: u2-users@listserver.u2ug.org Subject: Re: [U2] Phonetic Name Algorithms I have one that is suited to Pacific Islander languages, such as Maori, which tend to be vowel-heavy. ------- u2-users mailing list u2-users@listserver.u2ug.org To unsubscribe please visit http://listserver.u2ug.org/
