Comments on: Splitting names http://www.onlineaspect.com/2009/08/17/splitting-names/ a blog about building stuff on the web Mon, 09 Jan 2012 18:55:04 +0000 hourly 1 http://wordpress.org/?v=3.3 By: Josh Fraser http://www.onlineaspect.com/2009/08/17/splitting-names/comment-page-1/#comment-1226 Josh Fraser Fri, 10 Sep 2010 22:25:08 +0000 http://www.onlineaspect.com/?p=431#comment-1226 Nice! Thanks for sharing. One thing I've realized is that proper parsing varies a lot on the context of where the names came from and how they are being used. For example, in my use-case, anything in parenthesis should be ignored -- in yours, it's a nickname. I guess, ideally we should write a class where people can change that behavior w/ a single variable to customize it for their own purposes. Let me know if you're interested. Perhaps we could combine forces to see what we could come up with. Nice! Thanks for sharing. One thing I've realized is that proper parsing varies a lot on the context of where the names came from and how they are being used. For example, in my use-case, anything in parenthesis should be ignored — in yours, it's a nickname. I guess, ideally we should write a class where people can change that behavior w/ a single variable to customize it for their own purposes. Let me know if you're interested. Perhaps we could combine forces to see what we could come up with.

]]> By: Jason Priem http://www.onlineaspect.com/2009/08/17/splitting-names/comment-page-1/#comment-1223 Jason Priem Tue, 07 Sep 2010 18:44:07 +0000 http://www.onlineaspect.com/?p=431#comment-1223 Hey Josh, nice work. I just finished writing something similar, along with a test suite of names. It does pretty much what yours does, although it's object-oriented and captures nicknames and first-initials separately. Here are a few names your lib misses that <a href="http://jasonpriem.com/human-name-parse/" rel="nofollow">HumanNameParser.php</a> parses correctly: George (gob) bluth // gets "gob" as a nickname (not part of first name) smith, john // reverses around the comma carlos garcia y luz // gets "garcia y luz" as a last name e.e. cummings // keeps original capitalization I like your idea of matching all middle names as part of the first name; that way you never miss names like 'Billie Jo'. However, I'd argue that this is less of a problem than always treating middle names as parts of first names, since it's far more common to have a single-word first name. My lib is at GitHub, and of course it's open, so take or fork anything you like. Hey Josh, nice work. I just finished writing something similar, along with a test suite of names. It does pretty much what yours does, although it's object-oriented and captures nicknames and first-initials separately. Here are a few names your lib misses that HumanNameParser.php parses correctly:
George (gob) bluth // gets "gob" as a nickname (not part of first name)
smith, john // reverses around the comma
carlos garcia y luz // gets "garcia y luz" as a last name
e.e. cummings // keeps original capitalization

I like your idea of matching all middle names as part of the first name; that way you never miss names like 'Billie Jo'. However, I'd argue that this is less of a problem than always treating middle names as parts of first names, since it's far more common to have a single-word first name. My lib is at GitHub, and of course it's open, so take or fork anything you like.

]]> By: Josh Fraser http://www.onlineaspect.com/2009/08/17/splitting-names/comment-page-1/#comment-640 Josh Fraser Tue, 18 Aug 2009 20:06:11 +0000 http://www.onlineaspect.com/?p=431#comment-640 Glad you found this useful and good catch on the parentheses issue. Perhaps you can merge in your code for handling last name, first name? That's definitely a common use-case that I missed. I've set up Google Code and given you commit access at <a href="http://code.google.com/p/php-name-parser/." target="_blank">http://code.google.com/p/php-name-parser/.</a> Glad you found this useful and good catch on the parentheses issue. Perhaps you can merge in your code for handling last name, first name? That's definitely a common use-case that I missed. I've set up Google Code and given you commit access at http://code.google.com/p/php-name-parser/.

]]> By: Pete Warden http://www.onlineaspect.com/2009/08/17/splitting-names/comment-page-1/#comment-638 Pete Warden Tue, 18 Aug 2009 15:16:39 +0000 http://www.onlineaspect.com/?p=431#comment-638 That rocks, thanks Josh! I have very similar problems, but nowhere near so comprehensive a solution. In my case I'm trying to canonicalize display names from email address headers. One common case is that the name will appear as "Warden, Pete" - I try to detect and flip those, but I'm guessing that's not an issue for your data set? Also there's sometimes multiple words inside the parentheses, eg "Pete Warden (Mailana Inc)", but from inspection it looks like you're only catching the first word with your parentheses check? I'd love to see this on Google Code, there's some other functionality I'm working on that might fit here, like gender guessing from first names: <a href="http://search.cpan.org/~edaly/Text-GenderFromName-0.32/GenderFromName.pm" target="_blank">http://search.cpan.org/~edaly/Text-GenderFromName...</a> That rocks, thanks Josh! I have very similar problems, but nowhere near so comprehensive a solution.

In my case I'm trying to canonicalize display names from email address headers. One common case is that the name will appear as "Warden, Pete" – I try to detect and flip those, but I'm guessing that's not an issue for your data set? Also there's sometimes multiple words inside the parentheses, eg "Pete Warden (Mailana Inc)", but from inspection it looks like you're only catching the first word with your parentheses check?

I'd love to see this on Google Code, there's some other functionality I'm working on that might fit here, like gender guessing from first names:
http://search.cpan.org/~edaly/Text-GenderFromName…

]]>