Soundex implementation broken about text

[yorickpeterse]

Derp. Seems MySQL has an extra feature of sorts to return longer strings. In that case I would consider the second "bug" a "nice to have" instead.

[threedaymonk]

Re 1: This is as documented, though the documentation isn't very good and I agree that this behaviour is probably counter-intuitive.

I think it would make more sense if it processed only A-Z in the input, but that could be a breaking change for any applications that rely on the current behaviour (though I'd be surprised if there are any).

Re 2: As you've noted, MySQL is not standard. Soundex is supposed to generate 4-element codes. I'm not sure if that's really nice to have, or just buggy!

If I were to make any significant changes to Soundex, I think I'd want to add support for various dialects: Russell; Knuth; US government; Daitch-Mokotoff. On the other hand, I wonder how useful this is given that Metaphone gives better results for this application.

[yorickpeterse]

For me it's not super critical as I've since discovered the (double) metaphone implementation which worked better for my use case. Documentation wise, I completely missed that :/

[threedaymonk]

I'll close this, then, and make a note to improve the documentation in future.