I agree that text animators are somewhat limiting, especially when you want an effect to last for a certain amount of time (say 0.5s for any displacement). You need to do layer per syllable for that, just like you need to with .ass karaoke.
For not too fancy particle stuff, there isnt any need to write a script to do that all for you. Just copy all the text layers, place them in a precomp, use it as emitter for one (yes one) solid with Trapcode particular. Fancier particle stuff can be done with layer per syllable (but i already described all this in a post in the karaoke help thread).
Though i really like to see how your script will look like (with fancy GUI and stuff), i guess the english fansub world isn't really interested. You should try contacting some brazilian fansubbers, as they seem to be the ones that are actually interested in using AE for karaoke (like making a hand write the syllables as they're being sung
Must be something with path's and stuff.. haven't really tried). The english fansub world seems to care more about using .ass karaoke, which can be over-the-top and flashy too with aegisub's automation. (everybody knows that a quality fansub is all about flashy over-the-top karaoke
)
So I don't really agree with the people above me , AE karaoke
can be very nice and fitting to the song, in such a way that it actually makes the song better (but that goes for every karaoke), and a good script can help with that. It's not as if AE karaoke is always fancy and over-the-top. I just like working with AE, being able to preview and easily change values by hand.