2010/11/1 Jorge Llambías
<jjllambias@gmail.com>
>> In deconstructing "cakrespa", it thinks it splits as "calku zei respa"
>> rather than the correct "ca kerfa zei spati".
>
> That's not a very good example, since the {ca} falls off and the lujvo is
> actually {krespa}, which is correctly split. I wouldn't expect jvozba to
> correctly handle a fake lujvo that was purposely intended to fool it.
I would. That's the whole point of having a self-segregating
morphology, isn't it?
Yes. The reason I wouldn't expect that, however, is because it isn't a grammar parser. It is intended solely for the purpose of creating or splitting lujvo. It is not intended to receive any input other than lujvo. That is like, to be overly simplistic, feeding "(x^8+7)(y^4-6)" to an adding machine. Feeding a machine input it isn't designed to process is bound to fail everytime.
It also thinks that li'ela'i is a lujvo, rather than two cmavo.