Unicode and chunk expressiong

[pkc]

Dar wrote:

>  However, that does have its 
>own problems.  You end up with pieces of elements you don't expect.  I 
>ran a quick test with Tagalog and getting the last word got part of a 
>font element and the trailing </p>.  Of course, since it is all ASCII, 
>you can program around that.


That's quite right. In my experience so far, these trailing and fronting elements all occur very regularly and can be filtered out as one wishes. A line or two makes it all neat and manipulable. There is also the problem that unicode-derived ascii will produce "one" legible word for Transcript, no matter how many characters are involved.  A filter is required to space out betwen ";" and "&" which is also not a big deal. I am just not sure exactly what sort of functions Thomas wants to apply to his unicode chunks.