RE: Splitting file into N-sized chunks

> Everytime you output a tag, whitespace or text you will have 
> to keep track of its length.  Since you will always have a 
> static XHTML header and footer that wraps your content, you 
> might pre-calculate that length rather than calculate it on the fly.

I suspect that level of accuracy isn't needed. A heuristic that says 500Kb
of serialized XHTML = 250K characters in text nodes is probably quite
adequate for the purpose.

Much more of a concern is how to handle the breaks sensibly if the content
is non-linear.

Regards,

Michael Kay
http://www.saxonica.com/
http://twitter.com/michaelhkay