> A small problem with parsing:
> when a js or css file is in a link so that it can be 
viewed 
> (<a href='file.css'>read the source</a>), even in a 
> textarea or pre block, the file is parsed and this 
> generates a lot of html error files with any missing file 
> extension.
> Is it really useful?
Err not really.. but I will have to implement some hack to 
avoid these files (I'll have to track opening/closing tags)

> Java applet are sometimes parsed, sometimes not and 
values 
> are not parsed any more with the new versions 
> (value=text.html or value='text.txt' ...)

Values in these tags should not be parsed ; except when 
the "attempt to detect all links" dirty hack recognizes a 
link. The java parsing is quite hard anyway, and is not 
always working properly.