XOM API for Processing XML with Java

[Elliotte Harold <elharo AT metalab.unc.edu>]

I've had an item in the TODO list for sometime to consider implementing 
a suggestion of John Cowan's to always escape C1 controls (characters 
between 0x80 and 0x9F) even if they were available in the current 
character set. The reasons for doing this are to avoid conflicting with 
old-fashioned gateways and to make the files more human understandable 
in their raw source code form, since all those characters are 
non-printing. I was halfway through implementing this feature when I 
suddenly realized why I shouldn't do this so I'm going to document that 
decision here.

The basic problem is that these characters may appear in unescapable 
situations such as comments and processing instruction data. Requiring 
these characters to be escaped would basically make documents containing 
such constructs, unserializable; and I'm not willing to do that. Using 
these characters in comments or processing instructions or at all may be 
ill-advised, but it is not malformed so XOM needs to support it.

Longer term, it might be possible to escape these characters in a 
different context, such as directly in the Serializer, if it seems worth 
the hassle. I doubt that it is. They are very rarely used in XML. But 
even if someone comes up with a good reason why XOM should do this, I'm 
not going to implement this for 1.0. If there's a real demand for it, it 
shouldn't break anything in the API or implementation and I can add it 
in 1.0.1, but I suspect this is really just a corner case no one really 
cares all that much about.

--
Elliotte Rusty Harold