Binary Data

Michael Emmel (mike@jmaca.com)
Fri, 20 Feb 1998 16:10:56 -0700


Is it possible to include binary data in a XML document and follow the
spec.

<![CDATA[ ascii data ]]>

allows the inclusion of arbitrary ascii data except I do not think
uuencode or other binary -> ascii/UTF8
encoders will work without modification to eliminate the ]]> encoding.

Would this be possible.

<![BDATA length=1024[ binary data ]]> where the parser would ignore
1024 bytes and expect
to see a ]]> at the end.

The spec seems to imply only character data but does not disallow
binary data.

I assume a character encoding that did not use the ]]> sequence is okay.

I think the <![BDATA length=x[ ]]> tag is not.
You need let the the parser ignore and redirect x number of bytes from
the token stream. This would be equivalent to a "Java production" in
Javacc.
But I'm not sure if it is legal ???

So do I need to alter uuencode or some other encoding format to fit the
<CDATA format
or is it legal to include a binary section. And if not why not : )

Mike

mike@jmaca.com