xml-dev - XML design: why is > allowed in attribute value?

[Daniel Barclay <Daniel.Barclay@digitalfocus.com>]

I've been wondering something about the design of XML.

Why are unencoded greater-than (">") characters allowed in attribute 
values?

I would have thought that greater-than characters inside a tag
(that is, excluding the one terminating the tag) would have been
disallowed, to make it easy for a scanner to identify the ends
of tags without having to parse attributes and their quotation 
characters.


XML does require that less-than characters in attribute values be 
encoded.

It seems that the purpose of this requirement was to make it easy to 
identify the beginnings of tags by simply finding less-than characters,
without having to keep track of whether they appear in attribute-value 
quotes and therefore don't actually signal tags.  (Yes, I'm ignoring 
comments and CDATA sections.)

Is that reasoning correct?  If so, why wouldn't greater-than characters 
be treated similarly, to similarly simplify finding the ends of tags?

(Yes, I know a full parser has to parse everything, but some applications
(e.g., syntax highlighting) might just want to identify the beginnings 
and ends of tags.)

Curiously,

Daniel
-- 
Daniel Barclay
Digital Focus
Daniel.Barclay@digitalfocus.com