On Mon, 07 May 2007 19:38:15 +0200, Innovimax SARL <[EMAIL PROTECTED]>
wrote:
Does it mean a conformant implementation could support NO version of
XML?
Yes, in theory.
Isn't there any possibility to put it other way such that at least one
version must be supported ?
I'm not sure how that would be an advantage for people using wanting to
implement this API in some obscure language. At some point when we get
responseBody this will become a relatively simple API to do HTTP stuff
with. I don't think we should mandate XML support for that. It makes sense
to support it though, hence it already is a "SHOULD" for fostering
interoperability.
In that case, I don't understand why "version" is referenced, since only
"1.0" is used, but "1.1" is never referenced.
Ah, I see what you mean. I removed "some version" and mentioned namespaces
as well.
This is clear in the specification. It requires files to be namespace
well-formed.
I, indeed, find that through reading the rest of the spec. But since, it
is almost repeated each time in the spec, why not putting it here ?
Namespaces are now called out.
[[
There is a case-insensitive match of strings s1 and s2 if after
upercasing both strings (by mapping a-z to A-Z) they are identical.
]]
s1 and s2 because there is less confusion with letter a and b
uppercasing because it is used latter in the spec for method
Fair enough, done.
Just to be picky, the uppercasing instead of lowercasing has'nt been
included
Hmm. I didn't feel it was needed. However, it's changed now.
Thanks!
--
Anne van Kesteren
<http://annevankesteren.nl/>
<http://www.opera.com/>
