On 22/12/2007, at 7:24 AM, Bill de hOra wrote:
"updated=gt=2003-12-13T00:00:00Z"
use underscores instead for ranges, ie
updated_gt==2003-12-13T00:00:00Z
I didn't understand why '==' is used as a token instead of '='.
Perhaps the above suggestion helps with that.
Comparison operators are delimited with '=' so extension comparisons
can be easily added down the road. "==" is necessary so that you don't
confuse content with operator.
"By default, a selector is treated as an XML QName [W3C.REC-xml-
names-19990114] which selects any and all child elements of the
entry element which share the same syntax (i.e., the same prefix and
localname; the namespace URI is not considered), along with their
descendant content if any.
"
I'm no fan of Qnames/XMLNS, but that looks butchered to me. You at
least need to figure out what to do with atom in the default ns and
atom with atom:.
Yes, that seems to be where a lot of the pushback on the spec is
centred. Let me ask explicitly: does anyone like the namespace binding
between the URI and the document? I personally think it's not pretty,
but better than some other attempts at referring to QNames in URIs.
The alternatives, as I see them, are:
1) Specify this binding better (as you allude to)
2) Get rid of it completely, and require any non-Atom selectors to be
defined in fq:interface.
The spec doesn't say what should happen with acquired information or
assumed values (ie, author). I have to assume since you're treating
qnames literally, you're not catering for that either.
No, just haven't got to it yet.
I'm not sold on the subset XPath arguments elsewhere btw. This
language seems designed to match across branches not select subtrees.
--
Mark Nottingham http://www.mnot.net/
