[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: [xml-dev] Re: determining ID-ness in XML
That's a fair question. Earlier in this thread I said
that the term "xml processor" was quite vague. Is it
the parser, something a parser talks to what? In either
case, does it have to be cognizant of xml: and any name
that follows that? Where does awareness of xml: have
to be implemented?
The question is really "is an id or a name being proposed"?
SGML didn't require IDs of type ID or even CDATA to be
targets of hyperlinks. It didn't have a concept for
hyperlink: just unique IDs and a requirement for
IDREFs to have a target of type ID. If as you say,
this architectural fix is for a hyperlink target,
then an ID isn't required. Never was for any system
even before the web. Some SGMLers did use IDs for that
but for a practical reason: the parser would tell us
if the target was unique. Otherwise, we used namelocs
and application logic to validate. A nameloc, a hyperlink
target, can be implemented per application. In this
case, it doesn't need to be in the xml: namespace and
in fact, can be anything the application developers pick.
So far, there is no compelling reason for it to be either
Tim or James' proposed solutions. So far, the needy thing
is XPointer. So in the end, perhaps the processor that
has to care is not an xml processor at all.
There is no gaping hole in the web architecture. There
is the lack of a standard fragment name. Now you have
to decide if that includes uniqueness. And you are back
to square one.
Without a clear requirement, there is no work to be done.
From: Elliotte Rusty Harold [mailto:email@example.com]
Just so we're clear on what we're talking about, when you write "xml
processors" do you mean XML processor as defined in the XML 1.0
specification, and which common usage calls a parser? e.g. Xerces-J
or expat? or are you talking about something more general?