Re: Charter

On Wed, 2012-02-15 at 16:29 +0100, Robin Berjon wrote:
> I added,
> > "Applications must be able to distinguish well-formed XML from
> > corrected-to-well-formed XML." - I hope it's OK.
> 
> I'm less sure about this one. I see where you're coming from but I
> would hesitate to add it at the charter level where it becomes binding
> for all eternity.
> 
> My concern is this: if they can be distinguished at the format level,
> then there's a risk that they might not be compatible.

If one stream parses as XML and one doesn't, *someone* can distinguish
them.

> I wonder if it might not make more sense to handle this as different
> conformance products? "Strict XML processors" MUST handle
> application/xml content according to [XML10e5]. "User Agent XML
> processors" MUST handle application/xml according to [XMLAvK]. Or
> something like that.

I'm trying to make sure that we enable user agents to give warning
messages to developers and authors. One somewhat different strategy for
error recovery (unlikely to fly for a Web browser but very suitable for
an XML editor) might be
(1) parse as XML
(2) on error, offer to load a DTD (often the only way to deal with
"unknown entity" error)
(3) if that doesn't work, offer to try "automatic fixup"

In practice a Web browser probably doesn't want to do that, although for
new documents it'd be very helpful for authors, and for old documents no
longer maintained it'd still work. An editor needs to be "in your face"
about errors, like the difference between a C++ IDE and a
pretty-printer.

I think the issue for a charter is to mandate behaviour consistent with
the XML specification - do not claim something to be XML if it isn't.
I'll be happy with any plausible formulation that achieves this.

The meta-goal is to minimise what Anne called leakage - people creating
documents that "just work in my browser" - so that we don't "break XML"
for people doing civil engineering with it.

Sorry, long answer, middle of F2F here.

Liam


-- 
Liam Quin - XML Activity Lead, W3C, http://www.w3.org/People/Quin/
Pictures from old books: http://fromoldbooks.org/

Received on Wednesday, 15 February 2012 20:01:52 UTC