RE: unclarity about xml:lang position

BTW Greg's point is one that has been giving a lot of us some pause. You
really do have to know way the hell too much to implement these systems.
XML+HTTP 1.1+URL and then you can go do WebDAV. Ick....

A couple of folks in the IETF have been looking at coming out with a
re-written version of the HTTP spec. It would, of course, be non-normative.
The current spec bundles together a lot of features that are not necessarily
related. We figure we could write a spec on the absolute minimally compliant
HTTP/1.1 server in about 10 pages. Client is a bit harder because caching is
something of a nightmare. All though I bet we could identify the 90% useful
features in a page or two and just leave out the rest. I think this
simplification would be really useful but obviously it requires time. I
can't speak for the rest of humanity but I for one have absolutely zero
cycles at the moment.

URLs are just... well...evil. The spec seems soooooo simple. The code you
actually need to implement it however has taken use several years and
thousands of test cases to develop and as I'm sure Larry will point out
about 3 seconds after he gets this e-mail, we are still buggy as hell.

As for XML.... well... as momma says, if you don't have anything good to
say....

			Yaron (think S-Expressions =)

-----Original Message-----
From: Greg Stein [mailto:gstein@lyra.org]
Sent: Wednesday, November 18, 1998 6:26 PM
To: Jim Davis
Cc: Yaron Goland; w3c-dist-auth@w3.org
Subject: Re: unclarity about xml:lang position


Jim Davis wrote:
> 
> At 12:48 PM 11/18/98 PST, Yaron Goland wrote:
> 
> Maybe I'm the one who is slow. It appears that you agree that xml:lang may
> appear anywhere in the request body.  Right?
> 
> It further appears you disagree that the DAV spec is unclear.  I say it
is,
> because I know of at least one person who read it and misinterpreted it.

Read: Greg

:-)

> The spec says, I quote:
> 
> "Language tagging information in the
> property's value (in the "xml:lang" attribute, if present) MUST be
> persistently stored along with the property, and MUST be subsequently
> retrievable using PROPFIND."
> 
> and this led the implementor to conclude the xml:lang has to be in the
> VALUE in particular (as opposed to appearing in any position governing the
> value.).
> I can understand how he might take this position.  It's wrong, but I
> understand it.

I've been concentrating on the WebDAV spec and am not as fully versed
with XML as I should be. There have been a couple other XML issues that
I missed (e.g. can the server imply a default namespace, or must the
client always specify it explicitly, and the single vs double quote
issue).

> I say that it sufficies to show unclarity if even one implementor
> misunderstood the spec.  (I'm not talking about a casual, clueless reader.
> I'm talking about a person skilled enough to implement WebDAV.)

Thanx :-) ... the issue is that a WebDAV implementor also needs to be
*fully* versed in the XML specification(s). I'd also state that you must
be intimately familiar with the HTTP/1.1 spec and the URL
specifications. It is quite a pile of stuff to grok and pull together in
your head :-)

> ...
> Let me put it clearly.  Do you (or anyone else) OBJECT to adding to spec
> language such as Jim W proposed:
> 
> "Language tagging information
> in the property's value, or with scope that affects XML elements in the
> property's value (following scoping rules for the "xml:lang" attribute, if
> present) MUST ..."

Not from me, although I think it could be useful to add a section on the
Working Group home page that details the other RFCs and Drafts that a
person should be familiar with. (i.e. pull out the references section
and list them as "required reading" on the WG page)

Cheers,
-g

--
Greg Stein, http://www.lyra.org/

Received on Wednesday, 18 November 1998 21:35:08 UTC