W3C home > Mailing lists > Public > public-xml-processing-model-wg@w3.org > September 2007

Re: Towards a consensus draft (urgent)

From: Norman Walsh <ndw@nwalsh.com>
Date: Tue, 11 Sep 2007 08:38:32 -0400
To: public-xml-processing-model-wg@w3.org
Message-ID: <m2tzq1wf5j.fsf@nwalsh.com>
/ Murray Maloney <murray@muzmo.com> was heard to say:
| At 04:03 PM 9/10/2007 -0400, Norman Walsh wrote:
|>/ Murray Maloney <murray@muzmo.com> was heard to say:
|>| At 01:06 PM 9/10/2007 -0400, Norman Walsh wrote:
|>|>[...]
|>|>bearing in mind that namespace fixup says:
|>|>
|>|>    [Definition: Some steps can produce XML documents which have no
|>|>    direct serialization (because they include nodes with conflicting
|>|>    or missing namespace declarations, for example). To produce a
|>|>    serializable XML document, the XProc processor must sometimes add
|>|>    additional namespace nodes, perhaps even renaming prefixes, to
|>|>    satisfy the constraints of Namespaces in XML. This process is
|>|>    referred to as namespace fixup.]
|>|
|>| Can anyone produce a pipeline that exemplifies the process known as
|>| "namespace fixup"?
|>
|>I'm not exactly sure what you mean, but [...]
|
| What I mean is this... Above, you say "This process is referred to as
| namespace fixup." So, if there is a "process", surely XML Proc can
| be used to describe the process and implementations can run it.

No, I don't think that's the case, unfortunately. There aren't any
atomic steps from which to build such a pipeline. I suppose we could
add a "perform namespace fixup" step, but it would be a black box,
just as "perform XSLT processing" is a black box.

These are really problems that arise only when dealing with synthetic
infosets. Infosets that result from parsing (namespace well-formed)
documents don't ever require namespace fixup.

| I actually suspect that it is more of a rhetorical question; that it is not
| possible to write such a pipeline, and that "namespace fixup" is not
| quite a process so much as a set of guidelines.

Well. I suppose. It's not a single, defined process. It arises from
the observation that the infoset in hand couldn't be constructed by
parsing any actual document consisting of angle-bracket marked-up
text. The process is, "fix it so it could."

| If it is a process, then let's have the pipeline and declare victory.
|
| If it is not a process, let's admit it and provide whatever guidance we can
| and in as much detail as we need to make everybody comfortable with the
| probability of multiple implementations performing "namespace fixup"
| in reasonably consistent manner, where "reasonably consistent" is as
| adjudicated
| by bankers and other conservative businessmen, not jazz musicians.

I believe that 2.6.1 accomplishes that task.

| I take it that Alex is not yet comfortable. As a result, neither am I.
|
| I am not likely to be comfortable until at least all of the implementors
| in the WG are comfortable. I appreciate that this is slowing us down
| on the road to recommendation.

What concerns me most is that we seem to have arrived at an impasse.
Some implementors want to resolve the question by means "A" and some
by means "B" where A and B are mutually exclusive.

The chair welcomes suggestions for how to resolve this impasse.

                                        Be seeing you,
                                          norm

-- 
Norman Walsh <ndw@nwalsh.com> | 'tis expressly against the law of arms:
http://nwalsh.com/            | 'tis as arrant a piece of knavery, mark
                              | you now, as can be offer't; in your
                              | conscience, now, is it not?--Fluellen,
                              | Henry V

Received on Tuesday, 11 September 2007 12:38:46 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:21:54 GMT