W3C home > Mailing lists > Public > public-xml-processing-model-wg@w3.org > May 2009

TBL on XML processing model [was: XProc Minutes 7 May 2009]

From: Grosso, Paul <pgrosso@ptc.com>
Date: Thu, 7 May 2009 16:02:57 -0400
Message-ID: <CF83BAA719FD2C439D25CBB1C9D1D3020F854DF4@HQ-MAIL4.ptcnet.ptc.com>
To: <public-xml-processing-model-wg@w3.org>

> -----Original Message-----
> From: public-xml-processing-model-wg-request@w3.org 
> [mailto:public-xml-processing-model-wg-request@w3.org] On 
> Behalf Of Norman Walsh
> Sent: Thursday, 2009 May 07 10:58
> To: public-xml-processing-model-wg@w3.org
> Subject: XProc Minutes 7 May 2009
> See http://www.w3.org/XML/XProc/2009/05/07-minutes

>   13. http://www.w3.org/DesignIssues/XML

Sorry I missed this call--the discussion sounded interesting.

So I am in the middle of reading a bunch of the references
(and resources referenced from them, etc.).

In "The Pipeline Processing model" section of TBL's document
(url #13 shown above), he says (typos and all):

 The point of this article is that ... the pipeline model is
 basically broken.  Once you have things arbitraryily nested
 inside each other, there is no single pipeline which will do
 a general case.  And nesting things inside each other in
 arbitrary ways is core to the power of XML.

I don't really understand the second and third sentence, but
I'm pretty sure the first sentence is saying that we just
wasted the last couple years of our lives in TBL's opinion.

Am I reading that right?  Is that still TBL's opinion?

As far as his continuing discussion of XML functions (and again,
I apologize for missing the telcon, as I suspect I'd understand
this better if I hadn't), I see TBL giving examples that sound
like Henry's explanation (in the telcon minutes) of a topdown
processing where any element might trigger an xinclude process
or an xslt process wherein, in the processing of that xinclude
or xslt, some processed element in turn triggers another 
process (either another instantiation of the same process or
a different one).  Am I reading this right?

If so, is there anything in our current processing model work
that implements this sort of "multi-thread, recursive" sort
of processing whether an xslt process could call an xinclude
process which could call an xslt process which could call an
xinclude process?

Received on Thursday, 7 May 2009 20:07:14 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:32:47 UTC