stylesWithEffects / musings ?

Wed Feb 17 16:43:49 CET 2010

Interesting that there is no use of the ignorable namespaces.  My hunch is that this means there is no use of the new Office 2010 text effects in this document, and Word 2010 always writes out the StylesWithEffects part, but I'll verify that behavior.  Regarding the use of mc:Ignorable in the wml namespace, I'll look into that as well and perhaps we can discuss on tomorrow's call.

I agree with your comment about not validating parts that are not referenced via a known relationship type.  Such parts may not even be XML at all.  (As I blogged in a contrived example some time ago: http://blogs.msdn.com/dmahugh/archive/2006/11/25/arbitrary-content-in-an-opc-package.aspx)

Regards,
Doug

-----Original Message-----
From: Alex Brown [mailto:alexb at griffinbrown.co.uk] 
Sent: Wednesday, February 17, 2010 12:38 AM
To: Doug Mahugh; Horton, Gareth; MURATA Makoto (FAMILY Given); e-SC34-WG4 at ecma-international.org
Subject: RE: stylesWithEffects / musings ?

Doug hi

In my understanding:

Although the document does designate a couple of Namespaces as ignorable, no content in the document has those Namespaces and so MCE pre-processing has no effect for those Namespaces.

The document is invalid if we apply the wml.xsd schema (T), since that schema does not permit the mc:Ignorable attribute to occur (but perhaps if we follow the - informative - preprocessing description in Part 3 clause 13 this attribute is removed; the wording is not clear).

Murata-san believes such extensions are "anything goes" documents. I think that's right - but am unsettled by the idea that we have no control over "our" Namespaces.

So do we want to modify the spec to outlaw tampering with "our" Namespaces (even after MCE processing, assuming we can get that sorted out too)?

For the purposes of conformance testing, I think any document that exists in the package which is not formally referenced via a known Relationship Type, should NOT be validated - even if we think we know what it is (from its Namespace).

- Alex.

> -----Original Message-----
> From: Doug Mahugh [mailto:Doug.Mahugh at microsoft.com]
> Sent: 17 February 2010 00:51
> To: Alex Brown; Horton, Gareth; MURATA Makoto (FAMILY Given); e-SC34- 
> WG4 at ecma-international.org
> Subject: RE: stylesWithEffects / musings ?
> 
> Hi Alex,
> 
> Yes, that namespace is "one of ours" (SC34's), and its semantics have 
> not been redefined.  Rather, it has a couple of Office 2010 namespaces 
> (see 2010 in the namespace strings) that are designated as "Ignorable" 
> as covered in Part 3, Section 10.1.1.  As stated in that section, 
> "During processing, if a markup consumer encounters an element or 
> attribute in a non-understood and ignorable namespace, the markup 
> consumer shall treat that element or attribute as if it did not exist and shall not generate an error."
> 
> That part exists to enable Word 2010 to "re-hydrate" custom styles 
> that have been defined in Word 2010 (including new Word 2010 text 
> effects), after an editing round-trip through an MCE-aware 
> implementation that doesn't know about our new Word 2010 text effects.  Word 2007, for example.
> 
> So here's the scenario ...
> 
> - user creates a document in Word 2010, adds some new text effects and 
> saves that as a new style
> - Word 2010 saves the styles part with those new effects in ignorable 
> namespaces; Word 2007 will discard the ignorable namespaces, as 
> described in Section 10.1.1 above
> - Word 2010 also saves an identical copy of the styles part, which is 
> not used in the rendering the document; this is a custom part, with 
> its own non- standard relationship type and content type, and it will 
> be round-tripped by Word 2007
> - if the document returns to Word 2010 later, Word 2010 can re-hydrate 
> the custom-defined styles with new effects in it, even though those 
> effects were removed from the styles part by Word 2007
> 
> I think it's a pretty good example of the intended use of ignorable 
> namespaces, and also an example of how the flexibility of OPC can be 
> used to improve the interoperability experience.  And the semantics of 
> the WordprocessingML namespace are not changed in either the styles 
> part or the StylesWithEffects part, so no worries on that front.
> 
> - Doug
> 
> 
> -----Original Message-----
> From: Alex Brown [mailto:alexb at griffinbrown.co.uk]
> Sent: Tuesday, February 16, 2010 10:27 AM
> To: Horton, Gareth; MURATA Makoto (FAMILY Given); e-SC34-WG4 at ecma- 
> international.org
> Subject: RE: stylesWithEffects / musings ?
> 
> Gareth hi
> 
> The stylesWithEffects.xml non-standard (?) document I was referring to 
> has a root element <styles> with the Namespace
> 
> http://schemas.openxmlformats.org/wordprocessingml/2006/main
> 
> which is "one of ours", I believe ...
> 
> - Alex.
> 
> 
> -----Original Message-----
> From: Horton, Gareth [mailto:Gareth_Horton at datawatch.com]
> Sent: 16 February 2010 18:07
> To: Alex Brown; MURATA Makoto (FAMILY Given); e-SC34-WG4 at ecma- 
> international.org
> Subject: RE: stylesWithEffects / musings ?
> 
> Hi Alex,
> 
> The extensions have their own namespaces, e.g. for Excel:
> 
> http://schemas.microsoft.com/office/drawingml/2010/slicer
> 
> http://schemas.microsoft.com/office/spreadsheetml/2010/main
> 
> http://schemas.microsoft.com/office/spreadsheetml/2010/ac
> 
> The full schemas for the extensions can be found in the documentation 
> available on MSDN.
> 
> Obviously, they use existing elements and attributes as well, but I do 
> not believe they use them in non-standard ways - perhaps Shawn can 
> expound further.
> 
> Gareth
> 
> 
> 
> -----Original Message-----
> From: Alex Brown [mailto:alexb at griffinbrown.co.uk]
> Sent: 16 February 2010 16:51
> To: MURATA Makoto (FAMILY Given); e-SC34-WG4 at ecma-international.org
> Subject: RE: stylesWithEffects / musings ?
> 
> Murata-san
> 
> >> So I repeat
> >> my question, are such elements - used in extensions - subject to 
> >> the constraints of 29500?
> 
> > No.  It is subject to the constraints of the MS Office 2010 
> > extension specifications
> 
> A wise man (Rick Jelliffe) wrote recently [1] about another document format:
> 
> "the purpose of namespaces is to *prevent* third parties redefining 
> elements or attributes"
> 
> That's a view which I've certainly subscribed to in the past -- and 
> therefore would assume that we (SC34/WG4) has some kind of jurisdiction over "our"
> Namespaces.
> 
> I think it is certainly bad citizenry/ecosystem-wrong-doing, or 
> whatever, for parties to redefine Standard elements and attributes in 
> non-Standard ways (not least since some tools may use Namespaces as a 
> cue for how to process
> documents) - I'm surprised OOXML allows this!
> 
> - Alex.
> 
> [1] http://lists.oasis-open.org/archives/office-
> comment/201002/msg00024.html
> 
> 
> __________________________________________________________
> ____________
> This email has been scanned by the MessageLabs Email Security System.
> For more information please visit http://www.messagelabs.com/email 
> __________________________________________________________
> ____________
> 
> __________________________________________________________
> ____________
> This email has been scanned by the MessageLabs Email Security System.
> For more information please visit http://www.messagelabs.com/email 
> __________________________________________________________
> ____________
> 
> __________________________________________________________
> ____________
> This email has been scanned by the MessageLabs Email Security System.
> For more information please visit http://www.messagelabs.com/email 
> __________________________________________________________
> ____________
> 
> 
> __________________________________________________________
> ____________
> This email has been scanned by the MessageLabs Email Security System.
> For more information please visit http://www.messagelabs.com/email 
> __________________________________________________________
> ____________

______________________________________________________________________
This email has been scanned by the MessageLabs Email Security System.
For more information please visit http://www.messagelabs.com/email ______________________________________________________________________