RE: [oic] ODF 1.1 compatibility / manifest:version in ODF 1.2

[Chris Rae <Chris.Rae@microsoft.com>]

Hi Andreas - apologies for the slight confusion, I was  trying to answer Svante's question about how we ended up building our ODF 1.1 implementation slightly differently to others.

Unfortunately, when we built Office 2007, only ODF 1.1 was available to us, and now that it's out there we are not really able to change that implementation. A surprisingly large proportion of our users never apply patches for various reasons, and we're heavily restricted internally on the types of feature we can issue patches for. 

Because many Office 2013 users will be sharing documents with 2007/2010 users and a great deal of these files will work fine in our ODF 1.1 parser even though they are ODF 1.2 files, we made the call not to write the attribute.

If ODF 1.2 had been an entirely different file format to ODF 1.1 with no chance of documents being understood by a 1.1 parser, we would certainly have written the attribute and caused our older apps to reject the files.

Chris

-----Original Message-----
From: Andreas J Guelzow [mailto:andreas.guelzow@concordia.ab.ca] 
Sent: 13 November 2012 13:29
To: Chris Rae
Cc: Svante Schubert; dennis.hamilton@acm.org; oic@lists.oasis-open.org
Subject: RE: [oic] ODF 1.1 compatibility / manifest:version in ODF 1.2

On Tue, 2012-11-13 at 14:11 -0700, Chris Rae wrote:
> Hi Svante - as I work for the producer who's neglecting to write the attribute, I think I owe you an answer to these questions.
> 
> As I think has been said before on this thread, we are neglecting the 
> attribute because our older implementations will declare the files as 
> corrupt. The part of ODF 1.1 that caused us to declare the files 
> corrupt is the opening sentence of section 1.5, "Document Processing 
> and Conformance":
> 
> "Documents that conform to the OpenDocument specification may contain 
> elements and attributes not specified within the OpenDocument schema.
> Such elements and attributes must not be part of a namespace that is 
> defined within this specification and are called foreign elements and 
> attributes."
> 
> We were reading this to mean, "the only situation in which you can use 
> foreign elements is if they’re in a namespace not defined in ODF 1.1".
> However, I think others were reading this to mean "unknown elements 
> and attributes in namespaces which are defined in ODF 1.1 are not 
> foreign elements, and can occur anywhere".

I don't think we are necessarily reading this any different from you.
And there is really nothing wrong with an 1.1 implementation to reject an ODF 1.2 file. [Personally I would not call that file corrupt since there is the possibility of it being a newer version ODF file.]

The file in question is not an ODF 1.1 file. The sentence you are referring to is about ODF 1.1 files and ODF 1.1 files definitely should not contain any attributes in the ODF namespaces that are not defined there.  

I think the problem arises with an 1.1 implementation trying to read a
1.2 file. Rules in the ODF 1.1 spec do not apply to this situation.

Andreas  

> 
> -----Original Message-----
> From: oic@lists.oasis-open.org [mailto:oic@lists.oasis-open.org] On 
> Behalf Of Svante Schubert
> Sent: 08 November 2012 05:55
> To: dennis.hamilton@acm.org; oic@lists.oasis-open.org
> Subject: Re: [oic] ODF 1.1 compatibility / manifest:version in ODF 1.2
> 
> I am still puzzled, why any ODF 1.1 consumers should have a problem with a mandatory ODF 1.2 attribute?
> Even more a problem that creates a situation that only could be solved by ODF 1.2 producers neglecting the mandatory ODF 1.2 attribute.
> Can anyone point to the quote of the ODF 1.1 specification that is claimed to have led to this situation?
> 
> Thanks,
> Svante
> 
> On 07.11.2012 19:01, Dennis E. Hamilton wrote:
> > I think you have misunderstood what I hope I said.
> >
> > I believe adding manifest:version as a MANDATORY attribute in ODF 1.2 was a mistake.  There was no consideration that this would be a breaking change for no useful value.  In particular, down-level implementations that actually checked for ODF 1.1 schema validity would fail even though there might be nothing in the package that would be a problem for them.  In fact, implementation of the attribute by ODF 1.2 Producers did not even happen at first.  That might not have bothered their own down-level ODF 1.1 Consumers, but it did bother others.
> >
> > Furthermore, the ODF Package is meant to be semi-independent from its use to carry ODF Documents.  Adding this association into the package is brittle coupling.  
> >
> > I stand by my recommendation, whether or not the ODF TC considers an errata for the feature in ODF 1.2 (since it doesn't seem to break any implementation to make it optional except when the package itself is using an ODF 1.2-only feature), whether or not the OIC TC provides an advisory faster than the ODF TC can respond on this matter.
> >
> > The ODF Specification is not Holy Writ, and all adherence is voluntary.  I wouldn't volunteer for <manifest:manifest> manifest:version.
> >
> >  - Dennis
> >
> > -----Original Message-----
> > From: oic@lists.oasis-open.org [mailto:oic@lists.oasis-open.org] On 
> > Behalf Of Svante Schubert
> > Sent: Wednesday, November 07, 2012 01:55
> > To: oic@lists.oasis-open.org
> > Subject: Re: [oic] ODF 1.1 compatibility / manifest:version in ODF 
> > 1.2
> >
> > On 06.11.2012 18:48, Dennis E. Hamilton wrote:
> >> Unfortunately, there are folks who think it is a good idea.
> > Yes, I am still absolute sure, that evolving and extending an 
> > existing standards is a good idea in general and the version 
> > attribute a good idea in special.
> > But perhaps I misunderstood you here, anyway let us not distract 
> > ourselves it would be the wrong TC anyway and let us now just focus 
> > only on a single aspect. What is the root cause for the given issue?
> > You mentioned the previous ODF spec is wrong. Can you provide a 
> > quote on the ODF 1.1 part?
> > I did not find anything justifying the issue among:
> > http://docs.oasis-open.org/office/v1.1/OpenDocument-v1.1.html#outline:
> > 1.5.Document_Processing_and_Conformance
> >
> > Thanks,
> > Svante
> >
> > --------------------------------------------------------------------
> > - To unsubscribe, e-mail: oic-unsubscribe@lists.oasis-open.org
> > For additional commands, e-mail: oic-help@lists.oasis-open.org
> >
> 
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: oic-unsubscribe@lists.oasis-open.org
> For additional commands, e-mail: oic-help@lists.oasis-open.org
> 

--
Andreas J. Guelzow, PhD FTICA
Professor of Mathematical & Computing Sciences Concordia University College of Alberta