Re: [office] Thoughts on ODF-Next

["Andreas J. Guelzow" <andreas.guelzow@concordia.ab.ca>]

On Thu, 2011-01-13 at 06:57 -0700, Michael Brauer wrote:
> Dennis E. Hamilton wrote:
> > In short: I object to any introduction of speculative features by premature
> > appropriation of the ODF x.y schemas and namespaces without any version
> > control and without the full application of the standards-development
> > process.  
> > 
> >  - Dennis
> > 
> > More thoughts:
> > 
> > To suggest to implementers that a CSD that has never seen Public Review is
> > safe to implement strikes me as ridiculous.  It also creates a distorted
> > preferential access to the ODF x.y schemas and namespaces that seems
> > incompatible with any claim to an open-standards process.  
> 
> First of all, let me clarify that I'm not recommending to implement CSDs 
> in favor of approved standards. Whenever a vendor decides to implement a 
> CSD in an application, that application of course should also support 
> the approved ODF standards.
> 
> But I think we should encourage vendors to implement CSDs, since this 
> gives us early feedback to our specifications, and therefore helps to 
> improve the ODF OASIS standard. And early experiences with a 
> specification are part of the OASIS standardization process when it 
> request statements of use. And what could be a better use of a 
> specification than trying to implement it?
> 
> But I understand your concern. Of course, before a feature gets 
> implemented, it should be stable enough to be implemented. So, I maybe 
> should have been clearer, but the CSD's I'm thinking of should be as 
> stable as possible. That is, the new features they contain should have 
> been discussed and agreed in the TC. And we as TC should be of the 
> opinion that the CSDs are good enough to become an OASIS standard if we 
> wish so. That's why I suggest to have development phases of 4 months, 
> followed by editing ans stabilization phases of 2 months. I'm 
> definitively not thinking of approving working drafts as CSD after 6 
> months, regardless in which shape they are. And of course it may be a 
> good idea to wait for a subsequent public review before implementing a 
> feature.
> 

I concur with Dennis. Suggesting that implementers may use CSD's to
justify the use of the ODF namespaces for feature will just continue the
current nightmare of some implementations creating files claiming to be
ODF1.2 before such a standard has been approved. If an implementor wants
to use a feature proposed for a new standard it really should use a
foreign namespace. If in the end that use will match the feature as
adopted in an ODF standard it would be easy for any implementation
supporting the feature in ODF also to support that foreign element.

Andreas