[odf-discuss] strategy for ODF migration (was: A story of TIFF)

[Peter Vandenabeele]

On 2/4/07, Daniel Carrera <daniel.carrera at zmsl.com> wrote:
> On Sun, 2007-02-04 at 17:55 +0100, Peter Vandenabeele wrote:
> > Now, if we let the customer choose, a logic conclusion might be to ask all
> > suppliers of plug-ins and converters if they can provide the following:
> >
> >  A plug-in, converter that has an optional "full-fidelity" flag (like
> >  a check-box in the "Save As" dialog), to let it save files either in:
> >
> >   * "ODF" mode (full-fidelity == OFF):[snip]
> >   * "ODFX" mode (full-fidelity == ON):[snip]
>
> This is tempting, but I see a danger.
>
> * For any given person's perspective, it makes sense to use "ODFX".
> Since using ODFX does not _decrease_ interop (it just fails to add any).

My assumption (incorrect ?) was that in "pure ODF" mode the plugin or
converter would/should do a "best effort" to translate _every_ visible feature
of the document to something in ODF, that is maybe imperfectly formatted,
but at least it is there (using the strict boundary condition that only pure
ODF is allowed). That is a quality measure I would attribute to the "pure ODF"
mode of a plug-in.

I presumed that in "ODFX" mode, the development of the plug-in would rather
be optimized to implement full fidelity (escaping to "foreign data" if required)
and spend not that much effort to translate _everyting_ to a best
effort ODF presentation (if dumping it in "foreign data" is easier).

In this assumption (that is only a thesis at the moment), the result of the
plug-in in "pure ODF" mode would be more faithfully represented in e.g.
OO.o, than the result of the plug-in in "ODFX" mode. And that would be the
expected behaviour, because the parameter to optimize is different (full
fidelity vs.  full interop) and the boundary conditions are different
(extensions allowed vs. pure ODF).

Curious what other's opinions are on this. Would e.g. the Foundation plug-in
result in different "openly specified" elements, if it was restricted
to only allow
such "pure ODF" parts and the best effort translation under this boundary
condition would be a separate quality measure in the "pure ODF" mode ?

> This is a case of the prisoner's dilemma. If everyone acts according to
> their best interest, we all lose.

That is exactly why organisations (like Belgium Federal) make e.g. a formal
ODF decision. Of course the status quo and sticking to the current dominant
format is the easy way for each individual on the short term. But
organisations may decide to set group policies that enforce changes on
a global scale if they
think that is beneficial to the group. And one such obvious policy could be to
require a "pure ODF" format for all new documents (which means that features
that are created in MS Office, but can not be saved in "pure ODF" would be
discarded automatically).

> In addition, I think this is what's likely to happen. Everyone adding
> their own tags and attributes. And truth be told, this is not always
> wrong. But that's precisely the problem. It's easy to get started, and
> it's always tempting to add a few more tags that suit you...

I believe it is a reasonable requirement for an organisation to only allow
documents of a certain type, with only a defined set of allowed extensions
(where I would assume the organisation has a full specification of those
extensions).

Peter