Questioning Portabilty Assertions

["Dennis E. Hamilton" <dennis.hamilton@acm.org>]

I have been looking at more places where Portability statements are made and
I want to challenge their being there at all.

It is my impression that there is no testable conformance target that can be
specified for *Portable* OpenDocument Formulas.  The reason is that it
involves the behavior, and the intentions, of a player that is not
recognized as a participant against which testable conformance can be
asserted.  We have no OpenDocumentFormula producer target, and we have no
way of characterizing the degree to which what is produced is guided by a
human end-user to assure that their purposes would be satisfied across
implementations (especially since there is no standalone implementation of
OpenFormula except as a calculator, although that is an interesting case).

It seems clear to me that the specification of OpenFormula needs to be
sufficiently complete that someone with a portability concern could discover
practices that MAY improve portability with regard to a targeted
OpenFormula-hosting (e.g., in ODF 1.2 conforming spreadsheet documents).
The degree of successful portability will depend very much on the degree to
which implemented consumers and producers of the OpenFormula-hosting
specification are reliable (and transparent) with respect to a particular
portability-seeking practice.

While I think that portability in a given context is a wonderful
achievement, this strikes me as something far closer to the best practices
for document authors and guidance for practice-encouraging implementations
that Rob Weir has raised on the ODF Interoperability and Conformance TC.  

I don't believe there is anything that can be meaningfully said about a kind
of portability that is independent of OpenFormula-hosting specifications and
the conditions for their conforming implementations *combined* with whatever
the state of support for portability-seeking users is *other* *than* what we
already should be saying about what a compliant OpenFormula targets are and
what the clearly-defined semantics are.

For working purposes, I think we should move portability statements into
working-document notes so that the thinking and the objective behind such
statements is not lost in our work.   We should be looking at what can be
said about what is specified and the qualify of the specified behavior that
is required to ensure that the safe areas for usage are discernable in the
context of a specified OpenFormula hosting.  Then we can remove them or
provide them in a Committee Note and/or encourage the OIC TC to give this
priority attention.

 - Dennis

Dennis E. Hamilton
------------------
NuovoDoc: Design for Document System Interoperability 
mailto:Dennis.Hamilton@acm.org | gsm:+1-206.779.9430 
http://NuovoDoc.com http://ODMA.info/dev/ http://nfoWorks.org 

-----Original Message-----
From: Eike Rathke [mailto:erack@sun.com] 
Sent: Tuesday, February 02, 2010 05:45
To: office-formula@lists.oasis-open.org
Subject: Re: [office-formula] Groups
-OpenDocument-part2-draft18-editor-revision(OpenDocument-part2-draft18-edito
r-revision.odt) uploaded

Hi Patrick,

On Monday, 2010-02-01 13:50:04 -0500, Patrick Durusau wrote:

> We still need to decide what to do with the portable language.
>
> One option would be to move it to a normative annex that treats  
> instances of syntax. The no revision of the portable language speaking  
> of documents would be necessary.

I think moving the portable document text away to a separated annex
wouldn't be good. It is much easier to understand if it is kept within
context.

Earlier we agreed to convert portable documents language to portability
notes, and in a second step turn parts of those notes to normative text
where necessary.


[ ... ]