[LINK] re: Australian consultation on proposed OOXML standard

[Glen Turner]

On Sun, 2007-08-12 at 12:33 +1000, Tom Worthington wrote:

> The comments show a number of shortcomings in Standards Australia's 
> processes. But I suspect that much of the controversy about the 
> meeting is due to a misunderstanding over the fast track 
> international standards process used. Rather than asking "is this 
> good enough to be a standard?" the process essentially says "is this 
> so bad it should NOT be made a standard?".

Tom,

I'm afraid the Fast Track process actually work the other way --
there is procedurally no mechanism to stop a draft standard launched
on the ISO JTC1 Fast Track.

The current ISO ballot allows three possibilities

 - abstain.

 - approve -- the draft is published as is.

 - disapprove with comments -- the draft will go to a Ballot
   Resolution Meeting where all comments will be resolved until
   a consensus to approve the draft standard is reached.

My feeling is that Standards Australia want to vote "approve,
with comments" at ISO but as NIST noted when voting "disapprove,
with comments" in the USA National Body, such a choice does not
exist. An "approve, with comments" is precisely the same as "approve".

If there is a truly bad standard launched down the Fast Track there
is no procedural mechanism to abandon it.  Yet this is what really
needs to happen with OOXML and its authors send back to the drawing
board, probably asked to improve ODF rather than propose a second
format.

This is partly the reason Standards Australia are feeling such
pain on this one. Many of the participants want an outcome that
ISO's processes are not set up to deliver. And SA are trying to
avoid defending such a broken process.

> Rick seemed to be supporting OOXML and says it is not difficult to 
> implement.

Considering Microsoft paid Rick to write the Wikipedia page
on OOXML, I'm not surprised. I was surprised that he didn't
disclose this up-front at the Forum.

As for difficult to implement it depends what you want to
do. If you want to pull some fields out of a form then
ODF and OOXML are much of a muchness, although ODF fully
complies with the XML specification whereas OOXML may break
some XML parsers, especially in non-Latin countries.  My
understanding is that Rick's comments applied to parsing a
document for field extraction (or what we might call "work
flow processing" if we were sales people).

But if you want to build a competing word processor, spreadsheet,
etc then you are bound to examine each of OOXML's sometimes
insanely complex fields in order to achieve compatibility.

Worse still, OOXML references a heap of other Microsoft-specific
file formats. So math formula are in Microsoft's OMML rather than
in MathML, vector drawings are in Microsoft's DrawingML rather
than SVG, bitmaps are in Microsoft'e EMF rather than PNG.

So a competing product can't build on other people's work,
by calling an existing MathML, SVG or PNG library. It has
to construct supporting libraries for the Microsoft-unique
formats too.

In short, OOXML sets a really tall hurdle for a competing
product.  This was Lars from Google's point at the Forum.
If he were to support OOXML in Google Apps and Docs then
it would be a "career ending job" -- he would work on
OOXML compatibility for the rest of his career.  Apps
and Docs supports ODF today, it was about a six month
job to convert to ODF from the previous proprietary
format.

> But, like the reported position of National Archives of 
> Australia (NAA), I think it will be more difficult to support an 
> additional international standard format (OOXML), than it would be to 
> just use the existing one (ODF).

It's not clear to me if the Fast Track process can cope with an
irreconcilable comment such as "OOXML duplicates the field of
application of ODF".  I don't think it can.

Microsoft's response has been to claim that OOXML has a wider
field of application than ODF -- it claims OOXML can be used
as part of a XML workflow.  This is true, but it is also true
of ODF.  As you would expect, since ODF was written by OASIS,
who's major goal is an interoperating set of XML workflow
standards (SOA, SAML, etc).

> Rick argues that government agencies, such as NAA, don't have to 
> implement OOXML, as Standards Australia is not a government body.

There's a MoU between the Commonwealth and SA:

  4.2 The Commonwealth recognises Standards Australia as the peak
      non-government standards development body in Australia.

Noting that there is no comprehensive governmental standards
development body.  The quasi law-making power of Standards
Australia is recognised by the Commonwealth:

  5.4 Where Government seeks the development of Australian Standards
      for regulatory purposes, Standards Australia will endeavour to
      ensure that they are drafted in a form suitable for referencing
      in legislation/regulation and that they represent a minimum
      effective solution.

> However, SA is a government endorsed body for making standards and so 
> its advice is likely to influence Australian government agencies. If 
> agencies send NAA documents in OOXML, then it will have to accept 
> them (and most likely convert them to ODF for long term storage).

It's the OOXML-->ODF conversion that is the problem. As long as OOXML
is a vendor-specific format then that vendor has an interest in 
achieving compliance with international standards by allowing simple
document export to ODF.

Once OOXML is also an international standard the motivation for
Microsoft to support ODF at all disappears. Supporting ODF only
reduces the friction of a company moving away from Microsoft,
and that isn't in Microsoft's interest.


What really hacks me off is that ISO are so close to getting
what they want. Everything supports ODF today -- even Office
has an exporter.  Nearly every product has adopted ODF as
its native format -- the only major exception is Microsoft
Office.

Making OOXML a standard is a step backward from where we
are today. There will no longer be a common format for
moving office documents between applications.

Cheers, Glen

PS: I'll be writing to ACS shortly to encourage then to request
    a vote for "disapprove" from Standards Australia.