An Antic Disposition: Cracks in the Foundation - IBM takes over ODF

Wed, 24 Oct 2007 16:34:36 -0000

Comments:

An eMail went out today, October 24th, 2007, nominating IBM's Rob "Show me your garage!" Weir to be the new Co Chairman of OASIS ODF TC. So it's looks like it's true; IBM is moving to take over ODF and OpenOffice.

Not that that's bad. In the long run this is perhaps the best thing that ever happened to ODF and OpenOffice. There is no way IBM's Lotus Notes business plan for ODF-OOo could be any worse than Sun's plan has turned out to be.

~ge~

posted by garyedwards
So, South Africa was watching closely the failed effort in Massachusetts to implement ODF? And now they are determined to make it work?

Good thing they left themselves a “pragmatic” out; “there are standards which we are obliged to adopt for pragmatic reasons which do not necessarily fully conform to being open in all respects.”

Massachusetts spent a full year on an ODF implementation Pilot Study only to come to the inescapable conclusion that they couldn't implement ODF without a high fidelity "round trip" capable ODF plug-in for MSOffice. In May of 2006, Pilot Study in hand, Massachusetts issued their now infamous RFi, "the Request for Information" concerning the feasibility of an ODF plug-in clone of the MS-OOXML Compatibility Pack plug-in for MSOffice applications.

At the time there was much gnashing of teeth and grinding of knuckles in the ODf Community, but the facts were clear. The lead dog hauling the ODf legislative mandate sleigh could not make it without ODf interoperability with MSOffice. Meaning, the rip out and replace of MSOffice was no longer an option. For Massachusetts to successfully implement ODf, there had to be a high level of ODf compatibility with existing MS documents, and ODf application interoperability with existing MS applications.

Although ODf was not designed to meet these requirements, the challenge could not have been any more clear. Changes in ODf would have to be made.

So what happened?

Over a year later, a few days after Sun delivered their proprietary ODF plug-in for MSOffice, Massachusetts announced that both ODF and MS-OOXML would be recognized as implementable standards. (Good one Sun. Thanks a pant load.)

Recent legislative hearings in Texas concerning a bill that would mandate ODF as the only document standard, revealed something very surprising. There was direct testimony that out of the entire Commonwealth of Massachusetts, only 24 desktops had successfully transitioned to ODF. Whoops. Good thing they had the Sun ODf plug-in :)

With Exchange/SharePoint developer hubs going down everywhere, automagically converting binary documents to MS-OOXML, and the MS-OOXML Compatibility Pack plug-in effectively enabling workgroup conversions to MS-OOXML, things do not look good for the future pragmatic implementation of ODF in Massachusetts. Or anywhere else for that matter.

The needed changes to ODf emphasizing the need to meet the Massachusetts defined market requirements of high level of ODf compatibility with existing MS documents, and ODf application interoperability with existing MS applications, were not made by the OASIS ODf TC.

So i wonder just how closely South Africa has been watching Massachusetts?

Not that it's difficult to figure out. In Massachusetts it all came down to the sprawl of MSOffice bound workgroups where business process document exchange required this difficult "high fidelity round trip" conversion. Not only does the ODF plug-in have to perfect a consistently high fidelity conversion to ODF of these active documents, but the documents had to be "round trip" ready in terms of application use.

Round tripping is a killer for ODf, with or without the inclusion of MSOffice desktops in any workflow – document processing chain. The specification was designed to protect content during extensive document exchange and processing chain use, but not presentation!

During an effective document exchange, many applications must be interoperable at the file format level. For ODF this is limited to "content", with the assumption that each application will apply their own "presentation". Sometimes this is jokingly called the "let a thousand presentations bloom but don't touch my data" approach. It's very consistent with the way document processing experts view the world.

The "let a thousand presentations bloom" approach is not however consistent with public expectations for ODF.

The public expects PDF quality of interop between applications, where both content and presentation is preserved throughout the document exchange process - unless and until the owners of the document make changes.

This dilemma brings us back to that age old problem of interactive (as opposed to PDF static) document exchanges being totally tied to specific applications - right down to the version level.

What we found in Massachusetts is that the ODf implementation barrier was due to MSOffice bound workgroups and business processes. And it wasn't just the year long Pilot Study screaming this. They gave us 150 test documents representative of critical business processes where uncompromising fidelity had to be maintained. And yes, the documents were filled with our favorite structural conversion problems of lists, tables, fields, sections and page dynamics. These problems in turn can be attributed directly to the structural implementation differences between MSOffice and OpenOffice.

Let's take this issue to a higher level. The ODF barrier itself had two problem points which we've expresses as the critical market requirements for a high level of ODf compatibility with existing MS documents, and ODf application interoperability with existing MS applications.

The first was that it's impossible to fully capture MSOffice feature sets and specific structural implementation models in ODF. The ODf container is simply not designed for the high fidelity conversion of MS binary or xml documents, which directly reflect the MS application specific implementation model of basic structures.

Our solution was the ODF iX "interoperability enhancements" comprised of five generic elements for the structural basics of lists, tables, fields, sections and page dynamics. These extensions were critical to any ODF plug-in effort needing to establish the high level of MSOffice compatibility - interoperability demanded by Massachusetts workgroups wanting to transition to ODF.

I wonder if the South African workgroups are different? Maybe they avoided the past ten years of MSOffice's dominance as a developer platform? If that's the case, they can move directly to Linux desktops running OpenOffice ODf. There would be no barrier similar to what stopped Massachusetts, California, and the EU-IDABC.

The ODF iX eXtensions were not acceptable to the OASIS ODF TC.

The second problem point has to do with ODF Interoperability in general. To be blunt, there is none. Except at the application specific - version specific level of document exchange. And this is not what the public expected from ODF.

To fix ODF Interoperability, there has to be some level of compliance and conformance requirements. Today, ODf compliance is optional. Nor are there any test suite requirements, let alone anything as strict and demanding as the W3C's "CDR" for W3C Compound Document Formats.

Every ODF application has their own way of implementing ODF. There are no guarantees of interop, or even basic interop requirements.

Still, OpenOffice 2.3 workgroups can exchange documents effectively with other OOo 2.3 workgroups. Just don't try to exchange those same documents with a Lotus Symphony desktop because Symphony is based on a proprietary fork of OpenOffice 1.1.4.

The effect of this is that a Lotus Symphony desktop (ODF 1.0) can't join a OpenOffice 2.3 workgroup (which implements ODF 1.2) without seriously disrupting the workflow.

Now imagine the situation in Massachusetts where they expected OpenOffice ODF 1.0 desktops to be joining ODF 1.0 plug-in enabled MSOffice workgroups! Or how about Linux – OpenOffice desktops joining existing MSOffice directed workgroups?

To compete against MS Server side systems, Linux server systems must either figure out how to interoperate with the MSOffice – Outlook bound workgroups, or, replace those desktops. Massachusetts sent out a clear message; the disruptive cost of rip out and replace was far beyond the price they were willing to pay for much desired freedom to choose, freedom to innovate, and the sovereignty over their information future that could have been theirs. And Massachusetts understood, perhaps better than anyone, the value of complete sovereignty. They couldn't bear the disruptive cost though.

Compound the matter with the reality that today, there are few if any ODF 1.0 - ISO 26300 applications, at any measure of compliance. Near all have moved to some level of ODF 1.2.

Compound that problem further by understanding that ODF 1.2, as currently written, will not pass ISO certification. In May of 2006, ISO Directorate issued a statement that ODF will not be exempt from ISO Interoperability Requirements. The good news is that MS-OOXML also fails to comply. But where does that leave the world?

Confusion, continuing ODf implementation – interoperability difficulties, and the absence of a clear winner in the universal file format stakes will leave the marketplace with only one option: the system default provided by the MS Stack of MS-OOXML ready applications with insidiously proprietary smart tags, XAML, WFP, TFS, and other stack specific collections of dependencies.

This is the interoperability thicket that South Africa so boldly charges into. If the OASIS ODF TC, and the ODF community of applications do not get there act together soon, odds are that South Africa will find themselves in the same predicament as Massachusetts. All dressed up in ODf with nowhere to go but MS-OOXML.

People have accused us of overstating the ODF interoperability problems. The lord of the garage gestapo and his minions insists that these problems can be fixed. Let's hope they are right. Let's hope that South Africa can pull off the implementation of ODF on something more than 24 desktops. And let's hope that this time, the ODf community comes out of the blogosphere and shows up for this important implementation battle instead putting the entire burden on SA as they did in Massachusetts.

A few more failures like Massachusetts and Microsoft is going to be wondering why they ever went through the problems of submitting their proprietary xml to ISO to begin with. Unless and until ODF can successfully challenge MS-OOXML as a real world "implementable" alternative, we're going to find ourselves stuck in the MS Stack for years to come. ISO approval or not.

~ge~

.

posted by garyedwards

Highlights and Sticky Notes:

You must admire their tenacity. Gary Edwards and the pseudonymous "Marbux". The mythology of Silicon Valley is filled with stories of two guys and a garage founding great enterprises. And here we have two guys, and through blogs, interviews, and constant attendance at conferences, they have become some of the most-heard voices on ODF. Maybe it is partly due to the power of the name? The "OpenDocument Foundation" sounds so official. Although it has no official role in the ODF standard, this name opens doors. The ODF Alliance , the ODF Fellowship, the OASIS ODF TC, ODF Adoption TC (and many other groups without "ODF" in their name) have done far more to promote and improve ODF, yet the OpenDocument Foundation, Inc. seems to score the panel invites. Not bad for two guys without a garage.

Tags: foundation ibm odf odf-civil-war opendocument weir

Posted by: garyedwards

OpenDocument's feed | Diigo Group

An Antic Disposition: Cracks in the Foundation - IBM takes over ODF