I know Rob Weir has been traveling a lot these days lobbying against Open XML across the world, so when I saw this post yesterday I assumed it must be jetlag. I think he completely misunderstood some of the responses from Ecma related to the issue of harmonization, and has missed some significant developments in this area over the past year. I already posted earlier about my thoughts around harmonization, and work that is already under way in the German Standards Body (DIN) to help guide the way. As I said previously, it appears OASIS is already discussing with DIN taking a more direct role in the Working Group, as indicated by this discussion between Florian from Novell and Rob Weir.

Here is what Rob had to say though which had me confused:

Ecma rejected every single one of these requests. Ironically, their response was that harmonization was not necessary because there exist tools that will translate between OOXML and ODF. However, since these conversion tools are restricted in their fidelity because of the lack of these very features, Microsoft's argument is rather weak.

On the question of harmonization, we are either moving toward it, or we are moving away. The Ecma response does not move us toward harmonization, but starts down the road toward further divergence.

But if you actually read the Ecma response, you'll see that TC45's position is actually quite the opposite. Harmonization is not as simple as just adding a few tags here and there. It's going to be a lot of hard work, and the German Standard Body (DIN) is already working on the first step, which is to identify the differences. This isn't something to take lightly.

Here is Ecma's full response to this issue (emphasis added):

There are currently several XML-based document formats in use, each designed to address a different set of goals or requirements. These include ISO/IEC IS 26300 (ODF), China's UOF, and ECMA-376 (DIS 29500 – Open XML). All these formats have numerous implementations in multiple tools and multiple platforms (Linux, Windows, Mac OS, hand-held devices).

The Ecma Response Document from the Fast Track 30-Day contradiction phase for DIS29500 addressed the question of harmonization by explaining the differences between the ODF and Open XML formats as follows:

"... one must recognize that creating a single "merged" format to address the user requirements of both ODF and OpenXML is a much more difficult goal—one that is hindered by fundamental obstacles comparable to what one might encounter while merging HTML and ODF or HTML and PDF. This is because of sheer difference of scope, feature and architecture. Ecma believes that one format cannot simultaneously meet the requirements that would come from the merge of the two formats and the stringent requirements of backward compatibility that drive the design of OpenXML.

First, while both formats share the high-level goal, to represent documents, presentations, and spreadsheets in XML, their low-level goals differ fundamentally. OpenXML is designed to represent the existing corpus of documents faithfully, even if that means preserving idiosyncrasies that one might not choose given the luxury of starting from a clean slate. In the ODF design, compatibility with and preservation of existing Office documents were not goals. Each set of goals is valuable; sacrificing either at the expense of the other may not be in the best interest of users.

Second, the resulting differences are not merely variances in scope that could be resolved by adding capabilities to one or the other. They are structural and architectural in nature. Where functionality overlaps, the corresponding elements nonetheless differ in precise meaning, usage, capabilities, options, and interaction with other elements. Even more importantly, the corresponding elements do not exist in isolation, but are components of whole document models, with different rules and constraints for such things as page/slide layout, flow, style inheritance, event processing, relative positioning, calculation order, formula dependencies, chart construction, graphic templates, animations, and so on. The resulting variations are not merely cosmetic. They compound to create qualitative disparities that, although perfectly acceptable for much of the user base, can be significant for organizations that require high fidelity in layout, content, or editability. Differences between the implicit page style model of ODF and the explicit page style model of OpenXML, differences in the models for splitting table cells, differences in the style information associated with spreadsheet cells, and differences in the full formula specification used in spreadsheets are only small examples of the hundreds of explicit design decisions that ensure the information included in the existing formats is represented faithfully in the OpenXML format."

There are many translation tools already in existence that enable interoperability between different formats by providing useful translation capabilities between ODF, Open XML and UOF.

We note that the German national standards body, DIN, has a committee, NIA-01-34 (see http://www.fokus.fraunhofer.de/fokus/fokus/presse/meldungen_fokus/2007/05/DIN-E.pdf), that is preparing a Technical Report on the translation of documents between the IS 26300 and DIS 29500 formats. The members of NIA-01-34 include format experts from a number of countries, working together to define the numerous differences between these formats.

Ecma strongly supports any harmonization effort that enables better sharing of information and allows better translation between the formats in the following way: Ecma believes that the work of the DIN (NIA-01-34) committee is essential to any harmonization effort. The work of DIN (NIA-01-34) will enable the industry at large to understand the detailed differences between the formats. Based on this detailed understanding, the ODF and Open XML formats could be extended in the future in order to enable better sharing of information and allow future translations tools to provide even better translation and interoperability between the formats.

Harmonization would require functional changes to two International Standards and would fall under the JTC 1 procedures for new work within SC 34 and could be done in the future. Such work should not be done in this Fast-Track process and should not impede the adoption of DIS 29500.


So, as I said there are many approaches you could take towards harmonization. The key for any effort like this though is to first have a full understanding of the issues (in this case identifying the differences), and then you can start to design the solution. I hope that once Rob is done with his travels and anti-OpenXML lobbying (I hear the latest is a trip out to Asia to meet with some national bodies) he's able to get up to speed on the DIN work and as the head of the ODF technical committee he joins in the work towards a better understanding of harmonization.