SCOTT FULTON, BetaNews: You're dealing with a team that, just a few weeks ago, bound itself to a decree of interoperability. But are there times in your line of work where your duty to be open smashes head-on with your ability to provide a one-to-one mapping between the specifications you deal with and those in other people's minds?
DOUG MAHUGH, Senior Product Manager, Office Open XML, Microsoft: The way I would answer that is to step back and look at just interoperability between document formats in general -- let's say, between PDF and HTML. It's always very difficult to do this one-to-one mapping approach. You can always find things that map -- for example, the concept of a paragraph. UOF, ODF, [and] Open XML all have that concept of some structural unit called a "paragraph." So that's kind of low-hanging fruit, that's an easy one. But once you get into how to style text and how to organize the semantics or the structure of the document, each format reflects a slightly different philosophy. And I think that's what the folks at Beihang University came up against.
I was there doing a workshop on Open XML for the people you're talking about, the first week of April of last year. I was there during their research phase that led to the report. But [with regard to] that one-to-one mapping, I saw some of their work there, and it is a complicated, subjective task. If there's one thing that has often been missing in the public debate about it, it's the fact that there are subjective interpretations or decisions to be made about how that mapping could best work. There's not just one canonical mapping out there that everyone could use; it's very much a design question, and there are many different ways to approach it. So I think that's what the UOF group has run into [is] just the reality and the complexity of making all those decisions, given that they are subjective decisions and they all fit together in one big story of interoperability.
SCOTT FULTON: I got the feeling that, all through this debate, there were folks who came to the conclusion that since Open Document Format was already standardized, that would have become already the objective namespace for how things should be mapped internationally...and that anyone else's interpretation, be it China's or Microsoft's or Corel's, thus becomes a subjective interpretation that must be objectively unstrung, like undoing strands of spaghetti, in order to make it parallel with the objective interpretation. There's another line of thinking that says ODF itself is a subjective interpretation [one among many], but that the standardization process is really a ratification of that subjective interpretation as a viable approach.
DOUG MAHUGH: On the comparisons to ODF, one thing that I think is an interesting aspect is, look at Patrick Durusau's recent statements as the editor of the original ODF spec. I spent a lot of time with Patrick in the US V1 [INCITS] technical committee, and I know he has a fairly consistent view on that. He's a big fan of ODF, and he'll tell you point-blank that ODF is his favorite document format, and has some pride of authorship there. But at the same time, the idea of extending ODF to include everything Open XML does, his view is that they started from different goals, and that in the case of the Open XML format, compatibility with this huge corpus of existing documents was the fundamental goal, originally.
In ODF, there was a different set of goals that started with the StarOffice formats, of course, and then they tried to come up with a more generic, minimal subset approach, if you will. And it's not clear that it's possible to extend that as far as would be necessary to encapsulate everything Open XML does, technically, without breaking some of the design assumptions there. An interesting point of reference on that: The DIN working group in Germany that is an international consortium of people looking at these details of how those two formats might or might not map to one another, that's where we expect to see some of the most definitive work, defining what these philosophical differences might be, and what the technical solutions to some of those challenges might be. It's a big, complicated topic, and nobody right now knows exactly what the best mapping would be, but there is some debate about whether it would even be technically possible to extend ODF as far as would be necessary to encapsulate the design goals of Open XML, in addition to some of its original design goals.
SCOTT FULTON: I've read some of Durusau's recent statements -- but you know him personally, I just have bits and pieces of semantics to go on. I gather that his impression is that the process of putting a format like his and like yours under international scrutiny can only improve it, and can in so doing only help weed the garden, if you will.
DOUG MAHUGH: Yes, and I think you see that very much reflected in the evolution of Patrick's view. Keep in mind that, last July -- this is all public now, the V1 votes that took place -- Patrick, at that time, voted for disapproval of DIS 29500, and now he is very publicly in favor of approval of DIS 29500, and as he would tell you, that's based on his view that the process has worked well over the last nine months. He'll be quick to say it hasn't been flawless, and he has specific ideas of how little things might be improved, but overall, he feels that the DIS 29500 spec is in much better shape now than it was nine months ago, and based on that, that's why he's now recommending approval. This is how standards are supposed to work, we all get together and hammer it out.