Promoting ODF (OpenDocument Format),After the MTS 2007 -- ODF and OOXML
Return from Redmond: 10 fingers, 10 toes. I'm OK.
We is returned. On the final day of the MTS 2007, we met with Sam Ramji to further discuss the issues regarding the storm in a teacup known as ODF/OOXML. As an aside, Tan Loke Uei made a fantastic presentation on Mobile Devices. He was eloquent, the subject matter was presented in a totally fascinating manner and it was by far the most brilliant presentation throughout the entire three days. *gush* *gush*. We weren't there. We were at the time sitting down with Brian Jones and Gray Knowlton, Program Manager and Group Product Manager for Microsoft Office respectively.
During the meeting with Sam Ramji, we reiterated several things regarding the ODF/OOXML issue from the Malaysian open standards/source community perspective. It's a reiteration because we had already said the same things to the Microsoft Malaysia representative who accompanied us to Redmond. These are:
- To say that the ODF/OOXML issue is simply a vendor fight between Microsoft and IBM is disingenuous and disrespectful of the efforts of the community in pushing for the use of open standards. We made the point that as far as we are concerned, IBM can be here today and gone tomorrow but we would still always push for open standards. IBM is not the boss of us, and we are not the boss of IBM. The way that ODF is embraced by just about everyone who have no vested interest in the matter shows how strongly we feel about having truly open standards. Microsoft and IBM can go bop each other on the head as much as they want, for all we care.
- We continue to feel that Microsoft should have just extended the ODF format within the standardisation process in OASIS, instead of coming up with something else on their own. Now, Microsoft would say that the development of OOXML involved input from various entities, including the Library of Congress, the British Library, Barclay's and Novell but that still does not explain why Microsoft didn't just haul all these other chaps with them and joined the OASIS Technical Committee for ODF.
Sam raised the matter of how interoperability and standards are viewed, with the pragmatic approach being that if specifications are open (at the very least), interoperability can be achieved. Further, Sam noted that Microsoft had already made their Open Specification Promise, which is all nice and well. However, I couldn't help feeling that anything that Microsoft has done from an "openness" perspective are just reactionary, in response to external pressure.
We were asked the applications that used ODF by Sam, and we trotted out the ones that came into our heads, namely OpenOffice.org, KOffice, Abiword, Lotus and Gnumeric. From his perspective, the main application that uses ODF would be OpenOffice.org, which effectively represents the main competitor to Microsoft Office. While we just nodded, I wish that I had the presence of mind at that point to note that from a standards perspective, I would still view the issue from the flip side. Namely, it's not about the application per se, but about the use of the document format.
When the "community" press for the adoption of an open standard, we remain agnostic on the application that uses the standard. The standard allows for choice in application preference. This has always remained the overarching concern when ODF comes into any conversation.
Which brings me to the issue of choice, or rather limiting choice.
I mentioned that one of Microsoft's objections to the use of ODF (or to be precise, the mandating of the use of ODF -- which is another point of contention entirely) is that by doing so, the environment becomes anti-competitive and effectively limits choice. Sam explained that from their perspective, this is correct since ODF's genesis was from OpenOffice.org, and therefore ODF would be closely tied to the features and capabilities of OO.o. This would effectively put Microsoft Office at a disadvantage, since MS Office would have different features and capabilities which would not be catered for by ODF. While Microsoft would support ODF because the market would demand it anyway, they argue that ODF is insufficient for their needs. In short, the use of ODF would show a preference for OO.o, discriminating against MS Office.
I'm going to skip ahead here and mention a little bit of the conversation that we had with Brian and Gray, where they explained the work that they have done on OOXML, and how Microsoft have always been supportive of standards and their adoption of XML is only a little behind that of OO.o. While OASIS was formulating the Open Office XML format (the precursor to ODF), Microsoft was also supporting the use of XML in Microsoft Office.
There are more stuff which were discussed and maybe Dinesh would expand more because I'm boring myself to death already writing this. For now, I would just like to note that BillH, Sam, Brian and Gray all mentioned that they listened to their customers, worked with diverse others for the formulation of OOXML (resulting in 6,000 pages) and consider adopting ODF as their file format would potentially alienate their 400-500 million MS Office users, some of whom have built applications on top of MS Office. This is because some applications would break due to the shortcomings of ODF (in their view) and besides, they would need to expand a lot of effort and resources to get ODF into MS Office.
I've had some days to think this over and I keep returning back to the same old question: Why didn't they just embrace and extend ODF to fit into their requirements? After all, since ODF was developed within OASIS, they could just charm/strong arm their way into the TC, because they have very good reason to be included. Embracing and extending ODF would simply benefit EVERYONE. Preserve all the OO.o specific hooks and put in MS Office hooks as well.
Dinesh mentioned in his post, where Bill Hilf "clarified that the file format (OOXML) was a part of the software and that OOXML and the software (MS Office) are quite inseparable. Ergo, OOXML is an integral and inseparable part of MS Office. That's why they could not adopt ODF as the file format for subsequent versions of MS Office." Sam said that we were under the wrong impression. After thinking about it, I'm not so sure that we are. I remember asking Bill whether the file format can be decoupled from the application or not. That's the key part in all of this. Can it or can it not be decoupled?
Because if it can be decoupled, why not just get your bits into ODF? If it cannot be decoupled, why not just admit it? Having two standards to do the same thing just sounds ... redundant. And no, I still don't buy the argument that Microsoft Office/OOXML and OpenOffice.org/ODF cater for different needs. The applications do the same thing: Office Productivity. The file formats do the same thing: store information created using Office Productivity Suites. Anything else would just be splitting hairs. I may sound too idealistic but we've got to have something to aim for. If not for the ideal, then what else?
I'm ending this. I apologise if this post is a little scattered but this whole issue is complex, diving in and out between technical arguments and political ones. I haven't even touched on the thorny issue of the ratification of ODF/ISO26300 as a Malaysian Standard, but suffice to say that we've made our feelings very clear to both Microsoft Corp and Microsoft Malaysia. See point 1 up above for the gist of it. We are not amused.
Finally, for what it's worth, I would like to sincerely thank Microsoft for inviting us to the MTS 2007. While we may not agree on everything, it was still good to be able to speak candidly and openly with them. Had a good time, though that door still bothers me.
I'm not sure why decoupling the file format from the application is all that important. How many different people are writing application suites anyway? Basically only big companies have the resources to do so, and maybe some of the largest-scale OSS projects like KOffice. In any case, one would want one's file format to contain all the information that's needed in a logical, easy to get at, manner and as little extra information as possible to speed parsing.
It seems impossible and pointless to try to standardize both the display and content of editable documents (I don't think ODF tries to standardize display and OOXML explicitly contains some tags which describe display methods but which are not described like doSomeCrapLikeOldApp). If you do standardize the display, then there is absolutely no room for office suites to be different from each other.
If you have different semantics for the same elements, as you would have to for office suites to handle things differently in any significant way, then it's worse than having different sets of elements in the first place. It's far more confusing for an element to mean a different thing depending on which suite created it, and these suites have vastly different internal architectures so meanings /will/ be different. So any particular organization will standardize on one office suite so that multiple people can edit and print documents and still be assured that they will look exactly the same.
Given the fact that Office suites will be standardized within an organization, it makes sense to have different formats for each suite. We just need some lingua franca formats which can be used between different organizations. Thus, there's no real need to harmonize OOXML or ODF. Having OOXML and ODF as open standards does ensure that converters can be written between them and any other document format one would wish. I think it will be difficult for applications to be both differentiated and support a single standard. Thus, I see ODF either splintering as more people implement it, or new implementations of ODF being almost entirely identical to older ones in terms of features and style.
Posted by: nks | Friday, 06 April 2007 at 05:15 AM
Nks, that makes about as much sense as having different file formats for different web browsers. But if we had that, there would be no World Wide Web. Your argument is in reality an argument against having standards.
And the issue is not just about office suites. A wide range of applications support the open document standard. See http://opendocumentfellowship.org/applications (.) Just as having the standardized HTML file format for the Web led to a proliferation of applications that read and write that format, OpenDocument is sparking a renaissance of development in the office productivity space.
Posted by: Marbux | Saturday, 14 April 2007 at 03:04 PM
The OASIS committee was created for making an open format for OpenOffice. It was not created for making an ISO standard Office document format.
The chairman of the TC even signed it's messages with OASIS Open Office TC.
There is no indication that at the start of this committee (when there was still a relevant participation to be made) Microsoft was asked to join and contribute. The only document that revers to participation at that time is a "Call to participation for the Open Office XML format TC" only one month ahead the start.
Also the OASIS TC never considered compatibility with the format of the market leader which already had an XML format in existance at the time of the TC start.
Was Microsoft ever seriously considered as a participant in the standardisation of the Open Office format in 2002 ???
The formats are hardly the same and/or easy to be made compatible by extending.
OOXML is a lot more hierarchical structured format with its markup created in child and parent structures whereas ODF lacks this structure making it more flexible to combine all kind of markup structures and much harder to fully implement as all this flexibility has the potential for extreme complexity.
Posted by: Hmmm | Wednesday, 18 April 2007 at 08:24 PM