On the first of this month, after reading about the original dispute between ODF advocates and representatives from the GNOME community, I had this to say:
Wasted almost an hour today watching this trainwreck of a line of commentary. Unfortunate, and unproductive.
It may not comes as a surprise, then, to hear that in the month that has revolved by since, I’ve done little but hope that the entire affair would blow over. To own the truth, I’d hoped to have nothing whatsoever to do with this.
But given the fact that the wound seems to be festering and smelling faintly of almonds rather than healing, I feel obligated to comment. In spite of the fact that officially I have no standing with either side, and it’s likely that neither is particularly concerned with what I have to say.
A reasonable mind might inquire, then, what the purpose of commenting at all is, if the result is destined to be ignored. My response to that is simply that I believe commenting to be the right thing to do, in this particular case; besides which, there is always the chance that the studied attention of relatively impartial third parties might encourage further rational discource between the opposing camps. A slight chance, anyway, which is all the justification I need for a Q&A. So herewith, one man’s take on the divide between the various community advocates.
Q: To start, the traditional disclosure statement: any potential conflicts?
A: Possibly more here than any other single issue I cover. From the commercial perspective, RedMonk customers include vendors both for (IBM, Sun) and against (Microsoft) the Open Document Format (ODF), and have been involved in consulting projects around the format and related applications. From the non-commercial perspective, I know personally and have great respect for many of the individuals currently at odds with one another. Detailing them all would take more time than we have here, but suffice it to say that I have ties that transcend the boundaries of the current dispute. Finally, usage. Over three years ago, I switched to a Linux desktop based on GNOME, and have been using that ever since. During that same time period, I’ve been a user of OpenOffice.org, and ODF has become our standard internal file format.
So make of all of that what you will.
Q: Speaking of the current dispute, would you care to summarize that for those who don’t track such things in detail?
A: Certainly. But first, a bit of background. The dispute centers, in my view, around the support or lack thereof the GNOME foundation has provided for the standardization process of OOXML. One level up, OOXML, or Office Open XML, is Microsoft’s alternative to the aforementioned ODF, which is backed by a number of Microsoft competitors. The GNOME foundation is the organization behind the GNOME project, a desktop effort that provides a windowing environment and related tools for projects such as Fedora, Project Indiana, Ubuntu, and others.
Q: Isn’t ODF the open source standard?
A: Well, in and of itself, it’s just a standard, and as such can be implemented by open or closed source software with equal facility. Practically speaking, however, the closest thing that the ODF standard has to a reference platform is OpenOffice.org, which is indeed open source. But be careful of falling into the trap of conflating open source and open standards – they are two very different things, with occasionally overlapping but ultimately different goals.
Q: Where does ODF stand currently relative to its Microsoft competitor, OOXML?
A: A complicated equation that depends both on who you ask and the metric(s) used; there’s a lot of history that can be dragged out depending on how much time you have to answer the question. For our purposes, however, the most salient assessment is the relative statuses of the two competing standards with ISO, the international standards body that plays a major role in determining governmental standards adoption among other functions. At this particular point in time, ODF is approved as an ISO standard and OOXML is not. As reported on by Andy Updegrove, in early September OOXML was voted on within ISO and failed to achieve the 2/3′s majority required. Microsoft and other OOXML backers then had until February 2008 to respond to the comments provided by voters, at which point another vote will be taken. That upcoming vote is effectively the basis for the entire dispute.
Q: How is that one vote responsible for the current situation? What does it matter, really?
A: Well, consider the stakes. The Microsoft Office file format has traditionally been the means of lock-in for the suite itself. That suite is, along with Windows, not only one of the two most successful software products in history, but essentially a license to print money. It is no exaggeration that there are potentially billions of dollars at risk should the popularization of an alternative file format lead to a decline in the sales of Microsoft Office. The market for office applications, after all, is still overwhelmingly dominated by Microsoft, but the thinking is amongst many that if the file format office applications write to is independent and not controlled by Microsoft, then Office would have to compete solely on its merits rather than the ubiquity of documents stored in its signature file format. Both Microsoft and competitors such as IBM and Sun know this, which is why they’ve all been waging a nasty behind the scenes campaign for the hearts and minds of politicians and standards bodies across the globe.
At this juncture, then, ISO certification and the governmental seal of approval that comes along with it is the principal trump card held by Microsoft competitors. Microsoft and its partners would seek to nullify that card by gaining ISO certification for OOXML, making both formats equally acceptable from an international standards body – and implicitly government – perspective.
ISO Certification of OOXML would be a significant blow to ODF, because as Office is the market dominant product at the current time, OOXML is the path of least resistance for Office users. The incentive to transition to ODF is minimal, unless it is accorded a unique and differentiating status in the eyes of standards bodies.
In short, then, there are billions of dollars of pressure on both sides of the ODF/OOXML debate. Backers of the former want ISO certification denied to OOXML, while supporters of the latter are campagining heavily for it. Which brings us back to GNOME.
Q: Indeed? How do they fit in?
A: GNOME, as mentioned above, is a desktop project, and as a result sponsors development – using the term liberally – in a number of desktop related areas. More importantly, GNOME is self-described Free Software; it’s About page includes the following:
GNOME is Free Software and part of the GNU project, dedicated to giving users and developers the ultimate level of control over their desktops, their software, and their data. Find out more about the GNU project and Free Software at gnu.org.
Some of you may see the dilemma already. Should a free software project pragmatically support a Microsoft backed standard? Or should it idealistically shun it, perhaps to the detriment of users of the project?
Whatever you or I may think they should do, the project itself chose the former.
Q: How did they support OOXML?
A: Well, here’s how the GNOME foundation described their participation in their statement responding to critics:
Jody Goldberg is the lead maintainer of Gnumeric, a GNOME-based spreadsheet application, a position he has held for seven years.
Before June 2007, he worked for Novell, representing them on TC45-M in order to obtain further documentation of OOXML during its review process. In June, Jody left Novell and proposed that the GNOME Foundation facilitate his work with TC45-M by joining ECMA as a non-profit.
We accepted Jody’s proposal to make sure that OOXML was documented enough such that FLOSS implementations were possible without a huge amount of pain (as experienced by those working on DOC and XLS binary format support). The decision to participate in TC45-M was made by the Board as a direct result of Jody’s request. It did not involve any third party influence or financial considerations at any point.
Jody’s last interaction with TC45-M was in July, to deal with the latest set of issues he submitted regarding charting and pivot tables. While he is not participating in the current activities of TC45-M (which is focused on issue resolution for the ISO standardisation process), our membership continues so he can participate in the next review period.
During his participation in TC45-M (via Novell and the GNOME Foundation), Jody has raised hundreds of issues with the documentation of the format, which will demonstrate a significant, material, on-going benefit to FLOSS implementations of OOXML and as a result, to users of FLOSS products that require such interoperability.
In other words, the support essentially was designed to ensure that OOXML could be effectively delivered within the context of GNOME software specifically and Free Software generally.
Q: So what’s the problem?
A: Well, there is a contingent of ODF advocates that feels that any support whatsoever for OOXML is essentially a betrayal of the aims of the Free Software community. Here’s a snapshot of that viewpoint from Russell Ossendryver:
Having Gnome team members promoting the agenda of its main opponent, however, is not only counter-productive but also reflects negatively on the project and its credibility. GNOME is supporting its main opponent by explicitly participating in the official Ecma / ISO process; by participating informally at the conferences; and, presumably, by participating inside of actual development. It seems that Gnome is becoming Microsoft’s catspaw to damage and slow down open source and open standards.
Beyond the negative impacts, it has also been implied that this decision to “support” – I employ quotes there because I cannot agree with that assessment as I’ll get to in a minute – OOXML has been influenced by GNOME’s corporate interests. An assertion disputed, of course, by the GNOME statement quoted above.
Q: I take it that you side with GNOME in this dispute?
A: I do. I understand, certainly, the hopes of ODF advocates that OOXML not achieve its desired ISO certification, because I can appreciate how much more difficult it will be to propagate that format if its on equal footing from a standards perspective. The failure of OOXML to achieve certification during the September vote surprised me, but opened in my view a window for ODF. Here’s how I described it then:
There’s a second window. Apparently Doug Heintzman confirmed this notion to InfoWorld, saying that the recent ISO setback for Microsoft is “certainly related” to this announcement. Having missed the office suite window, the question now becomes what the ODF crowd can do with the time it has while Microsoft regroups to try and sufficiently address the comments to achieve ISO standarization for OOXML.
Crucial as this window is, however, for the projects in question, I cannot support or condone criticism of the GNOME foundations current approach. I think it’s misguided at best and inappropriate at worst.
Q: On what grounds?
A: Well, to begin with, there seems to have been from the start a poor understanding of what work, precisely, the GNOME foundation was supporting. Some of the initial commentary, particularly on list, seemed more focus on the fact of the participation over the nature of the participation. The GNOME foundation is adamant that it “is not pursuing or assisting with ISO standardisation of OOXML.” Unless somebody can actively refute that with proof, I don’t see the substance.
Further, it has been implied by some ODF advocates that GNOME should seek to influence Jody’s work, and steer him in a direction away from addressing current OOXML shortcomings. Setting aside the technical aspects to the argument, which I have discussed with Jody in the past, this strikes me as unreasonable in the extreme. While it was obviously GNOME’s choice to support or not support Jody’s efforts subsequent to his exit from Novell, attempting to compel volunteer developers in a direction they’re not interested in is an effort doomed to fail.
Lastly, I think the approach recommended by the ODF advocates in this case is likely to negatively impact users of the GNOME desktop. Even should Microsoft’s bid to have OOXML recognized as an ISO standard fail, it cannot be argued that users of the GNOME desktop are highly likely to encounter OOXML formatted assets regularly, given the volume of Microsoft applications that will be generating them. Without the work of developers like Jody, who is to say that Microsoft’s format – which is documented by north of 6,000 pages – would be implementable within projects like GNOME?
If you believe that users should be the first priority and you concur that OOXML is likely to propagate itself to some degree irrespective of the outcome of the February review of comments, then logically the GNOME Foundation’s behavior here is quite justified.
Q: What would you say to critics that would argue that you’re a Microsoft shill, or less stridently, that you’re too accepting of OOXML and inadequately supportive of ODF?
A: Well, I’d (re)acknowledge that Microsoft is indeed a customer, but point out that our two biggest clients – those at the Patron level – are in fact IBM and Sun. So presumably I’d be shilling for them if I was shilling.
But it’s also worth revisiting what I’ve said on the subject of OOXML previously. In essence, that Microsoft is to be commended for opening its formats, but that ODF was there first. Microsoft’s behaviors with respect to its formats are as logical as they are unfortunate from a customer perspective. Here’s how I put it in September:
The best means of describing my feelings with respect to the question of format are that I’m pro-user, and while Microsoft and I continue to respectfully argue the point, it’s my belief that users are ill served by multiple document formats. Ironically, Microsoft’s monopoly has been good for users in this single area: they’ve had one document format that became, by virtue of the product’s success, a de facto standard.
In my perfect world, then, I’d acknowledge and support but one format.
But neither GNOME nor any other desktop is likely to see that perfect world. As a result, I believe that projects should at least ensure that users of F/OSS will have the opportunity to compete effectively in a marketplace that is likely to demand OOXML support.
Q: Anything else you’d like to add?
A: Just that I’m not a fan of open letters, excepting those of the McSweeney’s variety. They rarely result in productive, rational dialogue in my opinion, and this as been no exception.