Saturday 22 September 2007

The impact of OOXML

(I mentioned some of the following ideas in a presentation I held on Thursday at KTH, and thought I would share them. As a side-note, there may be another FFII-style lecture here later this year, covering a wider range of topics but also some OOXML material again.)

OOXML, to me, is not a standard; I do not recognize the value of it as an Ecma publication, or indirectly in "de facto" terms, simply because the main problem is still there: nobody but Microsoft knows how the format really works, and so not all consumers are yet able to escape the lock-in that is Microsoft Office.

When OOXML was preliminarily voted down on 2 September, not much has actually become clear about its future; most votes have a chance of being changed, especially since we do not yet know what the final proposal will look like. Microsoft's manipulations will certainly continue until February, and thus the usual techniques for prediction may not apply. (Without the interventions of the company, I would say OOXML would be gone already - maybe not even voted upon.)

While one can expect that the standard proposal will require a significant amount of changes to be approved, such fixes could still be of trivial nature, relatively speaking. There are some rather tough suggestions around, though. One of them was submitted by France (J1N8726-03.doc) and others, suggesting that OOXML first be split into two parts; a "core" and "extensions", where the former is something ODF-like, and the latter is an add-on to address properties of the old file formats.

So, in theory, anything could happen; some imaginable scenarios being that:
  • ISO rejects OOXML. While this in itself would not exclude a new submission, that would get much attention with a new fast-track, or require much time without it; the proposal would likely be out of the picture in these cases.
  • ISO decides to split the proposal, and approve the "core" and "extensions" parts as "technical specifications". An "ISO standard" labeling is delayed further, awaiting e.g. merging of the "core" part with ODF.
  • ISO approves OOXML "as-is" (few, or no fundamental problems are solved). This would most likely affect the reputation of ISO itself (due to the scale of the abuse of the process), perhaps to such an extent that the approval would have little weight eventually.
In some sense, OOXML (as published by ECMA) consequently is less of a worry now - regardless of the voting results. However, an "as-is" approval could (through governmental policies etc.) raise a lot of "short-term" problems and further setback in useful development, and it would also have an impact on ISO. It's not just about OOXML, though: XPS is also knocking on the door, and it carries lock-in threats with it as well through software patents. Maybe there is more to come, still.

So, while the harm that could come of OOXML and the end result is probably limited at this point, it is important to fix ISO's - and its members' - working procedures (e.g. for the fast-track, voting rules, etc.) and patent policies, in particular. Right now, ISO allows for licensing terms that are incompatible with free software / open source, and XPS could invite significant trouble in this context. Also, the discussion of ODF vs. OOXML still has important lessons for reaching a new and improved single standard. The patching continues...

Thursday 13 September 2007

EU and terrorism

According to Reuters, EU Justice and Security Commissioner Franco Frattini proposes to "prevent people from using or searching dangerous words like bomb, kill, genocide or terrorism". I haven't yet seen any updates yet, so I'm not sure how this is progressing.

Nevertheless, this is probably my favourite of EU initiatives, only in direct competition with the "take off your tie"-proposal. You see, this one has a certain elegant characteristic; once enforced, any online material on the subject will become unavailable instantly, through magic! Clearly, the journalists - I mean terrorists - need to be stopped now, before they write something that could get the Commission - I mean population - to fear for its welfare.

Out of curiosity, I'd love a more detailed report on how recent EU proposals have improved overall security. I feel some politicians are not getting the credit they deserve; it doesn't make any sense, does it?

Tuesday 11 September 2007

Black sheep

Interesting what happens during one's vacation / work abroad. Miguel de Icaza, founder of GNOME and Mono, writes that "OOXML is a superb standard", even implying that it's better than OpenDocument. With all due respect to his work on GNOME (and Mono, though the conditions for such a project is another discussion), I perceive a sort of dark religious tone over this.

In support of OOXML as a second standard, he goes on to say that "it is always better to have two implementations and then standardize than trying to standardize a single implementation". I'll try to interpret this; it seems to me he's actually suggesting e.g. OpenDocument to be just a description of how Open Office works, and OOXML of Office (but no other products). I must say that this is not the best way to demonstrate an understanding of the standardization process.

On the other hand, I'd agree that Microsoft is in fact trying to design a format based on (only) their own product (they even made some extra effort to include Microsoft Office bugs in OOXML). However, something about Miguel implying that this backwards approach would be valid in the 21st century, doesn't quite feel right.

Basically, I think his first post in the thread can be summarized subjectively as follows:
  • People who think OOXML has fundamental problems are crazy, there is absolutely no merit to their criticism. I won't bother to tell you why, though, because they're crazy. Oh, and it's the best proposal there is, believe me!
  • I know there is OpenDocument, but one more standard can't hurt. Maybe everyone should have their own standard? (Hm, you say this removes the whole point of standards? No way, chill out!)
  • Patents [regarding Moonlight] is not a problem, just get all your software from Novell! (To be fair, though, this part of the response could also be seen as not meaning anything at all.)
Frankly, I'm getting somewhat bored due to the low quality of (most?) pro-OOXML arguments. Please, Miguel, try something better.

Wednesday 5 September 2007

Microsoft spin on OOXML

It's so easy to poke fun at Microsoft's futile attempts of PR spin, but I'll try to restrain myself to a few examples.

In their press release on the preliminary OOXML rejection, it is suggested that there is "strong global support" for OOXML. Tom Robertson is quoted as pointing out the "high quality of the Open XML format" and that "the results from this preliminary ballot are very encouraging".

Therefore, I'm confident that the 100+ comments from many countries are just minor problems (where such numbers were not reached, surely this is completely unrelated to e.g. Microsoft's vote buying), and that Tom is more than happy to deal with at least six more months of desperation and bad press. Well, to each his own. ;-)

Stephen McGibbon from Microsoft goes on to write in his blog that "it is very interesting to note that there is clearly more support for OpenXML already than there was for ODF", referring to a graph that shows no opposition of ODF! Did I miss something here...?

He also adds many nicely coloured charts that clearly show how committees with less or no expertise (i.e. those of non-P countries, especially the ones who have never participated before) tend to follow Microsoft's lead to a greater extent (surprise!).

Overall, I guess I could only be glad to see this fine commitment to honesty. Bravo, Microsoft! :-)

Tuesday 4 September 2007

Microsoft vs. FFII : 1-1 on OOXML

ISO has just confirmed that no criteria for approval of OOXML has yet been fulfilled. They have also announced that a Ballot Resolution Meeting will be held in February 2008 in order to (possibly) get another result.

Update 5 Sept 11:20: FFII just published a press release.

Monday 3 September 2007

OOXML, preliminary ISO vote results

ISO is expected to announce the vote results today. Meanwhile, here is a map showing inofficial estimations (may be incomplete relative to the figures here). According to the voting rules there needs to be a qualified majority (2/3) of all participating members that vote. Of those, currently there seems to be 12+ NO votes, and 6+ abstenstions, leaving at the most 41-12-6=23 YES votes. 23+12=35, and 23 / 35 is 65.7%.

If these figures are correct, it seems that OOXML is not approved at this time. However, the possibility of involving a so called Ballot Resolution Meeting (BRM) has been mentioned. Holding a BRM could mean some NO votes become YES votes, when underlying concerns are addressed. This could mean the process continues for several months - more on this later if so.

Update 4 Sept 14:25: more detailed reports confirm OOXML has not received support on any criteria. 73.91% of the votes of the participating members and 51.61 53.13% of all votes approve of OOXML. 3/4 and 2/3 majorities, respectively, are necessary. At the time of writing, ISO is yet to announce the official results, or any indications on whether a BRM will be held or not. At this point, my feeling is that (while gaining approval later is theoretically possible) holding one would be less than constructive use of their time.

(Sweden is not listed in the table of votes, since they decided to withdraw their decision completely. "Abstention" here means the respective member body has submitted their decision to abstain.)