I'm in Malaysia this week for several meetings, including a briefing set up by PIKOM, a Malaysian ICT industry association and member of the TC4 technical committee and ISC-G in SIRIM, the Malaysian Standards body that is responsible for evaluating DIS29500. PIKOM wanted to hear both sides of the debate around Open XML, and they set up a meeting for this purpose, which I was invited to attend.
I thought I was going to have an opportunity to see Yoon Kit and debate some of the BRM resolutions with him. YK is one of the most prolific bloggers on Open XML lately, and we sat near one another in the BRM in Geneva, so I thought it would be interesting to compare our perspectives on the work done there. Unfortunately, for reasons that aren't clear to me, YK didn't attend the PIKOM meeting last night.
I was told when I first heard of the meeting that it would be a debate on the technical details of Malaysia's 23 comments on DIS29500, to help PIKOM's members understand the issues of most importance to Malaysia. The concept was that IBM would provide five persons who feel Malaysia's comments hadn't been addressed, and Microsoft would provide five people who felt that Malaysia's concerns had been addressed, and we would debate the 23 comments at a PIKOM meeting on Wednesday evening. But when we arrived Yoon Kit wasn't there, and then I heard he had other plans and couldn't make it. Then on his blog post today YK said the meeting was restricted to PIKOM members, which doesn't make sense because most of the people in the room weren't PIKOM members and instead were representatives of big tech companies like IBM, Oracle, Google, and Microsoft, or members of the Malaysian open-source community like Yusseri Yusoff and Dinesh Nair.
I was disappointed not to have an opportunity to discuss the technical details with YK, or with anyone for that matter: nobody else at the meeting had any opinions on the specific dispositions to Malaysia's comments, as it turned out. But today I see that YK's latest blog post mentions that he hopes somebody who attended the meeting will blog about it, so I'll share my thoughts on the PIKOM meeting below. I think it would be great if Yusseri or Dinesh could share their perspective as well on the Open Malaysia blog or in the comments below.
The people in attendance at this meeting fell into three groups: PIKOM members, the anti-OOXML crowd, and the pro-OOXML crowd. I'll not try to name all the PIKOM members since I'm not sure I have everyone's full names, but there were five or six of them present. For the anti-OOXML crowd, there were Hassan Saidin of IBM, Jeremy Allison of Google (co-creator of Samba), Shane Owenby of Oracle, Yusseri Yusoff (Omnilogic), and Dinesh Nair (QubeConnect). On the other side of the table were Dzahar Mansor (Microsoft Malaysia), Oliver Bell (Microsoft Singapore), Jan van de Veld (formerly the Secretary General of Ecma), and Dave Welsh and myself from Microsoft HQ.
PIKOM had a specific view of how to structure the meeting. They wanted Microsoft's side to present arguments in favor of the dispositions to Malaysia's comments, and then they wanted IBM's side to present arguments against those dispositions. We started with a lively debate between PIKOM and Oracle/Google on this detail, because PIKOM wanted to restrict the discussion to Malaysia's 23 comments, but Oracle and Google wanted to discuss other topics. A compromise was eventually reached: Microsoft's side would present an overview of Malaysia's comments and dispositions, and then in the second half of the meeting IBM's side would present their view of the issues that they felt PIKOM should take into consideration.
So the first half of the meeting was me going through Malaysia's 23 comments, the proposed changes from Malaysia, and how those dispositions turned out at the BRM. My colleague Dr. Dzahar helped out on some of the details, and we went through all of the comments in numeric order.
I couldn't explain the reasoning behind Malaysia's vote, of course. So on comments where Ecma had provided exactly what Malaysia asked for, but Malaysia vote Disapprove on the disposition, while others voted Approve and the disposition passed, all I could do was encourage PIKOM to try to get more information in the upcoming TC4 meeting. Some of these were very puzzling: for example, on MY-0008, they asked for a very specific editorial change, Ecma agreed to make the specific change, but Malaysia voted disapprove. The Oracle representative suggested later in the meeting that PIKOM should note that Malaysia's voting changed to consistent disapproval at some point during the BRM, so apparently they felt this explaned such votes. It would have been great to hear from somebody involved in those voting decisions, because all we could do in this meeting was speculate.
Next up were comments from the IBM delegation. These were almost entirely from the Oracle and Google representatives, with a few observations later in the meeting from IBM and the Malaysian ISVs present. I thought we'd be getting into the technical details of Malaysia's dispositions, but instead these comments were entirely about the process. I was surprised by this, but I've heard this has been typical in some of the Malaysian meetings on DIS29500: more focus on process than the technical details. I know YK has opinions on the technical details, so I wish he had been there to debate them at this point.
Most of the process concerns expressed by Oracle and Google were about the size of the specification and whether it could be reviewed properly in the roughly 450 days between its submission to ISO and the end of the process next week. I'll not respond to those concerns here, as others with far more experience in these matters than me have weighed in on such claims countless times over the last year.
Oracle was quite clear in their view that this isn't a vote on the DIS29500 specification, but instead a vote on the Fast-Track process. In fact, Shane Owenby said exactly that in his closing remarks: "you're voting on the process here, and if you feel the process isn't working then you should vote accordingly." As I learned in a fun post by Yusseri last fall, this argument that if the process doesn't make sense to you then you must approve DIS29500 is also known as the Chewbacca defense.
Who's Lobbying Whom Again?
As a final thought, I'd like to respond to one thing YK speculated about in his blog post:
I don't know what took place in the meeting, because I wasn't there, but it seems Microsoft certainly pushed hard to justify the case for OOXML and why Malaysia should change its vote from "Abstain" to "Agree".
Actually, the concept of changing Malaysia's position was only raised by the non-Malaysian representatives of Oracle and Google who attended. They were quite forceful and emotional in their demands for a Disapprove vote from Malaysia, based on their organizations' concerns about the Fast-Track process.
As for my own attempts to influence Malaysia's position, I raised some questions about certain specific dispositions as mentioned above, and I was also the only person present who identified any specific shortcomings in Malaysia's dispositions. I said I thought that 21 of Malaysia's 23 comments have been addressed thoroughly, and that this showed a good-faith effort by Ecma to improve the spec based on Malaysia's feedback (similar to the improvements Ecma made in response to US comments, which have led many in the US to switch to an Approve position on DIS29500 in recent months), but I thought that Malaysia should ask for more on MY-0006 and MY-0016 during maintenance. If anyone else present had an opinion about the specific dispositions to Malaysia's specific comments, they didn't express it during the meeting.
Yoon Kit, you should have been there!