Tracking OOXML issues

No matter where one stands in the DIS 29500 debate, one fact is indisputable: the vast amount of human effort that is being expended. This struck me yesterday as BSI technical committees met to agree a UK position on Ecma’s responses – a task that is nearly complete thanks, in no small part, to the use of Mantis for tracking each of the UK comments and its proposed disposition.

The work however does not end there as the UK must finalise its view on other NBs’ comments too. As the JTC 1 Directives explicitly state, the reason why all NB comments are distributed is to allow all NBs to form an opinion on all of them:


By extension, of course, NBs shall naturally be considering Ecma’s responses to these comments too. It is this considered national position that delegations will be taking to Geneva:


So, NBs need to do their homework so that delegations arriving at the BRM in Geneva are fully briefed. The delegation should ideally know their national position on all 1,000 or so distinct comment/responses that could be discussed. It is the responsibility of the delegation to faithfully represent their national position (not individual divergent delegate views), and to be prepared to respond to any fresh issues that arise in line with guidance their NB has given them.

1,000 responses in 35 hours?

Given the five day time limit of the BRM, a frequently asked question is: how can 1,000 issues be addressed in the time, even if NBs already know what their position is? The answer, I think, must lie in paper voting. I am sure that the overwhelming majority of meeting resolutions will be decided by voting (as allowed for by the JTC 1 Directives), and delegations will be given lengthy voting papers allowing them to approve, abstain, or disapprove for any proposed resolution. The voting papers are likely to have three kinds of proposed resolution listed on them:

• Verbatim responses from Ecma's proposed disposition of comments (as contained in the document published by SC 34 as N 980)
• Ecma responses that have been amended by the BRM
• Fresh responses arising from BRM discussion

for the latter two types, consensus might well be reached during in-session discussion, in which case there is obviously no need to put the proposed resolution to the additional test of a redundant vote.

In ISO (and as adopted by JTC 1), the word “consensus” has a specific meaning:


Different meeting chairs take different approaches to determining consensus. In general, if the existence of consensus is not beyond doubt on any issue at the BRM, it will be deferred to paper balloting alongside the undiscussed issues.

Now, paper balloting follows normal JTC 1 in-meeting rules:


Tied votes will be assumed not to have passed (unless they get the “further discussion” the Directives require during the BRM, which is unlikely).

Some commentators have been suggesting the BRM will be some kind of biased pro-OOXML love in, but this is laughably wide of the mark. As I have blogged before, a clear majority of delegations are from countries that either disapproved or abstained in the September letter ballot; taking P-members alone, this is even more so, with over two thirds of delegations being such non-approvers. This will be no love-in: I am expecting some hard work and high-quality technical discussion!

As the BRM nears (one month from now I will be in Geneva) I have been busy working with ISO/IEC officials to make sure the meeting mechanisms are sufficient for the enormity of the task we face.

In my last post I mentioned how the BRM could get through the “straightforward” issues by discussing them quickly in batches. Following feedback on this, and further thought, I am tending towards a more radical plan: not discussing these “easy” issues at all, but instead deferring them to ballot papers that NB delegations can complete “off line” so that valuable session time is not consumed considering missing commas and the like. This leaves the sessions clear to discuss more important matters.

And what are these important matters? Well, that is a question only the NBs can answer — and to that end SC 34 has issued a request for NBs to submit a shortlist of the comments/responses that they would most like to discuss. These shortlists will be used for more detailed planning of the sessions, helping to ensure that we get the broadest coverage we can of the most important issues, in the limited time we have ...

With a new year it seems to be a blogging tradition to make some predictions for the coming 12 months. A number of predictions have duly materialised in the blogosphere about DIS 29500 standardisation. At the risk of pouring cold water on some of them, here is a round-up of the state of things in the run up to the BRM in Geneva.

BRM delegation sizes fixed

National Standards Bodies have submitted their proposed delegations for the BRM to JTC 1 and, as anticipated, some have been reduced in size to fit the venue, which takes 120 people.

In all, 40 countries (plus Ecma) are sending delegations, and 14 of them (of more than four people) have been scaled back. The result is that delegations will vary in size between 1 and 6 people — this (precedented) adjustment hardly represents the sort of “disenfranchisement” that was being anxiously predicted in some quarters.

Given the purpose of the BRM the pattern of attendance, as expected, reflects a skeptical stance towards the DIS: 27 delegations are from countries that either disapproved or abstained in the September letter ballot.

Critical role of the Heads of Delegations

It is something of an honour, in my opinion, to be appointed Head of Delegation (HoD) by one’s National Standards Body, and at the BRM this role will carry greater than usual weight. A few points are worth making about HoDs’ responsibilities.

First, the HoD will be responsible for registering votes at the BRM (and it is worth re-iterating that any voting at the BRM is voting on the text, not voting on approval of the standard);

Secondly, in any voting, the HoD (in common with the whole delegation) will be presenting their National Body opinion, not their personal one. This (which is normally the case anyway) is especially emphasised in the JTC 1 directives, which state that BRM delegation members are “representatives who are well aware of the NB’s position”;

Finally, the HoDs play an important role in maintaining order during the meeting, since they are responsible for running their own delegation and keeping it in order.

Random agendas?

The agenda for the BRM will be published soon, probably later this month. While the details are yet to be finalised it is clear that the agenda will not be “random” — in a five-day meeting of 120 people that would be a recipe for disaster. It is also clear that the mainspring of the agenda will be the body of comments submitted by National Standards Bodies in the 2 September ballot. A secondary consideration will naturally be the responses the submitter (Ecma) has made to those comments. Assuming comments are partitioned into common clusters, (e.g. by parts of the DIS) then within each cluster there are likely to be three categories of submitter response:

• straightforward acceptance (everyone agrees on the comment, its NB-proposed remedy is adopted as is);
• modified acceptance (the gist of the comment is accepted, the proposed response differs from that originally suggested);
• non-acceptance/non-response (the comment is not accepted).

There is little to be gained dwelling in areas where everyone is in agreement, and so I expect the meeting to move quickly through batches of comments which have straightforward acceptance. This will leave time for review of the more contentious responses. Inevitably, there is likely to be insufficient time to review all comments in detail – in this case, some will be held until later in the BRM; others will be deferred to maintenance (should the DIS get approved).

Kyoto aftermath

Following informal discussions with some NBs in Kyoto, I think it is perhaps worth re-emphasising that the BRM will be a meeting which concerns itself with the DIS 29500 text, the comments NBs made on it in the 2 September ballot, and Ecma's proposed responses to these comments. The FAQ already spells out some topics which are out-of-scope, and in addition the following should be noted:

• the JTC 1 Directives make clear that comments NBs may have after 2 September are out-of-scope: “comments received after the normal voting period will not be taken into account, except that they will be submitted to the appropriate SC Secretariat for consideration at the time of the next review of the IS in question” (13.6)
• the behaviour and conformance level of any applications which claim to implement DIS 29500 is irrelevant to the standardisation process. ISO/IEC standards have to be complete in themselves without reference to applications, to avoid the possibility of difference (and thus divided authority) between the text itself and any so-called “reference” implementation. In other words, MS Office will not be discussed.

Other Geneva meetings?

It seems the OpenForum Europe is organising a meeting to coincide with the BRM: same time, same venue. It’s not clear from the announcement what time of day this seminar is scheduled to take place, though an invitation is issued to BRM attendees. I will be disappointed if those involved in this event expect BRM delegates abandon their work in session to attend this meeting, as that can only diminish legitimate participation in this important ballot resolution process.

At the SC 34 Closing Plenary

The closing SC 34 plenary passed, as expected, without controversy – but instead with some warmth of feeling as the attendees expressed their appreciation for the outgoing SC 34 officers.

The Unclogging of SC 34

SC 34 seems to be getting over the worst effect of its influx of new members: ballots not getting passed because of insufficient voting. Some NBs attending these meetings were even able to submit votes for open ballots, getting us up to the required numbers just in time. Talking to them it seems the problem has (as usual) been “cock-up rather than conspiracy”, with problems occurring such as emails going awry in the bureaucracies of the new members’ standards bodies.

This means that my working group (WG 1) was able to progress its work on DSDL and the part I am editing (DTLL) is due to advance to FCD (Final Committee Draft) stage early next year. I was also able to continue work on an implementation of a DTLL validator in Java – this is open source and on sourceforge, though I don’t recommend checking it out unless people want a very early view, or want to help with the coding effort (yes please!)

Impressions of Japan

This was a very enjoyable and worthwhile trip, and it is reassuring to have seen SC 34 continuing to function effectively. The Japanese people are very friendly, the food fantastic, and the temples and gardens so beautiful (Kyoto has the highest density of World Heritage Sites of anywhere on the planet). I regretted not taking a more serious camera than my battered Canon compact. If there’s one place in this world that merits lots of megapixels and fine optics, it’s Kyoto.

And OOXML?

Though DIS 29500 was off the agenda, it was, out of session, on most people’s lips. It is my hope that good faith engagement with the process and principled discussion from all (what might be called SC 34 characteristics) will ultimately triumph, though (as patrollers of the blogs will know) feelings seem to be running high in some quarters.

No neutral observer is being so foolish as to predict what will happen – on the long flight home I was pondering this and found (possibly as a result of too much in-flight Sake) a Haiku coming to mind which seemed to sum this up:




Mirror Lake at Ryoan-ji Temple

Part of the Zen Garden, Ryoan-ji Temple

The maintenance of OOXML

There has been a bit of a hullabaloo about the supposed maintenance regime for DIS 29500 (should it become a standard). Neutral observers should know that much of this blogarrhea is “not even wrong”. Here's why.

First, there seems to have been some confusion about Ecma’s proposed maintenance agreement. The important word here is “proposed”, for while some commentators seem to have assumed that this agreement spells out what is happening, such a proposal does not equate to what will happen, what should happen, or even what necessarily can happen.

Secondly, Ecma’s proposal is (naturally enough given the complexities) incorrect in some aspects of its statement of JTC 1 process, and unclear in others. In particular, as was pointed out at the plenary (without contradiction) clause 2 of Ecma’s proposal is mistaken. It states:


However, what clause 13.13 actually states is:


and, as a matter of fact, JTC 1’s “designated maintenance group” is SC 34, and not Ecma. It is for SC 34 to decide how the maintenance of any future standard based on DIS 29500 is maintained.

Thirdly, the “JTC 1 rules” mentioned in the Directives clause 13.13 govern the maintenance and are clear and inviolable. In summary, they state:

• Proposals to amend the text, and acceptance of any such amendments, are subject to normal ISO voting processes (clause 15.5);
• The standard cannot be “stabilised” (i.e. no longer subject to periodic maintenance) except through approval in a JTC 1 ballot (Clause 15.6.2);
• For the standard to be stabilised it must have passed through “one review cycle” (Clause 15.6.1). In this review cycle the text would have to have been re-written to comply with ISO’s formatting and verbal requirements (Clause 13.4).

Fourthly, there are a number of alternative ideas being voiced how the maintenance will actually be handled, of which the proposal for a new working group is but one.

And finally, the rules for the Fast Track (which DIS 29500 is taking) differ from those for the PAS submission track (which ODF, ISO/IEC 26300:2006 took). ODF is being maintained by OASIS, as allowed by PAS. Any future IS 29500 will be governed by different rules as detailed above.

Battle Armour

On the left is a picture of a suit of Japanese samurai armour. It was suggested to me that this would be an appropriate outfit to wear for convening the BRM. Hmmm – it is slightly disconcerting just how many people are telling me this is the worst job in standardisation …

Be forewarned: the ISO process will “fail”

Speaking to fellow NB representatives, it is clear that lobbying (in many directions) is intensifying as the BRM nears. It would be naïve to expect anything else, I suppose. It amuses me the degree of self-certainty both sides have (coupled with very high levels of mistrust). One corollary of this is that they both profess that the only thing that can undo them is a “failure of process”. Tweedledum believes their DIS is so good that only a “failure of process” can thwart it; Tweedledee, however, is convinced that the DIS is so deeply flawed that only a “failure of process” could allow it to become standard.

And so, while I have up till now thought that a solid grounding in the JTC 1 Directives and meeting procedures would be a good education for convening the coming BRM, I am coming to believe that in fact the best preparation is being a father of two small children, both of whom are sometimes prone to intemperate outbreaks of sibling-rivalry. Inevitably, when this ends with one of them feeling they have “lost” a dispute, the complaint will be “it’s not fair”. I am fully expecting something equivalent when this standardisation process produces a result.