Alan Bryden, the ISO Secrerary General, has made some press comments about the OOXML standardisation project, which make interesting reading for those who have been following along. After noting the wrong-headedness of much of the anti-ISO commentary, he is quoted as saying of the standards process:


This is surely the key point: where there are improvements to be made, then the system admits for that change.

It is also noted that ISO/IEC 29500 is not the first controversial project ISO has had to deal with. I am told by ISO veterans that the recent fuss was mild compared to that surrounding ISO 10990-4, for example...

ODF – news?

The latest issue of the ODF Alliance Newsletter has been published (PDF) and predictably enough the lead story is on … OOXML (their favourite topic).

The (I thought) exciting news that Microsoft has announced upcoming support for ODF in its Office products is relegated to second place and gets the rather grudging headline “ODF Alliance calls on Microsoft to act on its commitment to implement support for ODF”. Ah – I didn’t realise the real story here was the statement the ODF Alliance had issued. Being an ODF supporter gets a bit confusing sometimes.

Anyway, the newsletter’s lead story mentions that four countries have appealed to ISO and IEC (as confirmed by the official press release) and adds for good measure that Denmark has “protested” and that there is legal action in the UK.

The appeals of South Africa, Brazil, India and Venezuela are for real, and the texts of two of these appeals are spread widely over the web (I believe Brazil’s text reached the blogs even before it reached Geneva).

Denmark has not appealed. What appears to have happened in Denmark is that an open source lobby group has written a letter to the Danish Standards body. That is not “Denmark protesting”.

And in the UK there have been developments in the so-called “legal challenge” to BSI’s decision on DIS 29500 …

Singeing the King of Spin’s Beard

What has happened in the UK is fairly well documented. In September 2007 the UK, as a JTC 1 P-member, voted on DIS 29500 and submitted a record-breaking 635 comments outlining technical faults. The (officially confidential) details of that September ballot’s votes have been widely leaked on the web showing the UK’s vote of disapproval at that time.

After the BRM, and following the vote modification process earlier this year, BSI published a FAQ explaining what had just happened. In this it stated:


A few days prior to this a report of IST/41’s meeting had appeared on The Register, leaked by persons unknown, apparently in an effort to “smoke out” (read intimidate) the committee members. Angry and/or veiledly threatening emails to committee members duly flowed, but overall the leak had no effect in changing any positions. (But who knows how the suggestion that the UK was changing its vote might have been received in other countries – the leaker may have unwittingly given DIS 29500 a boost!)

Following the end of the voting process, the UK Unix Users Group (UKUUG) issued a press release attacking BSI. The press release began:


and the UKUUG chairman, Alain Williams, provided the following quotation:


– pretty much a direct accusation that the BSI’s own published statements on DIS 29500 were untrue.

Now the UKUUG clearly looks upon itself as a pretty serious organisation. Its press release – not short on pomp – states that it “values intelligence, thoughtfulness and long-term thinking rather than immediacy and froth.”

In that light there are a number of things that are very puzzling about the UKUUG’s recent behaviour. Most of this stems from the fact that UKUUG was a member of IST/41, the very technical committee which recommended the final UK position on DIS 29500.

So, one might ask:

• Why is a Unix (Unix!) user group (user group!) so interested in DIS 29500 anyway?
• Why are its members’ funds being used to finance a preposterous legal action (their site rather desperately asks: “If you're looking to support UKUUG in it's [sic] efforts to examine the BSI's OOXML decision, you can join UKUUG or make a donation via paypal to paypal@ukuug.org.”)
• Even allowing UKUUG is legitimately interested, why did it not send (as it should have) a representative to the crucial meeting on DIS 29500 that decided the committee’s position? That was how it could have lodged its members’ views.
• Having disenfranchised its members and failed to influence the process legitimately in the meeting (UKUUG attendance would have made a decisive difference, I have no doubt), why then attempt to influence the process externally?
• Why question the veracity of BSI’s statements, when as a group UKUUG is entitled access to all and any documents and records which would clearly show it the truth of the matter?

Predictably, when the UKUUG’s “legal challenge” eventually found its way in front of a judge, it got short shrift. The Inquirer reports:


This brings a decisive an end to some decidedly rum behaviour from the purportedly “intelligent and thoughtful” organization that is UKUUG, leaving its members to ask some interesting questions of its leadership.

International Appeals

The fate of the internal UK appeal foreshadows (I suspect) what will happen with the International ones. I believe they will be considered properly; but they will achieve nothing. In fact worse, they will be harmful: Patrick Durusau has written how their net effect will be to distract from needed reform of the JTC 1 standardisation process. The delay in publishing ISO/IEC 29500 will also delay its much needed maintenance and frustrate the very large number of countries who voted to approve DIS 29500 on the basis that it would come under an SC 34 maintenance regime.

Ultimately the balance needs to be got right. I believe the legitimate concerns of the protesting NBs should be heard and acknowledged, but that unless substantial new evidence emerges in this exercise, we should all move forward without further undue delay. It will benefit nobody to go through a lengthy exercise of paying lip service to insubstantial complaints. In that case, okay, these four countries may be mollified – but the overwhelming majority of other countries who expended so much effort in approving ISO/IEC 29500 will look on with increasing exasperation if it becomes apparent the appeals process is nothing more than a drawn-out diplomatic fudge. The multi-month appeals time frames being talked of are completely out of touch with modern ICT standardisation realities, let alone any conception of a “Fast Track” that still remains.

In any case, the appeals have certainly not distracted NBs from continuing to notice that the text of the (pending) ISO/IEC 29500 Standard has not been circulated to them, as the JTC 1 Directives clearly require. They have also noticed that the existence of appeals is no excuse for the non-appearance of this text.

It is to be hoped the JTC 1 NBs will take steps at the next JTC 1 plenary to curtail the power of the standards mandarins and issue instructions to circulate the revised text.

- Alex.

XML UK are running a day conference entitled “XML in the Office” on Thursday 26 June at Reading Town Hall. My own presentation outline promises:


Anybody interested in the XML technology of OOXML and ODF should attend this event (and there will be no doubt be some talk about standardisation too, since XML UK fielded many of the technical experts that contributed to the UK's thorough National review of OOXML).

Registration form here. Hurry while there's space!

Call for Participation

Note there are still a couple of speaker slots and XML UK would welcome more participation. They'd be particularly interested in hearing from people who are willing and able to present on some ODF projects …

Two years ago today the first entry of this blog was posted. Eighteen years ago today I began a whirlwind romance with the woman who became my wife. It is probably more on account of the latter than the former that we are drinking Champagne tonight. But perhaps now, as the DIS 29500 project draws to a conclusion, a little blogging reminiscence is in order …

DIS 29500: some personal reminiscences

Looking back over two years of blog entries, it is clear that the dominant theme has been the standardisation of OOXML (ISO/IEC 29500). One of the first blog entries reported from the May 2006 Seoul SC 34 plenary, the meeting at which it first became evident trouble was brewing.

I remember the meeting well, how there were so many unfamiliar faces present. It was the first time I met Ecma’s Jan van den Beld and IBM’s Rob Weir (to name but two). I remember how, at the opening plenary Jan gave a presentation about Ecma and spent some time explaining how the name “Ecma” was not an acronym but a brand name, and how its capitalisation was important. SC 34 old-timers shifted uncomfortably in their seats – wasn’t this just the sort of corporate bullshit one came to standards meetings to avoid? I remember how Rick Jelliffe (struggling with the crapware on his new Dell laptop) got tapped angrily on the shoulder when leaving the room and told it was no surprise he was a “Windows user”. I remember how a pleasant conversation over lunch was interrupted by somebody scraping their chair across the floor to join our table and launching into a monologue about how we should be wary of OOXML, as it contained secret metadata fields the FBI used to track people down and shoot them.

Clearly something was up, and the forces of unreason were on the move. It was at that meeting that I resolved to do what I could to resist these forces – a route which led me to become convenor of the DIS 29500 BRM.

Looking back at my first blog entry, my overall reaction to what I wrote is “+1”. A couple of extracts are interesting in the light of subsequent developments:


This one counted against my claim to become BRM convenor. I personally believed the accelerated standardisation process were unsuitable then, and the following two years have merely reinforced that belief. Personally, I look forward to the day they (PAS and Fast Track) are scrapped as part of root-and-branch reform of JTC 1.


In the end, this turned out (I believe) to be a decisive factor in the approval of DIS 29500. Ecma made the road very rocky though, by initially producing a text that was so lousy with faults.


As it turned out, technical experts generally had rather more input into the process than I predicted. However in the light of decisions such as Norway’s this is an interesting anticipation of the fact that generally standards approval is ultimately a bureaucratic process.


Oh yes, for sure! Though there are signs the debate is maturing and that when it comes to document formats people are beginning to realise “it’s just stuff”.

Where is the final DIS 29500 text?

The final DIS 29500 text was meant to be submitted by Ecma, and then distributed to National Bodies by 29 March 2008. Ecma duly submitted a text on time but this text has remained with ITTF. National Bodies are unhappy about this, and at the SC 34 Oslo meeting 12 countries passed a unanimous resolution calling for the release of the text – to no effect.

The description of the Fast Track process in the JTC 1 Directives is generally pretty sketchy, but the closing stages of it are particularly poorly thought-through. Is it really sensible if Ecma’s efforts become, unchecked, the final IS text? Personally, I'd say not, and that an all-important QA stage has been omitted. ITTF are perfectly entitled to make special rulings (as they evidently have done, and not for the first time in this project) on the authority of the Secretaries General of ISO and IEC. It would be more sane, I believe, for them to have invented an FDIS stage for this project and have NBs submit editorial corrections. However I can see that the politics and practicalities of the situation make this difficult – it’s not hard to imagine every fault in the text being screamed about by the anti camp as a reason for halting the entire project.

Ultimately the situation raises questions which go to the heart of the relationship between JTC 1 as an entity, and its member bodies. Just who is in charge, the nations or the officials? The unfortunate state of the Directives have meant there have been too many occasions when officials have had to step in and save the nations from the folly of the Directives that they themselves approved. Like ODF and OOXML the Directives is (literally) a standard, a standard that has faults. Unlike ODF and OOXML, however, I am beginning to believe the Directives have got to a state where they cannot be redeemed by evolution and amendment. It may be time to start again from scratch.

Harmony? Patrick Durusau, Jean Paoli and Brian Jones

Since my last blog entry a lot has happened. The big story of course is that DIS 29500 gathered enough votes to pass as an International Standard. Although the margin of votes in its favour was (surprisingly to me) reasonably comfortable, the overall mood of acceptance seems less of a "yip yip yahoo!" and perhaps more of a "well, alright". Everybody is now watching very carefully ...

Resolutions

Now however, the whole process moves forward into a much more significant stage. At the just-finished SC 34 meeting in Oslo a number of resolutions were passed relating to 29500. The most significant of these is resolution 4, "Creation of Ad Hoc Group 1 on ISO/IEC 29500 Maintenance", and it's worth looking at it in some detail. I will go through the complete resolution below with some explanation of my own ...


simply states the fact that JTC 1 have handed full responsibility for the standard over to SC 34. We then have some informative text setting out SC 34's current orientation:


So, SC 34 has a plan: it envisages taking control of OOXML, then setting up a maintance procedure for ODF, and then working on cross-standard initiatives. As the first part of effecting this plan it is setting up a temporary group (in ISO parlance an "ad hoc group") to investigate how OOXML can be maintained. The rest of the resolution sets out the terms of reference for this new investigative group.


The first of these terms of reference set out the task that the group is to perform:


There are several things of note here. First is that the group activities are limited to OOXML only, and that the group will be purely advisory. It will advise its convenor who will weigh up the advice and report back to SC 34 who will in turn make a decision that heeds (or ignores) that advice.

Perhaps the most interesting thing here is the list (items "a" to "e") in which SC 34 gives a "steer" to the group on how it should be thinking about OOXML maintenance. What we learn from this is that the group should try to come up with a process that ensures maintenance activities are fully staffed, as open as possible, that keep Ecma and ISO versions of the standard in sync, and which sticks to the full rigour of normal committee working without any Fast Track or PAS oddities.

The later clauses flesh out some important further details.


sets the duration of the work. It will take place over the next 6 months or so and report in time for SC 34 to consider its findings at the next plenary.


simply states that the usual people will be able to participate in the activity (SC 34 committee members, etc.)


says it is I who will be convening it - an honour indeed!


points out that our super-efficient SC 34 secretariat (of the Japanese NB) will be providing much needed assistance.


sets out some meeting details. The "main event" will be a two day meeting in London in summer.

and ... that's it. The resolution passed with unanimous support.

Bringing Ecma in

The above resolution says nothing of Ecma, since maintenance is now an SC 34 activity. Practically speaking, however, the Ecma TC 45 team are the guys who know more that anybody what is really going on with OOXML, so SC 34 issued a liaison statement (a fancy term for a "message") inviting these experts to participate:



This encourages inclusion of the TC 45 people -- but they will of course have no power of vote at any level in the process.

Short-term maintenance

While all this investigative work is going on to determine the long-term maintenance procedure for OOXML, there is a more pressing task: recording the defects that are currently known and being found in the OOXML spec. With this in mind, a second ad hoc group is being established for collections of comments.



The main thinking here is that knowledge about defects must not get lost. This group, under the expert leadership of Murata Makoto, will make sure that does not happen, and that defects are carefully collected so that they might be fed into the onward maintenance process.

A little local difficulty

While all this work was going on, the OOXML backstory continued. In Norway there is a fuss that the NB (Standard Norway) did not adopt the recommendation of a technical committee. The situation was well explained by the excellent Lars Marius Garshol in his blog entry preceding the Norwegian vote. I can understand the crossness - I certainly would be miffed if our standards body (BSI) did not heed the advice of its technical committee, though in the UK case of course the technical committee recommended that BSI approves DIS 29500 ... What I find less understandable is how what is essentially a local spat (with SN rather obviously acting within its powers) has been blown up into something bigger ...


Steve Pepper

The demonstration passed off reasonably enough ... 100 or so geeks decamped from the adjacent open source conference to stand in the rain and wave placards while Steve Pepper had fun making a speech and brandishing a hairdrier (perhaps risking quips about 'hot air' in the process). There was some half-hearted chanting ("what's that smell? OOXML!"). Meanwhile some on-looking SC 34 people felt insulted. One neutral XML expert, who I know for a fact took a very close technical look at DIS 29500 asked "what are they saying? that we are incompetent? that we do not have the right to decide for ourselves?". Well ... were they?


Disharmony? the anti-OOXML protest

Standards of behaviour

Perhaps the most extraordinary document issued by SC 34 during this Oslo meeting was an "open letter", which received many signatures.


This content speaks for iself.