Critical questions for national bodies considering OOXML/DIS 29500
by Rick Jelliffe
Dr Sutor's questions revolve around two premises that he wants us to buy into. The primary premise can be characterized as this: it is better to have no standard than an imperfect standard; unless DIS 29500 has no problems at all you should vote NO or ABSTAIN That this is rubbish should be obvious to everyone: there are always trade-offs and room for improvement. His secondary premise is along the lines of If you support it, you should prove why your support is not suspect. To say this is nasty crap is to denigrate nasty crap.
Lets make it straightforward. Here are each of Dr Sutor's questions, with my comments.
* Was this specification appropriate for the Fast Track process? If not, it should not be approved in such a process and you should ABSTAIN or vote NO.
No Bob. The question is "Is this revised draft good enough and useful enough to be accepted as an International standard, in the reasonable expectation that there will be a good, agile and aggressive maintenance program and in the knowledge that there is work currently being performed on harmonization and testing?"
The issue of the procedure used is irrelevant to the desirability and usefulness of the final draft standard. You can take up the issue of Fast-Track with JTC1. You can get your National Body to lobby JTC1 to not allow standards of more than 100 page to be fast-tracked, or some other concrete proposal.
Think about it for a second! There has been an enormous amount of international scrutiny of the OOXML draft, far more than you would expect in a slow-tracked standard. You can look through an OOXML file and see by even the most trivial inspection that the standard does provide the lion's share of documentation on the XML and notations in the OOXML file; and you can go through the Editor's comments as accepted at the BRM and see that the lion's share of National Body comments have been taken seriously. Objectively, this standard has been really well reviewed, and the process has not prevented this review. NASA sent a man to the moon in about 8 years, for goodness sake.
Saying that there is a lot more work to be done is saying nothing: there is always more work to be done, and the sooner OOXML is a standard and under maintenance, the sooner these things can be addressed. The BRM addressed the big picture issues that need to be right before a technology is on the books: the organization, the conformance classes, the normative status of schemas, the conformance language like should/shall, and the need to have a clear slots for handling legacy/compatibility/deprecated kinds of issues going forward. The big picture is completely good enough for DIS 29500 mark II to be an ISO standard. The BRM dealt with small-picture issues, such as typos and wrong examples and a lot of word-smithing issues that are also needed for a draft to be good enough quality as a standard. And the BRM also dealt with some key internationalization, accessibility and modularization issues that are also a reasonable bottom line for a standard.
Which leaves a myriad of middle-level issues. A number (hundreds!) have been identified and resolved through the BRM. I am sure that many of these issues require more work, which is why moving on to maintenance rather than farting around in draft-standard limbo is the best way to go.
* At each stage of this process, was sufficient time allowed to develop contradictions, completely review the specification in its entirety, generate all appropriate comments, review all proposed resolutions completely and explicitly, and fully review the updated document? If not, you should ABSTAIN or vote NO.
Yesterday I quoted Jim Melton, the editor of SQL, on the XML-DEV mailing list, and I will put the same quote, because unless people realize that the vote to become a standard is the start of getting a standard perfect, not the end they will be sitting targets for FUD: in particular the kind of disinformation that says "This is our only chance, it has to be perfect." Here is what Jim Melton said:
Or perhaps most people were somewhat intimidated by the prospect of (thoroughly) reviewing a 6,000 page document. To put this in perspective for those who know SQL’s size and complexity, the sum of all nine parts of SQL is about 3950 pages. A ballot on SQL frequently receives several thousand comments, and we’ve been balloting versions of SQL for 20 years!
In fact, virtually every large spec I’ve ever had the “pleasure” to review leads to “thread-pulling”, in which every page yields at least “one more” bug, and following up on that one leads to more, and following up on those leads to still more, etc. I would personally be stunned if 30 dedicated, knowledgeable reviewers of a 6,000 page spec on its first public review were unable to find at least 3,000 unique significant problems and at least 40,000 minor and editorial problems. But that’s just me
No Bob. The question is "At each stage of the process, was this draft treated fairly and any differently from other ISO standards?
And another question is "How can I encourage my National Body to participate in the continued improvement of this standard (and ODF!) at ISO/IEC JTC1 SC34? It is very important and worthwhile, and it is so great that this we can proceed steadily on with no artificial deadlines using maintenance."
* Have all your comments been fully and correctly addressed? Are the changes reflected correctly everywhere necessary in the specification? Have you verified this? If not, you should ABSTAIN or vote NO.
No Bob. The question that should be asked is "Have enough of your showstopping issues been approved now to the necessary extent that the advantages of having a standard outweigh the disadvantages.?" It is a balance, a tradeoff. And "Have the championing stakeholders (ECMA and Microsoft) demonstrated a preparedness to engage and make improvements?" for which the answer is objectively yes, based on the BRM at least.
* Is this high quality technology? If not, you should ABSTAIN or vote NO.
No Bob. I'll visit the quality issue later. But the question is "Does this standard fairly attempt to meet its scope statements?" The purpose of DIS29500 is different to the purpose of IS26300 for example: the underlying technology is a fact at loose in the world: hundreds of millions of people use Excel and Word and Powerpoint and the scope of the standard is not to say "This is the best approach in the world" in some abstract way, or "This is the best thing for data interchange" but "This is the information that these applications produce and consume." Bob wants the data formats that millions of people use to be kept proprietary, to be caught up in some limbo of committee work and red tape: by his actions, he wants to opposite of openness.
* Can you can say that you completely understand the specification that emerged from the Ballot Resolution Meeting (BRM), with all its changes, and that it is now a very high quality specification? If not, you should ABSTAIN or vote NO.
No Bob. The question is "Does your national body believe that it would be better to have a standard using DIS29500 mark II than to not have a standard?"
Now I certainly agree that National Bodies should give a reasonable amount of technical scrutiny to any standard they vote on. But if you don't have the technical ability to do a reasonable amount of scrutiny, then you ABSTAIN. And voting NO is completely the wrong vote: if you are ignorant, you don't impose your view on other nations, you keep quiet. And you never, ever, ever try to block something merely because it does not interest you: you respect other National Bodies enough that if they want something but you don't, they can have it. That is the way that consensus works at ISO.
I'll deal with quality below.
* Are you fully confident that no additional problems were introduced at the BRM that your national body would insist must be addressed? If not, you should ABSTAIN or vote NO.
No Bob. The question is "Are you aware that your National Body can notify ITTF and SC34 of any perceived editing SNAFUs and that they will be discussed and, if true, fixed?"
The FUD over all these additional errors that will be introduced is premised on their being no procedures or maintenance to fix them. But it falls down, because in almost all cases the texts that are being introduced in the new draft are completely spelled out. For most of the Editor's dispositions, people have had months to think about the text; indeed in many cases the Editor's comments are just direct applications of the wording that the National Bodies originally suggested.
One of the big unknowns is the move to ISO-ese "should/shall" conformance language: however, since this move in no case changes the normative impact of a sentence (though in a few cases it might make the normative impact of a requirement more stark, it is to be hoped) it should not be a cause for any freak out. But it should be pointed out that ECMA did in fact make available at the BRM an unofficial version of the draft with all the Editor's response changes already there: so any national body can certainly avail themselves of that.
(Now I do agree that the Fast Track procedure is sub-optimal after the BRM: a longer time limit for editors to get his work done would be reasonable so that NBs can vote on the consolidated text, and to relieve ITTF of their overseeing role. But it is not in itself a reason to vote NO or ABSTAIN or YES.)
* As an international standard, does this specification inappropriately favor a single vendor and its products? If so, you should ABSTAIN or vote NO.
No Bob. The question is "Does this specification address more than the petty competitive rivalries of large US multinational corporations and give for the first time a voice to national bodies to provide the kind of documentation and maintenance openness that a market-dominating technology should have? Does this help the users in the world who have to, for whatever institutional or client reasons, use or integrate with Office and its files?"
* Are you 100% confident that there are no intellectual property problems that would prevent anyone from fully and completely implementing everything in the OOXML specification? Do you have this assurance from experts who are not from Microsoft or in their financial ecosystem? If not, you should ABSTAIN or vote NO.
No Bob. That would require that every person who voted had a law degree: no standard is voted for on that level. The question is "Are you reasonably certain that the intellectual property issues of OOXML are no different from those of other similar standards with similar licenses for which there has been no problem"
Lets get realistic here. The people who get sued for IPR infringements on software are the people who have made big enough money to be good targets. And they hire fancy lawyers and duke it out.
The mention of "fully and completely" implement is the giveaway here. Bob's comment is not interested in end users, sitting there creating Excel formulas: these number in their 10s of millions and have no voice and absolutely no chance of suffering IP attacks: they are consumers. And Bob is not interested in system integrators, who are faced with integrating into existing and future Office sites just as a fact of life. No, Bob's comment is interested in making IBM's niche requirements into something that should dominate other considerations. If Bob were remotely serious about this, why hasn't he raised exactly the same issues about ODF? And IBM's license is patently similar to Microsoft's: it is a bravado performance to criticize something as inadequate that is almost identical to your own company's license!
* Has the process as you have seen it been without undue and inappropriate influence by the supporters of OOXML? If not, you should ABSTAIN or vote NO.
No Bob. The question is "Has the process resulted in a standard that is good enough to start and get improved by maintenance at ISO? Has the normal ISO processes resulted in the various sides making their cases, problems being looked at, and serious attempts been made to resolve issues? "
I don't think people are fools Bob. They can see that the readiness of the NOOOXML people (who Bob is happy to share a stage with and happy for his staff to link to) to label any setback for their cause, no matter how trivial, as a de facto sign of undue influence, process iregularity, corruption and bribery, makes these claims completely without credibility. Where is the smoking gun Bob? When you look at these claims, what do you find: ooh some business who would benefit by having the standard goes to a standards meeting and says "This would be useful for us" and that is a sign of corruption?
And Bob, how on earth can you talk about undue influence when you later write Are you willing to stake your professional reputation on that action? Are you really saying that people who dare to disagree with you can expect to have their professional reputations pilloried? I have on several occassions had experts say to me "I want to support it but I am scared that I will get slammed the way you have been". They feel intimidated by the visciousness of the attacks. That these continue on, that they are repeated and amplified through the sockpuppets and into the community, makes it a reasonable expectation. But Bob are you really saying what it looks like you are saying: that people who hold a different opinion to you will be ruined professionally? Is this a warning or a threat? What is the difference between this and the old FUD "No-one ever got sacked for buying IBM"?
* Have the principles of balance and equilibrium in the standards setting process been violated to the benefit of OOXML? If so, you should ABSTAIN or vote NO.
No Bob. There is no principle of balance and equilibrium. You cannot object to a standard just because it might upset your company's marketing plans. The key question for a standard is "Is there a market requirement for this standard" which there clearly is; if there is not a requirement for the largest market dominating application to be roped into the standards process and allowing a non-corporate-dominated governance of the standard, then there is no need for almost any standard.
A standard is an agreement. It is an agreement that is encouraged by public policy because it creates a market. Where there is no universal agreement on a single standard, you have multiple standards. The Allied Tubemakers case is a clear example of a disruptive standard: the metal tube manufacturers tried to block the plastic tubes, but the other technology was ultimately allowed even though it disrupted the business plans of the existing participants.
Equilibrium my foot. There is absoluteness no business for a standard to be turned down because it might disadvantage some commercial enterprise or sector, not Microsoft, not Sun, not Red Hat, not pinmakers, and certainly not IBM. That is anti-standards and pro-cartel thinking.
What is vitally important for the standards process is not "equilibrium" but "equity": does every standard get treated the same regardless of its champions and opponents, and regardless of campaigns of FUD and vilification. The FUD and vilificiation campaigns are no more a reason to vote NO as they are a reason to vote YES. You vote to accept a standard because we would be better off with the standard on the books and under control than being proprietary and with no formal, neutral mechanism for public institutions and stakeholders to influence that technology.
* Were rules broken or changed during this process? If so, you should ABSTAIN or vote NO.
No Bob. Ah, the hanging chads. The question is "Is your National Body aware that if they think rules have been broken, they should raise this issue with JTC1? If a delegate is concerned, are they aware they can raise the issue with their National Body?" You vote for a standard on the technical issues.
There is no such thing as a protest vote on a standard, if your National Body is acting in good faith. The final draft text is adequate or it is not adequate: that is the question.
I was told that at the JTC1 meeting on the Gold Coast, Australia, late last year, the issue of irregularities came up, and no actual evidence for anything wrong was found. What you need is a bank account, or a confession, or something like that, to prove corruption.
I make a joke that a year ago there was only a handful of people in the world who were experts in JTC1 operations, and now there are hundreds of thousands! But when people who have never read the ISO and JTC1 Directives, and never participated in any meetings, opines on procedural irregularity, I hang my head in embarrassment for them.
The excessive interest in rules and procedure just shows a kind of win-at-all-costs mentality. The essence of an international standard is agreement and working through issues and problems, not trying to trump people by procedural vexations. The standards process is a kind of mediation, not a kind of court room or election.
You cannot keep a pig out of muck. In we dive, with Bob demanding we swear on the lives of our children. I cannot read this in any other way that a resounding thumbs up to the atmosphere of presumption of corruption and vilification on anyone who disagrees with his company's policy.
Hmmm, Bob is IBM vice President for Open Source and Standards. He sets the tone for his employees. It would be great to hear Bob say whether it is official IBM policy that, for instance, anyone who supports DIS 29500 mark II as a standard should be prepared to stake their professional reputations.
* If you voted YES on this, are you willing to stake your professional reputation on that action?
Is this a threat, Bob? (Cut to picture of de Niro from Taxi Driver Are you looking at me?)
If anyone in the world has "staked their professional reputation" it is me, but it is not that I am particularly obsessed or thrilled by OOXML but because I refuse to be intimidated. If DIS 20500 is rejected, I will think it is a missed opportunity but I don't see that it will affect my work or career at all. If DIS 29500 is accepted I'll have a beer, but it is not the main game for me.
Perhaps I am lucky to have enough runs on the board that this is only a minor issue and not at all a main part of my career: and I do think it is great that the big boys are competing over standards for XML file formats! Something no-one would have predicted a decade ago, despite our hopes.
I recommend people who are worried about this ignore it. You will get a far greater reputation for integrity by standing up to bullies and intolerant people than by kowtowing to them. (I can understand discretion being the better part of valour of course.) You find out who your friends are and you have the courage of your convictions.
It would be interesting to know who exactly is going to cause our professional reputation to be dragged down if we do have your own angles? Is it bluster, or is Bob announcing the formation of some team of crack character assassins, poised with poison pen ready: the Lotus Ninjas perhaps? :-)
* If you voted YES on this, can you personally attest to the high quality of the OOXML technology and the standards process it went through?
There is a great ISO standard on Quality, ISO 9126: software product quality. I much prefer the old version rather than the 2000 version, it is more pithy, but standards do get maintained! What IS 9126 is about it making subcategories of "quality" which are concrete enough that you can rank their importance and figure out how to measure them. It give three classes which can be used to classify these more concrete subcategoroes: internal quality, external quality, and quality in use.
Quality-in-use can only be determined over time, when the software (in this case, the standard) is deployed and used. We cannot say much about quality-in-use at this time.
Internal quality, in the case of a standard, would relate to wordsmithing and editorial issues, as well as how complete the schema was, whether the normative references were complete, how clear each sentence was, and so on. DIS 29500 mark II is not bad on this score.
External quality relates to how well it can be tested against external criteria. Now in the case of documents, this is where validation fits in. We can test documents using the schemas. We can test formulas against the BNF productions too, and even automate this. So all XML-based, schema using standards provide a significant base-line of quality: that is why SC34 (and W3C) has spent such a big effort in developing the modern schema languages. They objectively allow a significant amount of testing that documents do conform to the specifications; and a standard that has clear, objective, automatable, verifiable conformance, has reached a certain level of quality just by that. Now testing application conformance is a much more difficult proposition, which is one reason I (and some others) are not particularly keen on application conformance (this effects ODF and many other standards too): without a way of automating testing (or, at least, of expressing the constraints in a purely declarative form) it becomes a matter of human judgement whether the standard is enough quality.
However, I say that with a sharp proviso: a standard is not written for a layman or a novice: they are written so that a person who is aware of the application domain could read them. You don't learn typsesetting from DIS29500 or IS26300. You don't learn the mathematical properties of the spline or the delights of postfix programming languages from the ISO standards for PDF.
When considering "quality" of a standard ISO 9126 we have to consider it against its scope. The quality requirements for a standard that said "This standard is intended to express all information in all documents in the world" or "This standard is intended to express the most crystalline and pure forms for office documents" are different from a standard that said "This standard is intended to express the information concretely found in the documents made by a common application." In this last case, "high quality" will relate to accuracy and completeness, regardless of the underlying technology. A "high quality" medical textboook will not say "Let us assume the body is symmetrical, because it would be offensive to left-handed people or go against aesthetics to allow a bias like that": instead it will try to state what is in the body, asymmetry, naughty bits and all, because its purpose is to reveal. To speak of quality without relating quality to purpose and scope is to speak nonsense.
* If you voted YES on this, will you publicly explain why and also detail any current or planned commercial interests you have in common with the supporters of OOXML?
Would you like all our business plans too?
But seriously, I have been open, and it has gotten me into constant trouble from self-righteous liars and over-competitive boors and flambant trolls "with God on their side".
But what about you Bob? Will you reveal your salary and bonuses if the anti-OOXML campaign succeeds? And also for your staff? And also for any lobbyists, for example European ones who may have written white papers for Lotus products but now hide this? What about discount rates for hotel rooms, and other largesse? Readers may not realize it, but it is actually really easy to find scandal and point the finger if you try. I don't know how many blogs I have pulled before publication after deciding not to descend to that level: there is no shortage of ammunition but it just makes heat not light. And most of all, because it is wrong. (Fans of old things might consider here the ninth commandment: Thou shalt not bear false witness. And I Cor 9:10 on slanderers/gossips/revilers.)
What Dr Sutor is doing here, is to reinforce this meme that anyone who is in favour of XML must be having secret business dealing with Microsoft. When you read through comments on his and his coterie's blogs, it is a really common theme: because there is no possible way that anyone can reasonably be in favour of OOXML becoming a standard (which is not the same thing as requiring that anyone should use it!) therefore the only explanation is monetary gain, whether illicit or licit. It is so offensive and narrow-minded. I want to quote what ODF editor Patrick Durusau wrote recently:
Granted, I have a number of issues with the current OpenXML proposal but experts do disagree in good faith even within open standards development projects. If a proposal cannot progress until we all agree, then we risk proposals being held hostage to whim and caprice.
Dr Sutor's questions are hardly attempting to discourage the a priori suspicion of deals: avarice, greed, corruption, business, whatever. Why require evidence? Thoughtcrime is enough: if there is someone who says anything in favour of accepting DIS29500 mark II, the onus should be on them to prove they are not involved in dealings, nefarious or otherwise.
When IBM Dr Sutor writes about "commercial interests" I do not believe he is seriously asking for more transparancy from participants. I believe he is just trying to set up the expectation of deals (sinister or otherwise) to propagandize non-participants, so that if there is an unfavourable outcome the line can be spun in the popular imagination that "The participants had made deals". Fewer marketing opportunities than "We tried and we lost."
* If you previously did not support OOXML but recently changed your mind, will you publicly and in detail explain why you did this?
Ah. Of course, because anyone who dares have a different POV is suspect...
I think it would be great for people to state why they think it would be good to have a standard. But it is just sick to want this to be in an environment of accusation and suspicion.
* Do you personally feel that OOXML helps the ISO and IEC “brands” related to quality of technology and process?
This is the game plan of "If I don't get what I want I will take my bat and ball home". If the JTC1 process doesn't deliver what they want, then bringing the organization into disrepute will work instead. It is just like if an expert doesn't say what you want, you try to bring them into disrepute. I am sorry, but it just seems so cynical to me...if that is the kind of dog-eat-dog world that big business corporate types have to live in, they should do us all a favour and keep to themselves.
Am I being too extreme here? Paranoid perhaps? No, I think I am being anti-paranoid: the paranoia that says that anyone who thinks that DIS29500 mark II is acceptable as a standard must have been got to, bought off, or have some kind of deal!
Well, the thing is that these spin doctors know exactly the toxic environment out there. They know when they make a little ripple it gets caught up into a tsunami. They know you shouldn't falsely shout "fire in a crowded theatrer."
Knowing this, why doesn't Dr Sutor write, for example:
* If you voted YES on this, will you publicly explain why and also detail any current or planned commercial interests you have in common with the supporters of OOXML? And if you voted NO on this, will you publicly explain why and also detail any current or planne commercial interests in rivalry to the supporters of OOXML? And of course, no-one should be so rude, so insecure, so crass, so toxic, so venial, so mean-spirited, so drunk on their own self-righteousness, as to assume without evidence that anyone who has a different position from theirs to be guilty of anything other than having their own brain.
I'd be happy to co-sign an open letter that says that. How about it Bob?
6,000 pages for OOXML and "only" 4000 for SQL. Wow!
Plus there are many people who say that the specification is not complete yet...
|Rick Jelliffe, Geneva
Bertie: I have explained at length on other blog items why simple comparisons of 6,000 pages is misleading. The amount of normative text is more useful.
It is great to see people willing to stand up to the bullying and inevitable character assassination that greets anyone who dare question IBM and its NOOXML campaign. In case you missed it, Jan van der Beld, former Secretary General of Ecma, has also blogged about what he's seen from IBM (and now Google) in the world of OOXML lobbying at ISO. It's a great piece that highlights the hypocrisy of IBM pointing the finger at Microsoft and ISO for supposed "corruption."