CyberTech Rambler

November 21, 2007

Sorry, Matusow, I have to disagree with a few of your accussation of Bob

Filed under: Uncategorized — ctrambler @ 4:37 pm

After months of quiet after a flurry of activities on the ODF vs OOXML front, things are warming up as we approach the Ballot Resolution Meeting (BRM) in February. I am glad that the BRM coverner, Alex Brown, had posted a few neutral/procedural information such as this FAQ on his blog during this lull where everyone’s head is cooler.

The first major salvo in the debate, as far as I can tell, came from Bob Sutor’s piece which Jason Matusow find it necessary to correct. While readers of this blog knows that I almost always disagree with Matusow, it is not characteristics of Matusow to point finger at people. I read Matusow’s opinion and have to say I do not agree with him this time as well.

Before I dissect Matusow’s clarification on where Sutor went wrong, let’s set one basic rule. I am not going to comment on the Sutor’s or Matusow’s opinion. Sutor is not impressed by OOXML, is a very prominent member on the ODF camp. This means most of his opinion is going to be “Do not accept OOXML” type. Matusow comes from the opposite background. There is no point criticizing them on “why cann’t you just stay neutral”. Rather, I will concentrate on where I think they got too far, or factual things that I think they overlooked.

Sutor: The BRM will result in a changed specification – those changes may themselves result in new problems.

Matusow: “..[T]he whole point of the BRM is to change the specification in the interests of improving it and driving greater consensus for adoption”

The only agreement here is the specification will be changed. I do not think Sutor will disagree with Matusow that the purpose of the changes, come BRM, is to make the specification better. So, I believe the argument here is whether the change is for the “better” or “worse”. Bottom line is changes, any changes, at any stage of the process, need not necessary be in the interest of improving it. A lot of time changes themselves results in new problems. Hence, I think Bob’s comment, taken at face value, is fair, especially he had chosen very carefully to insert the word “may”. I will argue that, if we factor in his bias against OOXML, this statement even “fairer” because he is duty-bound to highlight potential problems. Matusow’s argument that editorial process in place, both at ECMA and ISO and all those comments received, including IBM’s , only means that it is likely that the specification will be better, not a guarantee that it will be better.

Sutor:The specification is long, there are thousands of comments, there is no way for a national body to consider the dispositions in the time they have.

The essence of Matusow’s counter argument is that there had been a lot of time spent before September’s ballot and since to digest the comment. As such, the time is adequate. There is one problem with Matusow’s statement. Those “thousands of comments” are filed in September. Presumably, the pre-September digestions had reduce millions of comments (including frivolous ones) down to those “thousands” that need further work. Therefore, the only available time to digest through the “thousands” of comments is between September 07 to Feb 08, approximately 6 months. Since a lot of these comments are likely to overlap, and that pre-September time spent on the OOXML will help (but not to the extents Matusow is implying), I cannot answer the question on whether there is enough time to digest.

Sutor:There are questions about the BRM process, attendance, time to consider all comments etc. – the process is bad.

I agree with Matusow that Sutor went overboard here. This is actually an attack on ISO process in general, on the coverner Alex Brown in particular because it is saying that they lost control of the process. What is clear is everyone, including IBM and Microsoft, will try to game the process. ISO and Brown’s job is to ensure fair play. Until it is demonstrated that they do not, we have to refrain from criticizing them. The path leading up to September votes open a lot of question on NB’s involvement and potential gaming inside National Bodies. It has nothing to do with the BRM process. It is wrong for Sutor to say the process is bad today unless he is willing to show proof of this. By the way, may be the gaming had started as Rui Seabra is accusing Portugal of playing “musical chair” again.

Sutor: If you are not satisfied for any reason, vote against Open XML. The specification should be perfect or you should reject it.

Sorry Matusow. First sentence is fair. Reading ISO Fast Track rule suggests that both explicitly and implicitly, this is how the game should be played, if you are an unbiased National Body. But Sutor’s second sentence is too strong. As Matusow points out, no specification, including ODF is perfect. I, however, think that Matusow is wrong to imply that since ODF moves from 1.0 to 1.1 then to 1.2, it must therefore be immature to present ODF 1.0 to be approved by ISO. I have to accuse him of spreading FUD here as I am sure he understands the evolution of standards. It will be the same thing as me accusing Microsoft of releasing Windows 3.1 as it is immature compared to Window Vista.

Sutor: Don’t set a bad precedent with this specification.

Do not set a precedent with ANY specification so if this specification can be a bad precedent, do not set it. Matusow argues that this will not be a bad precedent because (1)Open XML has received more scrutiny and (2) dedicated engineering attention than any specification in JTC1 (if not ISO or IEC) history. First point is true. Second is not. Compared to ODF, where there were two independent implementations before it was proposed to ISO, OOXML have only one, Microsoft Office, today. The best engineering attention is one you get when you create independent implementations, so the second argument fails by default. Both, however, does not guarantee the final product will be a good specification. If this is a bad specification and ISO approve it, it will be bad precedent.

Moreover, if the intense effort surrounding a standardization process is inappropriate  legally, morally or spiritually or the mere invocation of the is inappropriate, *and* the standard is allowed through, then a bad precedent would had been set.

Finally,

Matusow:This blog is not about me and Bob, it is about IBM and Microsoft having differing views on the BRM – so let’s keep the comments focused on the substantive issues rather than about the people (thx)

Sutor has this disclaimer on his blog: “The postings on this site are my own and do not necessarily represent my employer’s [IBM] positions, strategies or opinions.” While I cannot find a similar statement on Matusow’s blog, I will be wrong to assume he is talking on-behalf of Microsoft. Whatever the case, I agree there is no point talking about two person and will extend it and say there is no point framing the question as IBM vs Microsoft as well. Let’s do what Matusow say, focus on the substantive issue, not people (or company).

Advertisements

3 Comments »

  1. As for quality and length of the specification and number of comments — I’m aware of no engineering practice where we measure quality by the number of defects found. On the contrary, quality is typically measured by a prediction of the number of defects _remaining_, and this prediction is based on a model that accounts for complexity, length of testing, test coverage ratios, trend of increment defect find rates per incremental test time, etc.

    So that’s the question — how many serious defects in OOXML are remaining?

    Comment by Rob Weir — November 21, 2007 @ 6:08 pm | Reply

  2. Rob,

    In engineering, we cannot say a work is of good quality by measuring the number of defects, predicted or not. However, with sufficient number of demonstrated defects, we can say for certain a work is of bad quality.

    It may or may be not be possible to nurse a bad quality work back to health, but if someone wants to try, let them.

    Similarly no amount of comments is going to guarantee good quality work, especially if the working party choose not to address them, but side step and ignore the issues. Comments are also interesting thing. Too little of them means insufficient attention, too many is an indicator of bad quality work. One would like to receive somewhere in between the two extreme, in the region I label “just right”.

    Counting on the number of comments alone, OOXML in September days is simply unacceptable, even after taking into account the amount of duplicate comments.
    Hence, we know today that pre-September vote there is a awful lot of defects with OOXML. That was cut-off point then.

    I am assuming that in the days coming up to BRM, the appropriate people are working to reduce them. So my next cut-off point is when the BRM starts. I am a outsider, and OOXML project team is not going to give me access to their work-in-progress, which is a shame.

    It’s their duty to work on reducing the defect. It’s their duty to demonstrate, beyond reasonable doubt, that all defects reported are resolved. JTC/NBs are there to judge whether the remaining flaws/defects are serious enough to reject the proposal. Can they do it? In a totally impartial ISO cloud, I doubt it.

    If you ask me, there are several fatal defects in OOXML, any one of which if not addressed, must result rejecting OOXML.

    Comment by ctrambler — November 21, 2007 @ 6:33 pm | Reply

  3. While we don’t agree on things, I really like the balance of this post – thx for the thoughtful writing. I do have a few comments:

    First, I was certainly not attempting to “point fingers” on my post. I respect Bob and his thoughtfulness on standards work. I do think that your conclusion about Bob’s disclaimer (and I have the same one on my blog) is off target. Those disclaimers are there for legal reasons, but in no way diminishes the fact that Bob runs the team at IBM, and I am part (not being a VP) of the team at MS, doing work on doc format standards. Thus, the opinions represented in both blogs are absolutely in line with the positions of the respective companies.

    Second, I think it is a fair point you make (and Bob did) that the effect of the changes is important. I made the point that the Project Editor is working to improve the spec, but to have those improvements not make it a new spec. That is a subtle thing, but very important.

    Third, it is clear that the sheer volume of comments and dispositions will be a significant load of work for review no matter what. But, the experience gained during the considerations leading into Sept. 2 will be helpful for those looking at the dispositions. That said, the fact that there are waves of dispositions coming out early, and that the work is being grouped into logical segmentation, etc. will all help. At the end of the day – there is no requirement that a NB has to agree with every disposition of every comment they made. Each NB will look at the dispositions and state of the spec as a whole and determine if that, balanced against what is happening in the market, and what they view the future of spec to be, etc. is enough to merit a change of their vote or not. This is not another ballot – it is really meant as an opportunity for those who voted no to determine if they have seen sufficient improvement to vote yes (I know the rules allow for yes to no as well, but that was not the intent of the BRM process).

    Fourth, you have misunderstood the point I was making about ODF and its progression. I have NO PROBLEM with the fact that it was ratified as an ISO spec. My point is that following the logic presented by Bob (and others talking about the maturity of the Open XML spec), then ODF should not have been submitted to JTC 1. Clearly, that is not the case, nor is it against typical standards practice to put specs in that will significantly change over time.

    Sorry this is such a long comment.

    Jason

    Comment by Jason Matusow — November 22, 2007 @ 4:50 pm | Reply


RSS feed for comments on this post. TrackBack URI

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Create a free website or blog at WordPress.com.

%d bloggers like this: