Openness must be done and seen to be done

I am paraphrasing the famous principle and foundation of any judicial system.

Why did I bring this up. Matusow believe that there is no hanky-panky in Portugal. He says everything is above board. I am sure it is, but that does not mean some important principles had not been subverted. Modern days corruption practice are all “above board” on the surface, just looks at how much is paid to “consultants” for their dubious services in securing contracts. Hence, the question in Portugal is whether the TC or whatsoever is open enough. The argument from the other camp is an important principle for the series of meetings, “openness”, “had not be done”. Therefore, at the bare minimum, if the TC do indeed wholeheartedly believe it has openness, it must demonstrate it under the “seen to be done” principle. Merely followed the correct procedure does not ensure openness, and it is always more difficult to argue openness if it is up to the existing members to set up their own rules, rather than have the rules laid down for them from an oversight body.

The question here is why the participation of two major IT companies are rejected by the TC, especially when one is discussing document formats where both companies has the expertise to contribute, and they are major players in the opponent camp. I don’t care about the first-come-first-serve rules, because although it is generally the best practice, it is just not the correct practice here. Some places must be reserved for whoever who have a stake in the discussion and the best practice procedure is to send them the invitation and reserve places for them. Then, for the rest of us, it is first-come-first-serve.

Space limitation argument is a joke. There is definitely rooms that can accommodate more than 30 people in an institute where meetings are common. Achieving the beauty of 20 member limit while excluding big players, those that can and is prepared to try make a difference to the office application landscape (the reason for all the file format war) does not seem a justifiable aim. To permit existing members to choose who to admit looks more like a cartel than openness when it can be demonstrated that the limit of 20 members is excluding a lot of interested party, especially your competition/opponents.

Do not shut out Microsoft simply because it is Microsoft

PJ of Groklaw wrote a very passionate article about why she will not want Microsoft to participate in Open Source. For once, I have to wholeheartedly disagree with her. First and foremost, Open Source is “rule-based” and as long as people/company involved are willing to play by the rule, they will be accepted. No emotional preference, but selective exclusion of others on past behaviour, no one single person’s opinion (or a small group for that matter) counts.

OSI (Wikipedia’s description) started with a extremely strong anti-Microsoft stand. One of its founder, a passionate guy who do not mince his words, even posted the so-called Halloween documents on it.  Over the years, it serves open source well. I find it a relieve when OSI decided to move the Halloween documents elsewhere a few years ago. Why? Those documents are not compatible with the principle of open source. Open source embrace all who is willing to play by its rule.

It is my opinion that one or two of the Microsoft’s share source license will probably qualify. A few, including those highlighted by PJ, are clearly out. If they qualify, I will admit them to the hall of fame of open source. Only proviso is a name change. They cannot and should not be associated with the other Microsoft’s Share Source Licenses to prevent confusion and abuse by Microsoft. [I am not one that would like to bring Microsoft’s history  with Open Source as a reason to reject them, but I still need to give history the proper weighting]. Microsoft have to clearly separate whatever OSI approved license from the other share source license and never ever refers to the OSI-approved license as Share Source License. A gag order on Steve Ballmer’s  will be required to stop him from misrepresenting the situation to stop jeopardizing any good work done between Microsoft and Open Source.

What do you want from your file format?

Rob Weir produce yet another excellent article on what a file format is. He focussed on the word “optional” this time but the most important underlying message is what a file format means to you. More precisely, what does different file format mean to different people. Even if you are pro-OOXML and does not like Weir at all, strip away all the politicking around ODF vs OOXML, you still have to adhere to the principle that a file format meant different thing to different people.

Take me for example, my interest in file format actually fits what MS had been trumpeting in public about “interoperability”: Not to create a rival office application, but to exchange information with other applications use ODF or OOXML. Specifically, I want to use OpenOffice.org spreadsheet or MS Excel as a possible way to edit data before they are sent to my application, and my application will spit out data that my users can use in other applications (mainly statistical packages). Thus, to me, the application developer, using ODF or OOXML give me the advantage of a feature-rich editing environment for my application, and enhance my relationship with my users if I can smooth their transition to their favorite statistical package. This relieve me from user-interface tasks, and allow me to concentrate on my application, my strength. This is something we do too scarcely in academia. We concentrate on analysis (rightly), but (wrongly) ignore user-interface tasks. This is wrong because we are now in the age where we can harness other applications to do the user-interface for us. So here is my expectation as application developer: reduce my workload and to give my user better access to my software.

Do my users care about what I get from using ODF/OOXML. No. To them, they want to be able to edit data easily. It will be advantages to them if they can create input data to my application in their favorite spreadsheet application. This give them more ways than I could with my own native user interface to manipulate their data into a form that my program can accept. There will be catcha here and there, such as “you cannot embed a picture” in your spreadsheet, but overall, I am sure they appreciate the advantages of using spreadsheet programs instead of my clunky user interface. Exporting data to Spreadsheets format that their statistics program can understand save them time in format conversion and reduces the chances of human error, such as transcribing 90×12 matrix, again another thing they will appreciate. Smoothness of data transfers from other apps to my apps and to another appes is what they expects. Using ODF/OOXML will meet their expectation.

It’s true that overall, their appreciation is, unfortunately for me, is not he strength of my application vs the competition, but the smoothness my application operates. I see a lot of good academic analysis software that fails to be adopted by other academic researchers because of poor user interfaces. Partly this is the fault of the software creator, but we can remedy it.

FUD or just market advertisements?

On Microsoft Watch, Joe Wilcox lists what he believe is Microsoft’s FUD on 27 June 2007.

Reading it underscore the difficulty of defining the word FUD. They are just marketing statements/advertisements to me, full of feel good factors that most people will “see through” as marketing statements. There is no “Fear” or “Doubt” elements, some “uncertainty” because we do not have Microsoft’s internal figure/memos to back up their claim.

Mud slinging, and beware of throwing accusation at others (that might come back and hurt you) (Updated 20070831)

By way of Jason Matusow webpost, I see we now have the viewpoint from Microsoft on two issues currently burning brightly in OOXML vs ODF debate.

First, from Stephen McGibbon, who replied to Rui Seabra recollection of what happened during the coffee break in the infamous coffee break meeting. This post is difficult to read because one do have to read beyond the attacks McGibbon leveled at others, including calling groklaw croklaw and an attack on Seabra’s conduct at the said Portuguese meeting. Those were unnecessary and if anything, can be interpreted as McGibbon is angry about it, or worse, cannot tolerate negative comments. My initial reading of Seabra recollection of the “coffee break” moment was “nothing new” and “nothing worth any attention”. Since McGibbon kick up a fuss on it, I reread it and still conclude the same way. I cannot understand why he indirectly call Seabra a hypocrite and why he believe Seabra was speaking “for the community” (As in Free Software community). He is speaking for community and it is ANSOL, the body he is representing in the meeting.

On the subject of Portugese meeting, it still leave the question on why the chairperson choose to exclude Sun and IBM. unexplained. Let’s make it clear that McGibbon, being simply an invitee to the meeting, rightly did not comment of the conduct of the chairperson. However, that is probably one of the biggest issue that needs explaining for the meeting.

There are two more things that I choose to connect with a relationship to in McGibbon’s posting. I will do it in reverse order the way it was presented. First, he claims Sun supports OOXML. I do not believe so. He based his response on Sun’s email clarifying Sun’s position on V1 voting in the state. Please read the whole email before forming your own opinion. I believe Sun’s response of “Conditional Approval (Disapprove with Comments)” are simply technical speak where approval does not really mean approval in the box-standard English, and for that matter, neither is Disapprove for that matter. To me, the two operative paragraphs are:

"This change is important to Sun in ensuring that our position is
correctly represented.  We wish to make it completely clear that
we support DIS 29500 becoming an ISO Standard and are in complete
agreement with its stated purposes of enabling interoperability
among different implementations and providing interoperable access
to the legacy of Microsoft Office documents.

Sun voted No on Approval because it is our expert finding, based
on the analysis so far accomplished in V1, that DIS 29500 as
presently written is technically incapable of achieving those
goals, not because we disagree with the goals or are opposed to an
ISO Standard that would enable them.  Sun voted Yes on Conditional
approval (disapproval with comments) because this is the only one
of the options we were given that would guarantee that the
specific changes already agreed upon by consensus in V1 would
actually be implemented."

Omitting the second paragraph, the way McGibbon do, does not paint the full picture. Opening him to accusation of selecting the truth. (Update 20070831: I belatedly found this pdf from Sun’s Bosak clarifying SUN’s position)

I am linking this to McGibbon’s claim of what Gary Edward said about Sun earlier in the post. I do not have any way to verify that Gary Edward’s original post is what McGibbon says it is. Hence, I have to evaluate the claim based on other circumstantial evidence. Reading the “updated post” does show Edward was rather unfriendly and a bit angry at Sun for something. This lend credibility to McGibbon’s claim. Unfortunately, given that McGibbon omitted the second paragraph above, I have to reevaluate the credibility issue.

And when you throw mud at people, make sure you know what you do. Accusing IBM of not supporting ODF as the default native format on ODF is Lotus Notes 8 is one thing, not realizing the default file format is actually the precursor of ODF and which is the file ODF is build on is another.

Second, Doug McHugh article on “Spinning the ISO process in the US“. When accusing Andy Updegrove of spin, the post give the impression that he attended the meeting. Which, on interrogation by Updegrove in the comments section, turns out to be not true. Mud boomeranged again.

Several Interesting News from the cyber graphvine

No. I am not starting a daily/weekly/yearly news digests. However, these are several developments over the past few days that deserve some attention, but not sufficient for a full-blown blog posting each:

• Sun exec accuse Microsoft of patent terrorism : It’s not terrorism, it is old fashion extortion. The word “terrorism” is overused and, in a lot of case, abused. In this case, I can see why the scale of extortion can lead to accusation of terrorism, but let’s not glorify this. Giving it more column inches with headline like this simply help Microsoft to spread its extortion message.
• HP secure a 5000 seats Ubuntu Deal: The news itself is interesting. As I recall this is the first public announcement of Ubuntu deployment at this scale. A triumph for Ubuntu and it shows that HP is willing to support Linux if nudged properly. The most interesting part is the comment that Canonical would like to keep deals like this under wraps until it is signed, just in case Microsoft comes out at the 11th hour to snatch it away. Microsoft take notes: your “11th hour strategy” works. IT managers take note: Even if you do not have intention to go Linux, having a serious looking migration plan to it can get Microsoft to slash price dramatically at the “11th hour”.Another interesting part which is what was not mentioned, i.e., it is not surprising that Canonical wants to keep this type of deals under wraps, but the customers also agree to keep it under wraps. This means one useful judgment of whether someone is serious about migration is to see whether they tell Microsoft about it. Using this yardstick, and believing the spin that Canonical indeed have several more deals like this, Canonical, and by inference, Linux on desktop, has serious customers on board. We may be witnessing the true dawn of Linux on Desktop.
• Skype has to include the text of GPL in its VoIP phone package in Germany. This news itself is mundane. It is rather stupid for Skype to believe that throwing in a small piece of papers with links to website is suffice. The interesting part is the court rules that having access to source code on line is insufficient in the case where hardware incorporating the software is delivered. There is no detail on how the court believe the access to source code problem for hardware can be resolved, at least in Germany. Asking Skype to include the source code in the package itself is problematic, as it means at times it may be forced to include another optical disk in the package and this drive up cost, lowering open source competitiveness compared to proprietary software. The good news is, however, that this just means less profit for Skype, but still larger than that can be achieved with proprietary software. I think it will resort to Skype agreeing to send, by post, the source code to any customer who requested one.

Put out the evidence and let the readers decide.

Matusow finally responsed to insinuation that Microsoft is ballot-stuffing. I got a feeling he is trying to play a word game here. I mean the stuff you expects PR people will do.

First is where he drawn the line on vote-stuffing

“I think one situation I would disagree with is if one party had wholly or partially-owned subsidiaries brought onto committees to vote as a block – that would seem wrong to me.”

[Only would, meaning he is only 99% sure that this is wrong? OK. This may be my bad English at fault. I’ll give my English the benefit of doubt and say he says this practice is 100% wrong.] This is  undeniably wrong. Like vote-buying, so obvious that only the arrogant company will do it and expects to get away with it. All giant IT companies, including Microsoft, will NOT engage in such practices. If they do ballot-stuffing, it will be more subtle.

The important omission in this statement, to me, is to say that if a company brought its business partner onto committees to vote as a bloc, this is not wrong. This I whole-heartedly disagree. It achieve the same thing and is therefore equally wrong. In fact, this is the more likely situation for ballot-stuffing. More difficult to discover and easier to deny. I mean, in all cases, including this, the best the injured party can do is to insinuate, isn’t it?

It is however equally wrong to say that the reason for the explosion of membership in one committee when it is handling important and/or controversial issue is a clear sign of ballot-stuffing, even if the new members voted en-bloc. As I said tirelessly in several blogs and this, one expects the parties who might be injured by an adverse decision to take action to protect itself. As such, one should see a natural, grassroot-based surge when someone feels threaten. In OOXML approval case, adoption of OOXML by ISO is important for Microsoft’s partners who had invested a lot in OOXML. Hence I do expect a surge. The only question is, is the surge a bit too big to be “natural”? It’s a question for the reader.

Even if these parties fade away after this one vote, it will be wrong to say it is vote-stuffing base on this fact alone. It just means they are only interested in OOXML. If they contribute to the discussion, their participation will be well-deserved and they should be welcomed. If they did not, then put more weigh on vote-stuffing.

And let technical committee stays technical only, business committee stays business only. Any party with business agenda infiltrating technical committee is and vice-versa is vote-stuffing, as it does not have any business there.

Is Microsoft doing ballot-stuffing? You decide. I can only presents the evidence and what I think, but it’s your call.

Back to wordsmithing by Matusow. He opined that all opinions, even those expressed using the same letter counts because the sender took the time to do it. Yes, all opinions should count. However, he miss out the part discussing whether all opinion should be given equal weights. I think this is a more important issue than to simply says that all opinions should count.

Possible hanky-panky in South Africa?

PJ is one of the more professional blogger in the cyberspace. I’m not. So while she was careful with her wording when she express that she thinks she spotted that Microsoft will try to overturn the results of recent South African No vote on OOXML from Jason Matusow’s blog post, I do not have such quirm. I call it possible hanky-panky. May be we see another strategic positioning, as happened in US.

South African nut will be harder to crack. The committee voted no. The body who commission the committee will have to answer why it did not follow its recommendation. The bar is higher now. But perhaps Andy Updegrove has the best reason why South Africa will not change its mind: They had long indicated that they are going to vote No by default on the Fast Track process. This means that they already set the bar high for themselves to justify a yes vote. OOXML is not one of the stronger case for the Fast Track process.

Good question from Dr. Nagarjuna

By way of Groklaw, I came across an exchange between Dr Nagarjuna of FSF India and Mr Kapur of Microsoft India. I find the exchange interesting because it shatter one of seriously wrong assumption I have with OOXML and I am pretty sure I am not alone.

The assumption is “I will be able to convert from existing Microsoft binary formats to OOXML **MYSELF**”. The key operative word is “myself”. I realize that I cannot do it myself but have to run back to Microsoft. Even if OpenOffice.org provide a converter, it will be based on years on reverse-engineering rather than Microsoft truely open up the format. Without this, a lot of promise of OOXML that I once thought will surface just vanished. One of which is you still have to go and buy Microsoft Office or someone who is willing to sign an NDA with Microsoft to convert your legacy document to OOXML first. This means, one way or another, you got to pay since open source people are generally adverse to signing an NDA.

Even if you get one-to-one mapping, there is still another important hurdle. Without legacy version of MS Office, you cannot implement so-called “optional” tags like “autoSpaceLikeWord95” that requires you to refer to legacy application.

Mr Kapur said that “Implementing backward compatibility is an application function not a file format specification requirement.” Since he says backward compatibility is not a requirement, then we have to insist on removing ALL legacy looking tags such as “autoSpaceLikeWord95” and throw out VML as they are backward compatibility issues and made it to the specification. To a  certain extent, the 1900/1904 date system will have to be throw out as well as they are for backward compatibility and including them will make sure the problem will not go away.

What Mr Kapur cannot do is to choose when to put backward compatibility in. I am going to be generous, however, by allowing him to choose one more time.

Good strategic move…. (Updated 20070723)

A small town crook like me trying to speculate on why Microsoft and Co decided to vote the way it did is always futile. I knew it before I started. But it was an interesting exercise nonetheless.

Not surprisingly, Microsoft was first to announce that INCITS Executive Board decided to put out a ballot on “Approval, with comments”. This is the continuation of the saga of V1 committee. If you believe the conspiracy theory, which I do, you must give your hat up to the victor, Microsoft. They achieved their objective of “Approval, with comments”.

Its a good move and finally explained why “Abstention, with comments” was not accepted by the OOXML bloc. Why settle for second best when you can create the climate for “Approval, with comments” (and get it)? Plan A was definitely for V1 committee to recommend “Approval, with comments” to INCITS, but is not likely to work. Plan B is to create the climate for “Approval, with comments” to be adopted by INCITS. To achieve this, make sure V1 cannot form a consensus as indicated by a 2/3 majority. Coz if it does, INCITS will adopt V1’s concensus and iNCITS will have a lot of explaining to do if it don’t. However, if there is no consensus, INCITS will have to make its own decision. Besides all the lobbying effort that can be done, one important factor is that INCITS will, in principle, follow the majority vote. Otherwise it will have a lot of explaining to do.

So, sadly for me, although in principle, 2/3 majority voting works at V1 level to prevent vote-stuffing. In practice, vote-stuffing works.

Although it is not yet check-mate, it is extremely close to it. Good move Microsoft! (and I do not meant it sarcastically).

[Update 20070723

Andy Updegrove gave some insights into the INCITS meeting. He characterised MS motion for a vote to send out “Yes, with comments” as unexpected needs some more clarification. Given that V1 failed to reach a consensus, one would expects INCITS to decide on the wording of th ballot paper they have to send out. I am going to take it as true that such a motion is not on the agenda beforehand and that there is nothing on the agenda about the recent v1 decision, but it should not come as a complete surprise. It is possible that a motion for vote to send out “No, with comments” was on the cards of one or more of  the attendees and they were simply a bit too slow in putting the motion in. Equally possible is that the decision to send out ballot paper is going to be on the agenda for the next meeting. We just don’t know.]

]