Massachusetts vs Microsoft (Last Updated 10 Oct)

The Massachusetts Technology Leadership Council hosted a meeting in which representatives of the state discuss the state’s recent proposal to standardize on Open Document Format (The mp3/ogg is 2 hours long. A summary is available here). It helps me understand the whole issue of “Open Standard” from Massachusett’s view point. It centered on the question of “sovereignty”, i.e., the state ability to provide unencumbered access to all public documents. With this, the state decided to use only “open standard”-based document format. The state criteria had caused it to exclude Microsoft’s XML Schema. That decision is a surprise.

Hence, not surprisingly, the meeting focuses very heavily on why Microsoft’s XML Schema is not acceptable. The state’s representatives really know what they are doing and answered all questions adequately, demonstrating that they had thoroughly research the issuesbefore forming their decision. Unlike, I must say, Microsoft representatives who came with a prepared script and either cannot or will not adapt that script to argue their case in the light of changing circumstances. Most damning of all is their failure to do their homework on Open Document Format. This is demonstrated by making false claims that Open Document Format cannot store voice, image and video, and saying that no software capable of using Open Document Format. These were, of course, corrected by others immediately. However, damages to their credibility was done. (Perhaps Microsoft sent the wrong people to the meeting? [Update, David Berlind thinks so as well, more on the links section.)

They also attempted to bring in their strong suit, i.e., the “cost” of conversion. I understand why they tried this, which unfortunately, was not the issue of interest right now (how to maintain “sovereighty”?) but a procurement issue. At one stage, after Secretary Kriss and Bricklin outlined why the Microsoft’s Office XML Schema is not open (no thrid party participation in modifying it; openness restricted to reading public record [For the record, I believe Microsoft had modified it to read and write public record], forbidding partial implementation [This is a big problem reading the whole complete document is most of the time unnecessary.]), that was the trigger (suprise, surprise) for Microsoft representative to roll in the “we need to protect our intellectual property”. Most Unfortunately, he did it in a accusary tone. Although I know this is classic Microsoft redirection: introducing issues that at face value, looks to be related but cannot withstand scrutiny. However, to Secretary Kriss credits, he did not take the bait and set the record straight.

At one point, Microsoft explained that while they are a member of OASIS (the body that set the Open Document Format), they had chosen not to participate and not to use it in Office 12. Moreover, as they are in a product development push, they imply that they cannot spare time to incorporate Open Document Format. I find this is irrelevent because it is a Microsoft issue: They decide for themselves whether to participate in any body and have to live with the decision. The fact that they are in a product development push is at best an excuse to delay implementation of Open Document Format should they choose to implement it, not a valid reason for the decision not to implement it. They did point out that since Microsoft XML Schema is in XML, a converter may be already available. [The ETRC requires Open Document Format as the native save format. This is a bummer for Microsoft. They argued that they cannot store everything they in Office 12 in Open Document Format. OK, why not just engineered a “degraded” version on CD/DVD for the State Government. Afterall, the effort is minimal, simply write a converter to filter to and from Open Document Format and attached it to the “Save” menu. It is certainly much easier than creating the crippleware known as Windows Starter Edition.]

Microsoft and their associates certainly did pull all their stocks in argueing their case. However, I have the feeling that while the whole room is dicussing the issue in question, access to data collected by the government (“soveriegnty”), Microsoft is focusing on pushing its own separate agenda. It is also addressing the issue by trying to frame it their way. [Pamela Jones of Groklaw have a different view, saying they were pressuring and threatening the officials. Se links below.] Instead of understanding and addressing the issues on why their format is not accepted, they chosen to use the “why chose them, not me?” attitude and formulated their questions this way.

One good thing in the discussion is that it had not degenerated into “Massachusetts vs Microsoft” all over again (remember the Windows Monopoly anti-trust case?). Representatives of other companies and organizations back Massachusett’s view by bring in supporting information. This includes Adobe’s confirmation that the restriction on Adobe PDF is minimal (simply that you cannot call it PDF if your derivative does not satisfy interoperability criteria), IBM’s contention that it uses file format pragmatically, not religion, and that the “sovereignty” is not an issue recognized by various governments besides Massachusetts. Overall, as expected, Microsoft did not get its way. I expect the state to put a few arrows into Microsoft, but the fact that they took arrows from other companies, they came out more bruised that I thought they will.

One last thing, I want Secretary Kriss, Mr Quinn and his team to be in charge of my IT. Citizen of Massachusetts, you lucky bastards!

Good links for discussions on the topic:

Groklaw: PJ’s article
Newsforge: [What has Microsoft done for Massachusetts lately?]

Massachusetts Government:
ETRM (Enterprise Technical Reference Model) version 3.5
Comments receive for ETRM
It is unfair to suggest that Microsoft’s comments is by far the longest without explaining the context. Negative comments are usually longer because the author must argue his case. Microsoft had most to lose therefore you do expect a long reply. However, note the strange emphasis on procedural issue. PJ of Groklaw read Microsoft’s stance in the meeting as pressuring the state and threatening them with lawsuits. It is a fair interpretation, and possibly true given the emphasis on procedural issue.

One legitimate question is raised by people with disability. Using Open Document Format might seriouslly complicate their life, as screen readers cannot read Open Document Format now. The state will need to address this issue whether or not is there a American With Disabilities Act. Hopefully. Hopefully with the wide adoption of Open Document Format,
this is a short term issue.

The schedule proposed by the state is a very aggressive one, as a few people points out. The state will have to come out with a realistic schedule, such as that proposed by David Wheeler.

[Update 10 Oct]
Massachusetts has published a FAQ that addresses concerns raised by the comments above. It answers Microsoft’s complains well, besides explaining how they will try to meet the disability issue. To me, the latter is more important than Microsoft’s or any other peoples’ comments.

[Update 29 Sept]
ZDNET : Did Microsoft send the wrong guy to Massachusetts’ ODF hearing?. Berlind argues that either Steve Ballmer or Bill Gates or someone allowed to make one the spot decision should had attended. It is not without precedence that MS top exec drop in when things do not go their way (see Munich case).

[End update]

Notable Quotes:

PJ of Groklaw or maybe someone she is quoting from:
“I think a good test of the license is this: if Microsoft had to license someone else’s file format as a mandatory format in Office, and that format was only available under the same terms Microsoft is offering the Office schemas under, would Microsoft’s legal department accept the terms as-is or not?” [article]

Stephen Walli:
[On the difference between a de facto standard and a de jure standards developed in a consensus-based process] Another categorization attempts to discuss the difference between de jure standards developed in a consensus-based process and de facto standards. A more accurate statement might be de facto technology that describes a market dominant product, rather than a specification for interoperability open to all implementers. [article]

FUD == “British Computer Society: The Trouble with Open Source” ?

Thank god the article “The trouble with open source” is simply a “Member’s view” and just that. At its heart, it is an advocacy piece: Keep open source inside academic and outside software industry, and can the government please please launch an investigation on all these negative impact of Open Source to British Software Industry. The main flaw with the article is that the author had tried to frame its main arguments, the rest of the article simply suggested that the industry faces chanllenges from open source software. Hence it fails to provide compelling reasons for any government investigation.

Lets disects the four main problems with open source, unframing the arugments as we go. First, on intellectual property, the question on whether an employee have the rights to contribute the software he or she develop while in the employment of others is a question that affects ALL software professionals, and therefore, cannot be potray as an open source only problem. No matter how many laws he quote in as many jurisdiction as possible, every software professionals are governed equally. I am not a lawyer, but I am sure the court will take a common sense approach when confronted with this problem, especially it the software is developed “after hour” and not in line of the employee’s responsibitiy to his employer. His piece also ignored the growing number of software professionals employed by companies specifically to contribute to open source projects.

Second, His view point on “conceptual Integrity” is already answered by Mr Eric Raymond in the book “The cathedral and bazaar” by Eric Raymond. Moreover, it is wrong for him to suggest that “conceptual integrity” is the exclusive province of a company or a brilliant man with his team in support. There need not be a central design concept that need to be slavishly adhere to as advocated by Mr Marshall for a software to succeed: the coming together of TCP/IP, email and webserver to form the entity we know as the Internet prove my point. In fact, this rigid adherence to a blue print (design concept) can inhibits innovation/creativity. Moreover, nowadays, there is a shift to developing a structure where software can be developed piecemeal and joined together. JavaBeans, .NeT Framework and Eclipse Platform are good examples. If adherence to a central design concept is important, the “community process”, in the example of Linux Kernel Project, shows that this is not an exclusive province of company or software industry. By the way, the internet relies very heavily on open source software.

Peer review is important. By revealing their source code, small open source team get the opportunity to have their code reviewed in a way not possible with a small team. The quality of peer review depends on the expertise of the peer. Restricting source code access as in the case closed source program, can only reduce the quality of the code. In contrast, with important open source project, the code is reviewed by renowed experts in the field. One such example is of course the Linux Kernel Project.

Thirdly, some software professionals called themselves hackers because they perceive their effort in modifying small pieces of program code into a bigger software is like “hacking” the software. Every programmers therefore started life as hackers (and probably remains as hackers), as we all start by modifying softwares, a small piece at a time. Getting credits for one’s contribution is the motivation for most open source contributor. This, and the fact the visibility and readability of source code make it easy to review unprofessional practice, is a higher incentive to behave professionally. At least higher than someone who can hide behind anonymity of proprietory software.

Fourthly and finally, while money is a very good incentive for innovation. It is not the only way. Using open source code, by reusing existing source code, an open source programmer can concentrates on writting the little extra code that realize his creation and unleash his innovative potential. Open source and proprietory software are often created to do the same task, therefore they share a lot of similiarity. This, does not mean however that open source, or indeed, properitory software, is a facsilimile of the other. Moreover, open source do promote innovation, examples include the “Union Filesystem” from Knoppix and Gimp and its ScriptFu programmability, to name a few.

Mr Marshall fears of small proprietory innovative software company being driven out of business by open source. The truth is, small business fails because it cannot compete, and the competition may come from other properitory software company, especially the larger one, or the open source company. Most importantly, if these companies can continue to innovate, they have no fear.

The authors highlights the benefit open source brings to academia but believe it is not appropriate for software industry. Why shouldn’t the software industry copy or learn a thing or so from a strategy that had helped the academia? It is true that traditional business model for software does not sit well with open source, but business model do evolve. Companies that fails to evolve its business model, like companies that fails to innovate, will fall sideway. MySQL, RedHat, Novell, Trolltech, SleepyCat and SendMail are examples of a few companies that found a model for themselves. None of the company above have nonscalable business model.

He says academia requires industry to distribute the fruit of its research. While industry is one of the most easy way of distributing the fruit of research, there are other channels, such as directly from the internet. Hence, it is not the industry that needs academia or vice-versa, this is a symbiotic relationship.

UK government, like that of the state of Massatucetts, or any company for that matter, is entitled to formulate their own procurement criteria for to suit their software need. However, as public body, they need to explain their policy decision and why they vaiour one type of sovware over another. As long as their explaination is valid and is clearly in the interest of their citizens, they performed their public duty and we should let it be.

In a free economy, the government should not favour one type of business model over another but instead let the market decides. Hence, it is wrong for Mr Marshall to ask for government help to investigate and combat the so-called “negative” impact Open SOurce might pose or to build on the opportunities that OS had created (I assumed, from my impression on his article, he wants to limit the assistance for open source limitted to academia only). The government should only investigate when there is a distortion of the market, such as the abuse of monopoly power. Open Source is changing the industry, the same way proprietory program changed the industry 25 years ago. It is not a distortion of the market, until than, the government should wash its hand.

—

Are Free Software Foundation (FSF) and/or Electronic Frontier Foundation (EFF) pressure groups? Mr Marshall thinks so. I reexamined the boundary I imposed between advocacy group, pressure group and lobby group. For me, an advocacy group uses legitimate tactics including refering to the law courts, education and PR campaign to advance one’s view. It becomes lobby group when it tries to buy influence, e.g., by contributing campaign fund or outright bribery. It is a pressure group when it assert “pressure” to get its way, such as threatening to withhold something they know is important to others to impose its view. Using this definitions, FSF and EFF are just advocacy groups. This is my view, and you can use FSF and EFF’s websites (pay particular attention to their activities) to make your mind up.

links:
Slashdot discussion

Is having different edition of the same thing a good idea?

Raymond Chen’s blog on why is there many versions of Microsoft Windows (e.g., Microsoft Starter Edition, Microsoft Home Edition, Microsoft Tablet Edition. See also Ken Fisher’s description of Microsoft Vista versions) at his blog drew a lot of responses. Among them, David Berlind suggested that Microsoft take a leaf from Linux’s distribution model. In Linux, as Berlind point out, only the kernel is essential, you pick-and-mix software on top of the kernel.

I think the closest analogy in Linux for “version” is distribution. Here, we find RedHat, SuSE, Debian, Knoppix and Ubuntu as a few example of linux distributions. For the purpose of this blog lets just ignore the incompatibility between the different distributions and argue why having different distributions is valuable and how the way Microsoft does different versions do not satisfy these criterias.

We once have a PhD student who is good with Linux and build us a custom Debian server for our website from ground up. The server was truely lean and mean: No unnecessary software and I mean it. The glitch is, it is difficult to maintain the server. You cannot update the server using standard Debian tools but have to download the source code, resolve conflicts, compile and test it. Therefore it is not surprising that the server works well for us until we added more people iinto the webserver management team. No other people on the team, including myself, have the knowledge necessary nor the time to rebuild the server everytime we need to apply an update. This approach flops and the minute he left us, I reformatted the server using a standard Debian distribution. This demonstrates what is to me the most crucial decision for using a distribution: Ease of maintaining. A distribution have its own standard practices and if everyone adhere to it, it make maintenance, including upgrade, much easier.

Computers nowadays have different roles to play. A computer destined for the desktop have different hardware and software combination. Thus, having different software distributions tailors their software collection for different needs is a cheap and good way of extracting values from these computer. However, experts earn their keep by fine-tuning the computers to make it more effective according to the needs in hand. I am working happily with an unaltered Knoppix. However, for the aforementioned webserver, there is no need for a graphical user interface. It also make no sense trying to configure the keyboard properly (to enable hot-keys). Hence we chosen to build the webserver using the most basic package available, i.e. directly from a Debian boot disk. For our data analysis machines, we use a modified knoppix distribution because we need to provide a rich user environment with our own sets of software combination not found elsewhere. I could had built all three systems from the minimal Debian boot disk. However, for my personal computing need and the data analysis machines this means I have to spend a lot of time adding the necessary software. Hence, I decided to use the knoppix distribution as “seeds” to populate the computers. Hence, this demonstrate the next biggest advantage of distributions: supplying different seeds on which to build up the computer.

Most people do not use operating system distributions as-it-is, but rather customize the distribution to a certain degree. For instance, when you receive your Windows Machine from your vendor. It is not immediately usable. Rather, you have to install a few other softwares to make your machine useful to you. Now imagine a corporate/academic scenario where it is a good idea to keep 100+ computers in identifical configurations. This is the scenario we faced for our data analysis computers. Our IT management team therefore decided to create their own custom distribution, based on knoppix, to ensure the same software is deployed throughout the organization. This show the ability to create one’s own distribution is important as well.
Windows Server, I believe, has similar capabilities to replicate machines over-the-net. However, this is different from creating one’s distribution as we cannot burn the distribution into a CD and take it to another machine.

Chen also argued that only developers will want every software available for Windows to be installed on their computer. That is not true. As a developer, I only want software my users are going to install on their computers, plus some development software on top of it on my computer. Having every software available on my development computers makes my life difficult, since I can accidentally use a piece of software not available on my user’s computer.

However, as a technical geek, I do not like the idea of creating different editions to suit the business model of charging different price to different people, or to force customer to fork out more for the same software (see this comment on Chen’s blog). While this approach make business sense by carving up and/or creating artificial market, to me, this is like importing South Africa apartheid system based on how much the customer can pay. This is worse when it sucks valuable technical resources from advancing software to work on artificial barriers (crippling) existing software. One good example is Windows Home Edition where I think resources had been diverted to create the “three applications at anytime” barrier. Companies are better of differentiating their softwares by advancing the software and not creating articificial markets by carving up existing one. The latter only provide short term gain.

To summize, the value of distributions are:

1. Creating a common practice for installation. This makes managing and maintaining computers easier.
2. improve efficiencies by aggregating softwares for different purposes.
3. forming the seeding points for customization

and I believe creating distributions should not be done for the sake of satisfying an artificial business need.

In his blog, Chen argued that there is no “one-size-fit-all” and therefore there is a point for different distributions. I agree that there is no “one-size-fit-all”, their strategy only go half way to address the issues of versions/distributions, namedly the first point above and arguably, some of the second point. Taking WinXP for example, the difference between WinXP Starter Edition and WinXP Home Edition is simply an commercial exercise to create a low end market. The distriction between WinXP Home edition and Professional edition is so vague that it simply confuse the end users and frustrate sysadmin by complicating the management of computers. While Microsoft claims that one of the advantages of using Windows on servers is their familiarity to the desktop. Reading the comments from Chen’s blog shows that this might not be true for Windows 2003 Server. Rather, we could be seeing the a split for destop/server edition in the making.

Can Microsoft or proprietory software companies creating something close to that of Linux? I think so and this include incorporating licensing revenue in the distrbution as well. It is necessary, however, to dump the current distribution model. Rather than creating different distributions themselves, Microsoft should put a price tag on every bit of Windows Software, throw them into a bin, allow third party to take parts out of the bin and build their own distributions, and charge for the privilege to do so. This does not, however, means that Microsoft should not provide a selection of distributions themselves. One other advantage of this approach to Microsoft it will tell them what sells and what does not.

Will Microsoft or other proprietory software companies adopt such a model? My prediction is no and the reasons are business-related rather than technological issues.

Copyright and Photocopier, VCR, DVD etc

As any good travellers do, before I depart for my holiday in Ireland, I decided to photocopy some important information about my travel details. Although I used the office photocopier for the purpose, I do fully reimburse the University for the privilege of using the photocopier. The cost, as I put it, is 6p per page.

That got me thinking, why did it cost 6p per page? For every copy I make, the paper and toner powder together cost less than 0.7p, The rest represents the cost of maintaining the copier and operation profit for the company we outsource the maintenance of the photocopier to. I have no beef with that. However, I have a problem knowning a small sum of money, either directly from the 6p I paid, or indirectly paid by the University on my behalf, to so-called copyright licensing fee collecting agency. Why should I pay license fee to them? I owned the copyright for the information on the piece of paper I printed. I did not give them the rights to license the material on my behalf.

I am sure everyone is aware of notices placed near photocopiers talking about restriction on what you can copy. For UK universities, most of us are photocopying under license from a particular license fee collection agency. This agency shall remain anonymous in this blog because I am not against the agency’s practice, but the whole issue of collecting licensing fee where no licensing fee is owned.

First and foremost, I will like to stress that I strongly believe in just reward. Hence if you photocopy/use material copyrighted by others, you have to pay the copyright owner. Equally, nobody should be allowed to demand payment for something you have not used.

It is true that in an academic environment, we used to photocopy a lot of articles under the copyright license. However, the use of photocopiers for this purpose had been superceded by printers. [ Note. There is one exception: photocopiers in libraries. Legally, photocopiers in libraries are treated differently in copyright lawsbecause of their public role. Therefore I excluded them in this blog]. The rights to print a copy is normally covered by a different licensing scheme, i.e., subscriptions to the online library providing the copy. Photocopiers are generally relegated to printing duplicates of important documents generated in the course of business, where the ownership of the documents in most case lies with the person who do the photocopying, or the organization he is attached to. Even in the case where the copyright does not belong to the person or organization, it is highly doubtful that the copyright agencies have the right to collect royalty on behalf the copyright owner. Hence, why should we pay royalty (more like a tax) to them?

That agreement is outdated and should be renegotiated. Academic photocopier usage is getting closer to that of normal business practice. Any renegotiation should take that into account.

The actual agreement between the copyright agency, the university or the photocopy company we use is of course written in such a way that says that the agencies are collecting their dues only and nothing else. This neatly let those agency sidestep the issue of of them collecting license fee illegally. However, as everyone who use a photocopier know or should know, part of your charges goes to those agencies no matter what you are photocopying. This presumption of “photocopying licensed material” is a good compromise in the 80s and 90s, but the situation had changed and hence, no longer valid and should be changed.

Do you know that the empty VCR tapes you buy in the shop is very likely to include a royalty fee (a.k.a. tax) paid to the entertainment industry just because you might use it to copy a copyright movie? How many of us use it to copy movies? Most of us use blank tapes to record TV programs and Home Videos which we have the legal rights to do so in the first place. Copying movies using VCR is technically challenging for most of the population. Even then, not every copying is illegal. There is a case to be made for “fair use” if I simply copied the movie as a backup.

Why is this issue important? The “tax” on photocopying and empty VCR might one day make its way to all blank DVD, CDR etc that you buy. If this happens, you paid someone unnecessarily to record your home video on DVD, make a backup of your software you purchase (which you are entitled to under fair-use). Better still, you paid the wrong person when you download and burn “Free and Open Source Software” which the owner explicitly give you permission to do so.

Personally, I do not like this tax, It treat me like criminals by presuming I am guity and worst of all, do not reward my good behaviour of not copying copyrighted material where I should not.

Another case of MS Imcompetancy : Its Recruitment Letter to ESR is insulting

According to Eric Raymond’s blog, Microsoft tried to recruit him. Did this surprise me? No. Companies are always head-hunting for talented staffs and staffs that can “add value” to the company. Mr Raymond can certainly “add value” to Microsoft if he accepts the offer. However, as anyone who heard of Mr Raymond will tell you, it will surprise a significant number of people in the open source community if he jump ship. Mr Raymond is not someone who mince his words and is quite brash so his reply (on the same blog entry as the recruitment letter) will give you a good taste of what he thinks about the offer.

Microsoft should be congratulated for taking such a bold step if only … they give more thoughts into the offer and give Mr Raymond the proper respect!

I have my doubts on whether the letter is genuinely from Microsoft or people acting on behalf of the company. However, for the purpose of this post, I will assume that this is a bona fide solicitation from Microsoft. I will put aside the debate on whether Microsoft is foolish and/or arrogant because you will get these messages on Mr Raymond’s blog and it is just a matter of time before this story is discussed on SlashDot. Instead, I will focus on why I think Microsoft botched the process and did not show Mr Raymond any respect.

Firstly, did you notice that the recruiting offer letter is a generic one. While people like me expects this type of letter from any head-hunter, people the statue of Mr Raymond deserves a offer letter written specially for him. This is especially important because the headhunter needs to show that he is serious about recruiting Mr Raymond, someone who is known for his hostile view about Microsoft.

Secondly, after reading Mr Raymond’s description of what the officer said in their phone conversation and after filtering off the flowery language in Mr Raymond’s reply, I do not think the recruitment officer have any idea who Mr Raymond is. Well, If I were the recruitment officer I will explain why I sent the letter to him as soon as Mr Raymond identified himself on the phone! May be we should not be too harsh on the officer for not knowing who Mr Raymond is. There is, however, no excuse for the staff member(s) who recommended Mr Raymond’s on for not briefing him and do the necessary background research. If this indeed did not happen, that recruitment office is totally incompetent and a waste of money.

That recuitment letter sounds to me like they are not serious and simply casting a net and hope to catch some fishes in the process. All in all, this approach by MS shows their incompetancy again. The other two cases that is still fresh in my mind are the Microsoft Virtual Earch removed Apple’s HQ from the map and resurrected the World Trade Centre in New York (They used seriously out-of-date (4+ years) map) and Microsoft backtracked and finally offer to repair PlayStations that scratch CDs after initially claiming that scratches is not a problem since it did not lead to failure to read the affected CDs.

And Mr Raymond, I can see MS sending the same recruitment letter to Linus Torvalds and Richard Stallman. They are extremely likely to demonstrate the same incompetency and disrespect they show you.

P.S. A big thank you to Microsoft for giving me a good story to launch my blog.