That has become the essence of Microsoft’s opposition to the UK government’s open standards policy, according to Steve Mutkoski, worldwide policy director of international government affairs at Microsoft. You can’t put your finger on it. And what’n’ever it is, it ain’t what the open standards movement says it is.
It wasn’t always so. How we got talking about this, in the atrium of Microsoft’s swish London headquarters, was that the software giant had used a curious tactic to defeat UK open standards policy. Computer Weekly was trying to shed some light on it.
When Microsoft, the open movement’s most powerful enemy, saw the UK hoist the open standards banner it did not try to have it pulled down. It tried to have the colours changed.
This is what the public consultation that closes Monday is all about. Microsoft and its cohorts in the proprietary software business tried to persuade the UK to their wording when it codifies open standards in official policy. The government asked for public support: is an open standard what we say it is or what Microsoft says it is?
Microsoft had used a dastardly ingenious tactic: steal your enemy’s language of identity and you render him powerless. The open movement had evolved its own meanings for the words it used to define itself. ‘Open standard’ was the most important and most powerful, particularly since governments had started trying to put it on official headed paper. Its meaning had been forged in opposition to the proprietary software model. It and UK policy had been defined in explicit opposition to Microsoft’s business practices. The movement had defined its own terms. Now Microsoft and its cohorts in the International Standards Organisation and Business Software Alliance wanted them rewritten on their own terms. They sought to impose their own meanings on their enemies’ language, and when it was written on official paper their enemies would be metamorphosed into their own likeness. It was voodoo magic.
That was however no reason to dismiss the proprietary lobby outright. It had been so persuasive in private meetings and letters that the government pulled its first official definition of an open standard. If Microsoft had given a good reason it deserved to be exposed to the light of day.
So far we only had the what, not the why. Leaked correspondence revealed Microsoft and its partners had insisted patent holders should be permitted to stake royalty claims over open standards just as they did over every other sort of standard. They claimed the right to label proprietary, patented standards as open standards. It seemed nonsensical. There had to be a reason why.
Computer Weekly got its chance to find out when Microsoft’s director of policy flew in from Seattle on one of those unusually sunny mornings in March. Mutkoski had come to personally lead the lobbying. Neither of us knew then how the lobbying was to backfire. But the important questions would still need answering.
So I asked the most pertinent of them. How could it be justified for a patent holder to stake a claim over a standard of software interoperability?
“I can explain that you,” said Mutkoski. “But I will also start out with a proposition. Why should it not be the case?”
It was perhaps too much to expect a straightforward answer. This was what in debating circles was known as shifting the burden of proof. The whole point of the interview was to question Microsoft’s assertions about open standards. Mutkoski was trying to shift the burden on me to prove the opposite of his case. That wasn’t my job, or my cause.
The opposite case, for open standards, had already been made by the coalition government, and other governments before it. They made it openly. It seemed valid. They promised liberation in a world where public computer systems were hog-tied by Microsoft’s desktop software monopoly and squished under the weight of Oracle’s market power. Microsoft might be the only official monopolist in the room. But 70 per cent of all software bought by government was Oracle. This much was fact.
I put the question again. How did Microsoft justify its attempt to impose royalty claims on open standards?
The answer, said Mutkoski, was that the software industry supported a variety of business models. One of those business models involved people selling proprietary software licences. Those people should be included in the open standards definition.
This was still no reason why. And it was again nonsensical. The point of classifying something is to distinguish it. Create a subset of standards called open standards. Allow all varieties of software standard to be thrown into it and you no longer have a distinct subset. You have a hotch-potch.
But Microsoft might have a valid case waiting to be put cogently. Could he give an example that demonstrated why such royalty claims should be permissible?
“Here’s why,” said Mutkoski. “You are a products company. You are looking to create great, innovative products that sell. You can either do the R&D yourself, which is going to cost money, or you could license the technology I contribute. I need to get a return on my R&D. I think that’s the paradigmatic why.”
And it was a why. But it was the why of the proprietary software licensing model. There was still no sensible reason why this model should be imposed on the open movement.
I put it to Mutkoski again. For what reason should anyone be permitted to claim royalties over a standard of software interoperability?
He summoned for his answer the example of those telecoms and electronics-derived media standards that were later to be barred from the UK consultation: mp3, H.264, GSM and 3G. His case was that these standards had been successful despite being bound in hardware-derived patents. He denied there was any reason to distinguishing between hardware and software. His said that in both cases it was simply valid for someone to claim royalties over a standard.
Yes, but why? What, say, are the H.264 royalty claims for? If we knew that then perhaps we could at least get the heft of a justification, if not the sense.
“I’d have to look there. There’s no way I can answer that even in an hour. These are areas that are so complex.”
“Do you know what the patent claims are for in H.264?”
“I know what some of them are”.
“Can you give me an example?”
“Not off the top of my head”.
“This is crucial. This is what I want to understand.”
“Maybe you need to talk to a patent lawyer. I’m not involved in any of the organisations well enough to answer the kinds of questions you have about specific patents.”
“Well then how can you say for sure that it’s desirable to allow royalty claims over standards if you don’t know what the basis of your argument is?”
“I think you are asking me to prove a negative. I would flip the question round and say to people who don’t want to recognise royalties why they think it’s desirable to not have royalties.”
He was trying to shift the burden again. And anyway, Mutkoski was a lawyer. He’d been a software licensing counsel and attorney at Microsoft for years. I wasn’t asking you to prove a negative, I said. I was asking you to prove your assertion. You are asking me to a negative, I said.
We had reached a dead-end in Microsoft’s argument: the proprietary model should be imposed on the open model because the proprietary model. It was argumentum ad morantium. The Cabinet Office had fallen for this.
Mutkoski denied it was so. Innovators must simply be permitted to claim rewards for their efforts, he said. That meant claiming royalties over standards as well as the software implementations that used them. Convince me, I said. You convince me, he said. You can’t prove your case, I said. You’re tying to shift the burden he said. We were going round in circles, flirting around the software patent debate. That is what it was about for Mutkoski – the legitimacy of software patents.
We were having this silly Dervish debate because Microsoft and its cohorts had protested over the UK’s attempt to define an open standard as something distinct from proprietary standards. This opposition seemed fatuous as well. If an open standard was distinguished by its not being proprietary then it would by definition not entertain proprietary claims. If its distinction was not that it was not proprietary then it was pointless distinguishing it all: just call it a standard and stop wasting our time.
Yet Microsoft, Oracle, the Business Software Alliance, the International Standards Organisation and the British Standards Institution all told the UK that proprietary standards ought to be included amongst those it called open. There was by their reckoning nothing to distinguish proprietary and open. Microsoft had even mocked up a Cabinet Office Procurement Policy Note, in official wording, that included an open standards definition with a proprietary clause stuffed in it. Standards constituted of proprietary licence terms known as FRAND were, it insisted, open.
When Computer Weekly asked these organisations to release the open standards definitions they had been pushing on the UK, they all refused. They weren’t all that open themselves. The information got out anyway in a leak and a Freedom of Information request. When CW pushed them to justify their position, they either withdrew it or, as did Oracle, refused to speak. As has already been written, ISO said it did not even know what an open standard was, though it had been telling government how it should write its definition.
What has not been revealed before now is that BSI went further in distancing itself from the whole affair.
David Bell, BSI head of policy, told Computer Weekly: “We don’t talk about open standards because it’s not a concept that we – it’s not part of our terminology. ‘Open standards’ is just not part of the vocabulary we use.”
Bell’s was the only sensible contribution the proprietary camp had made to the whole debate. Perhaps Mutkoski would make a similar climbdown when asked to clarify Microsoft’s definition of an open standard.
He didn’t quite climb down. He transmogrified. There was now no way to define an open standard for certain. It meant different things to different people. “From my experience,” said Mutkoski, “its a spectrum”.
Spectrums are what psychologists use to identify ambiguous mental conditions like autism. The human mind is so complex it can be hard to determine whether someone conforms to something so simplistic as a label. So psychologists determine how someone’s behaviour maps to a spectrum of autistic traits. This is how Mutkoski wanted the UK to classify open standards. He reckoned they had five traits. Four of them concerned the way an issuing standards body dealt with them: how overhead costs where covered, that sort of thing. One concerned the actual substance of the standard: and only then to specify that it could be proprietary.
Mutkoski’s point was you couldn’t put your finger on it. Even if Microsoft itself had a proprietary claim over an open standard, it might still be deemed open on the strength of its being formed in an open forum where the voting was arranged equitably and there was a liberal supply of cheese biscuits.
But who in this ambiguous world would determine which standards should be classified open and which proprietary, I asked. If every standard should be treated as an individual case, this spectrum approach implied an unreliable procession of professional assessments, tests and second opinions that might produce unsatisfactory diagnoses.
Let the market decide, said Mutkoski. And when the market failed? The market would correct itself. The standards ecosystem was sophisticated. It was self-correcting.
What went unspoken between us was the unsavoury example of the market failure exemplified in Microsoft’s own .doc format, or Oracle’s idiosyncratic implementation of Standard SQL. Those market failures could only be seen at a greater perspective than that conceivable by the standards ecosystem itself. The market failures had not corrected themselves. Government had decided to intervene. And even then only to spend its own money more wisely.
The spectrum seemed just another dastardly way to pull the rug out from the open standards movement anyway. A software standard is not a complex system like the human brain. It’s not hard to put your finger on it. If you decide that an open standard is an amorphous concept and let anyone define it how they want, you make it impossible to say what it is for sure. It’s then whatever you want it to be. Policy formed in this way would have the structural integrity of a freshly laid cow pat, or the semantic certainty of gobbledygook – at least for as long as it took the proprietary camp to have established in practice that all open standards were now proprietary.
To be fair to Mutkoski, there might still have been something he was failing to get across. He was better at expressing his ideas on paper, he said. He gave me a copy of a paper he had written about open standa rds: ‘Defining Open Standards: A Comparison of Policy and Practice’.
The paper starts out by misrepresenting a famous paper by a Yale law professor on the “I know it when I see it” hardcore pornography ruling of the US Supreme Court, as though to give his case against open standards the appearance of authority.
Nevertheless, Mutkoski’s proposition is that open standards have a fundamental problem: they rely for evidence of their own integrity on nothing but self-referential confidence. The problem with his critique is that it denies the open movement the right to self-determination. (Also ironically for a software patent expert, the paper he misrepresented is about freedom of speech). His analysis reduces the open standards policy efforts of the UK, Europe and India into a dissembling knot of gripes and picked hairs.
It’s hard to put your finger on what an open standard is because different organisations have defined it in different ways, the paper says. What it does not say is this is most true of those who drafted open standards definitions after capitulating to the proprietary lobby. It’s like saying laws against embezzlement are unworkable because some countries have loopholes in which gangsters operate freely. Just ask any gangster, he’ll tell you its unworkable.
This left just one question for the proprietary camp. If an open standard was one that included proprietary licence terms, what was a non-open standard? Surely there would be nothing to distinguish them?
“I think that most people would probably agree that a non-open standard is a specification created by a single company, held within the company and not shared with anyone else, not available for licensing, not available for implementation,” said Mutkoski.
But that wouldn’t be a standard at all, I said. “It could be though,” said Mutkoski. Ah so its whatever you say it is. I know it when I smell it.