The UK Patent Office is aware that patents *may* be damaging to pure software, but believes that the "technical effect" rule will prevent this. Unfortunately they were utterly unconvincing when pressed for detail.
-- Even more unconvincing if you look at some of the patents the UK PO has granted.
I got a reply.
It was a very patronising "we don't think it will pass and even if it does, we're happy with the wording". Completely ignored my points.
As per the other post on this thread, it would be good if you could send me the full details, to j.heald (at) ffii.org.uk -- or if you could put them up on a website somewhere. Thanks!
Not good at all. The text of Lib Dem party policy is actually supposed to be a position which is very skeptical of software patents.
If you could scan or transcribe the letter, and send it to j.heald (at) ffii.org.uk, it would be really useful to identify exactly whose briefings Chris Huhne's assistant is basing her reply on.
She got back to me a few days ago saying that there is no risk from these software patents... and that the reporting about them is all "wrong".
Which party ? This would be really important to know.
Patent opposition procedures are no silver bullet
on
How to Fix U.S. Patents
·
· Score: 4, Interesting
According to the article:
First, create incentives and opportunities for parties to challenge the novelty and nonobviousness of an invention before the PTO grants a patent.
A lot of people in the States seem to think that European-style opposition procedures, where private companies can challenge patent applications before the PTO grants them, are some sort of magic solution to the patent mess.
They aren't, at least not by themseves.
Opposition procedures can help, but Europe still grants its share of daft patents.
More worryingly, the number of oppositions at the EPO has been steadily falling over the last ten years, though there is no evidence that EPO quality is improving.
Instead, companies seem to be deciding that it's simply not cost-effective to put in the resources to do the EPO's job for it. If you're the size of Canon(Europe) for example (who I've heard this argument from), you've got a pretty good arsenal of your own patents you can hope to counter-sue or cross-licence with, and if the bad patent does come to court, you have the resources to fight it at that stage.
The people the worst patents really impact are SMEs, who have to settle, because they can't afford to fight them.
oliverthered wrote (calculating votes by parties):
there are 278 allied for patents.
and 269+ allied against.
Actually, it's more complicated than that.
At the first reading debate, in all three of the largest parties (Conservative bloc, Socialists and Liberals) split on the directive; with the only the smaller ones (Greens, Nordic Greens/Communists, and Euroskeptics) solidly anti-swpat.
In fact, the "official" line of all three biggest parties was substantially in favour of the Commission proposal (ie pro-swpat).
The vote only went the anti-swpat way it did because so many individual national party groups inside the big blocs, and so many single MEPs came to understand, individual by individual, just why the Commission proposal was so deceptive, and is such a trojan horse for all-out software patenting.
This time, both the Socialists and the Conservatives have chosen MEPs as their official experts who were "swpat-skeptical" last time.
But, again, every national delegation goes its own way; and then there are individual MEPs who do their own thing: the major political blocs tend not to vote solidly along party lines.
So every single MEP contact matters.
And the pro-swpat people have stepped up their lobbying like you wouldn't believe.
Latest story we have heard is that there are now a couple of Microsoft lobbyists working the entire parliament, trying to speak to every single MEP or assistant, a bit like Jehovah Witnesses...
The parliament is our best chance, to put over a coherent swpat-skeptical position. But it is critical for everybody to contact their local MEPs, to make sure that they understand the issues, and haven't been confused by the other side.
The Council of Ministers' first reading text had been scheduled for fast-track approval before the end of the year, probably by Agriculture and Fisheries ministers.
Infoworld's report is claiming that the Polish decision may still be negated, if Belgium changes its vote.
According to Mark MacGann of EICTA:
"In May, Belgium voted to abstain, and though I cannot speak for the Belgium government, we have been extremely encouraged by meetings we have had with officials in Belgium and are cautiously optimistic that they may change their vote to yes," MacGann said.
Should Belgium decide to approve the Council's version of the directive, Poland's change of heart would not keep the "Patentability of Computer-implemented Inventions" from being formally approved and the legislation would then move to a second reading in the Parliament, where the contentious debate would continue, MacGann said.
Classically, every system can always be viewed as being in one microstate. Then there is no such thing as entropy. Obviously, that would be a confused and useless view.
On the contrary, it's a most valuable view, and very helpful for seeing why unitarity and/or determinism is fundamental to the Second Law, not in opposition to it.
It reminds us always to remember that the entropy is not a property of the universe itself, but rather it is a property of the description of the universe -- coarse-grained and inevitably simplified -- that we have chosen to adopt.
So, in the simplest terms, we think of the universe evolving from one of a set of initial microstates M1 through a complicated black-box operation to one of a set of subsequent microstates M2.
Because of determinism, each initial state in M1 evolves to exactly one subsequent state in M2.
But our description of the initial state -- in terms of macroscopic variables &c -- is not sufficient to identify the microstate. Our description is missing some of the information, and this is the entropy S1.
If we could perfectly map our whole initial distribution of possible states through the black box, microstate by microstate, then our final entropy would still be exactly S1, reflecting the deterministic evolution of that initial distribution of states.
But inevitably we can't follow all of the shuffling in the black box in that detail, so some of our initial information ceases to be useful -- with the result that at the end of the process there is more information we are missing, so S2 >= S1.
So the Second Law inequality rests on two things: the total amount of information there is to know remains the same (because of the determinism); but the amount of useful information we actually have has fallen (because we couldn't follow the shuffling) -- and that is why the difference between the two, the entropy, the information we don't have, has increased (or at best remained the same).
The second law does not conflict with the assumption of determinism: it depends on it.
This carries over directly to quantum mechanics, where the meaning of unitarity is essentially a guarantee that volumes in the phase space are preserved -- a grid of microstates maps forward to another grid of microstates the same size.
Again, this does not conflict with the second law; it guarantees it.
In terms of the accounting, it's very important that the microstate of the Hawking radiation does represent information about the state of the universe, but information that we don't have.
Black holes radiate (no pun intended) a black-body spectrum, which is a spectrum of maximal entropy. This had been proven several different ways by the mid-seventies. If black holes destroyed information, which radiation, containing no information, would be the end of the story.
Um, no.
Maximal entropy = maximum number of corresponding microstates.
The universe is in just one of those microstates, not any of the others, so in selecting that microstate the Hawking radiation does actually represent an real flow of information.
If this is enough to guarantee that the Second Law of thermodynamics is obeyed, as the previous poster suggested, ie that
Entropy rate of the Hawking radiation + change in entropy of the black hole > all the entropy of particles falling into the black hole
then there's no really fundamental reason why the whole thing shouldn't be compatible with a more fine-detailed, deterministic quantum description for the whole process.
Can anyone here confirm that second-law inequality ?
The European Commission (ie the Executive branch) is not volunteering to suspend its antitrust sanctions for the full several years' duration of the Micosoft appeal process.
What the Commission has offered is to suspend the sanctions only for a few months until the Court rules on whether they are acceptable as "interim measures".
The European Court of Justice is expected to take several years to decide on the Microsoft appeal as a whole.
But the ECJ ruling on appropriate "interim measures" is expected much sooner, literally within months.
Richard Stallman will be giving one hour talk this Friday on software patents
The talk will then be followed by round table comments and discussion by Euro-candidates from all of the political parties.
Come along, and tell them what you think.
Richard Stallman
"The Dangers of Software Patents"
Friday 21 May, 6pm
Cruciform Building, Lecture Theatre #1,
University College London,
Gower Street.
The event is free, and all are welcome.
It's a 300 seater lecture theatre, and this has had to be organised at the last minute, so help us get the word out. Let's show the candidates that swpat is something we really care about.
The original German proposal was to add the words shown in bold into Article 2(b):
2b. A technical contribution means a contribution to the state of the art in a field of technology which is
new and not obvious to a person skilled in the art. The technical contribution shall be assessed by consideration of the difference between the state of the art and the scope of the patent claim considered as a whole, which must comprise technical features, irrespective of whether these are accompanied by non-technical features, whereby the technical features must predominate. The use of natural forces to control physical effects beyond the digital representation of information belongs to a technical field. The mere processing, handling, and presentation of information do not belong to a technical field, even where technical devices are employed for such purposes.
The Commission "compromise" on Art 2(b) was to delete everything from "whereby" onwards, leaving just the addition "new and".
But apparently this was enough to win over Germany, Poland and Latvia, which was enough to prevent the whole Irish draft being sent back for the re-write it deserved.
The story also has an article in the Economist this week.
The article mentions an interesting theory, that instead of an external meteorite triggering mass eruptions, it might be the volcanic eruptions that came first. The eruptions were powerful enough to fire a great gob of rock into space, and each big crater is where it re-impacted.
On this view the eruptions would be the prime cause of the mass extinctions - at Permian, Cretaceous and Triassic - and the impact craters just a side effect.
IBM has publicly expressed its disapproval of software patents, citing, among other things the cost of litigation.
No.
IBM Europe make it quite clear in its official response to consultations on the patentability of software in the EU, that it supports the patentability of software, and wants to see the software patents upheld, that have been controversially granted by the European Patent Office:
Patents abound in all areas of technology regardless of whether the technology is implemented using computer programs or not. Innovation in technology cannot be differentiated on the basis of whether computer programs form part of the invention. To adopt a more restrictive approach to patenting would prove a disincentive to innovation in almost all fields of technology to the detriment of research and development in Europe.
But IBM Europe thinks that things have gone too far in the United States, particularly as regards the patenting of business methods:
We argue that to require no more than a "useful, concrete and tangible result" in the broad sense currently being applied in the USA invites the patenting of ideas that may have been visualised as desirable but have no foundation in terms of the research or development to turn them into practical reality. IBM spends very substantial sums to develop products for the marketplace and must ensure the supply of products that are reliable, cost effective and meet the needs of the customer.
The quid pro quo in the grant of patents is the disclosure of technology that rests on the research and development activity generating the inventions. The rationale that applies for protecting technological innovation by patents is therefore absent for those business methods where no such technical contribution is made. It is important that the level of protection granted by a patent is commensurate with the technical contribution the inventor has made to the art. The danger of opening the door to the unrestricted patenting of business methods is that patents may be granted that foreclose business ideas with no requirement to disclose the technology that makes them practicable. Thus, whilst IBM supports the patenting of computer program implemented inventions based on technical innovation, we see no benefit to commercial activity in Europe from the patenting of commerce itself.
What Kodak is saying is that they have developed a novel means to do the HOW. That there are other means to do what WHAT is irrelevant here.
So you cannot say that CORBA is prior art. You might be able to say that a specific implementation of CORBA is. But then again, without going deep into the details of the HOW, you might not.
Below is Claim 1 of patent 5,206,951 filed August 31, 1992, a continuation of an application filed Aug. 21, 1987.
As you can see, it's pretty generalised stuff, not about low-level details of the implementation at all.
Did a draft of the CORBA standard, before that date, mandate the ability for object managers to be able to carry such an operation ?
If so, then it's prior art.
CLAIMS
1. A data processing system in which data is represented as typed objects, the system comprising:
(A) a plurality of object managers for performing operations with respect to objects, each object manager including
(a) means for performing at least one operation with respect to at least one corresponding type of object,
the means for performing at least one operation being responsive to a request to perform an operation of the at least one operation with respect to an identified object of the corresponding type for performing the requested operation with respect to the identified object,
(b) means for identifying in a first, corresponding type of object a reference to a second object,
(c) means responsive to the identification of a reference to a second object for generating a request for an operation with respect to the second object,
each request including an identification of the second object and an operation to be performed with respect to the second object,
(B) means for receiving from a requesting object manager an identification of a second object and an identification of an operation to be performed with respect to the second object,
(C) means for using the received object and operation identifications to identify an object manager that includes means for performing the identified operation on objects of the type of the identified second object, and
(D) means for invoking the identified object manager, and
(E) means for communicating to the identified object manager a request to perform the identified operation on the identified second object,
wherein an object manager can both request invocation of other object managers and can itself be invoked by other object managers.
You cannot patent an idea. You can only patent the IMPLEMENTATION of an idea. The title and summary of the patent describe what it does. The body of the patent describes how it is done. The HOW is what matters in a patent.
Well, strictly speaking the CLAIMS are what matter in a patent.
The author has to give a description of one way what is claimed could be implemented, that is good enough that somebody with normal skill in the art could create a working version of the invention ("sufficient disclosure").
But what the patent grants is a monopoly not just over this specific mode of implementation, but over any implementation covered by one of the claims. In contrast to the specifics of the decription, the claim is drafted to be as widely drawn as it can be without running into prior art.
So it's more than a little misleading to think of a patent as only covering one specific implementation.
In practice actual patent claims tend to be a lot broader and wider and more abstract than this -- much closer in fact to the original poster's notion of patenting/all/ implementations of the particular idea (if the claim is well drafted).
what they are saying is simply that you cannot use patents to violate the anti-trust rules.
There's also a further problem with Articles 81 and 82.
Property law is assigned to member states, and the European courts have decided that by analogy this applies to Intellectual property too.
But articles 81 and 82 are European law.
There are big restrictions on how European-level procedures can affect areas still reserved for Member States' law.
This means the EU antitrust authorities have to take a lot of care in what they do concerning intellectual property, unless they can show direct impacts in another area where they are allowed to take action.
This was the reason the EU's antitrust remedies were blocked in the IMS "bricks" case, and now why the EU's competition directorate is being *very* careful in the Microsoft case.
So it would appear that US corporations are subverting international processes for their own benefit. This is exactly the same as the Australia-US situation, where compliance with draconian US IP laws HAVE BEEN MADE A CONDITION of the US entering into a Free Trade Agreement.
Though it's good that the Australian Computer Society (ACS) seem at last to have woken up to the dangers. (Report , 4 May).
The heart of points 1,3,4 and 6 seems to me that this directive merely confirms existing practice, and so amounts to "no change".
But that's actually a very UK-centric view. The UK patent establishment has been pro swpat for a very long time (the patent office even granted some in the late 1960s, before they were apparently forbidden by the European Patent Convention in 1973); and the UK was pretty much the first state to buy into the EPO's new definition of when a computer program isn't actually a computer program.
But elsewhere in Europe the position is a lot less clear, with many courts, including the Patent Court in Germany (BPatG) very skeptical about what software should and should not be patentable; other states where the doctrine hasn't been tested at all; and even some patent offices, eg Finland, were still refusing all patents where the innovation was entirely in software until as recently as last year.
Note, too, that even many patent lawyers in favour of software patentability believe that the Commission version of this directive is the wrong way to go about things -- entrenching a doublespeak where an algorithm as such is absolutely not patentable, but the operation of an algorithm may be patentable, as may any machine readable expression of the algorithm (program claim) which would put the algorithm into effect; furthermore, through these mists of doublespeak the Counci/Commission version does absolutely nothing to clarify the very confused (and even intentionally ambiguous) line between what is and what is not "technical".
But to get to the heart of your post, if we've had software patents in the UK for so long, and the sky hasn't fallen, what's the problem ?
Well, actually I think we've been incredibly lucky so far, but I think things are changing:
1. The number of software patents being applied for is exploding.
2. They are starting to be used much more pro-actively.
It may be almost 20 years since Vicom, but it's in the last few years that software patent applications have really started rocketing. We estimate that well over 70% of all swpat applications ever have been filed since 1998, and the numbers are still going up exponentially, with a doubling time of about 4 years. In fact, most of the patent applications are so recent that probably a third are still waiting to be decided.
One of the biggest drivers for this explosion is fear. There are now so many other swpt applications being filed, that companies feel they have to file their own "defensive" applications just to establish their right to breathe. (Small companies can't do this of course). And the net effect actually seems to be a reduction in innovation -- a transfer of resources away from R&D and into legal costs.
The second issue is that the patents are starting to be used much more aggressively.
There seems to be a real change in atmosphere. Management of patents is becoming much more active to generate more income from them. Companies hit by such demands in turn try to monetise their own patent assets. And so the whole climate gets much more aggressive.
The main thing that has limited trouble so far is that most commercial software companies have able to go happily along ignoring patents, because they were very seldom prosecuted.
But this is starting to change. And as a Microsoft rep recently told me, if it does, there's going to be a bloodbath.
So that's why I think we're living on borrowed time with the software patents already issued, and I only hope we can be successful in defusing this ticking minefield.
As for Open Source, what has changed is that I think it has now become important enough that a number of commercial companies are being made genuinely uncomfortable by the competition from it. And I think the SCO case has shown that you actually can find people to sue. Now it's very easy to see what algorithms an open-source program is using, and that's why I'd say the prospects of a major patent case against one of the leading projects in the near future is probably a dead cert.
(But maybe not while the EU Directive is still not finally decided ??)
Secondly, the EU directive _actually reduces the degree of software patentability_, because currently after IBM, claims to "programs on a carrier" are allowed, but the directive removes the ability to claim this.
No, article 5.2 of the new Council draft overturns the parliament text, and explicitly permits program claims.
Fifthly, the EU always maintains more stricter examination than the US: business methods per se are _not_ patentable in the EU, and equally, flakey software patents have a harder time getting through. Stop transposing the failures of US into the EU.
Patents directed at improving methods of doing business have previously been disallowed by UK case law. This will be overturned by the directive, which will bring the UK into line with EPO practice, allowing patents for improved business methods which contain a "technical contribution".
The EPO's standards of what constitutes a "technical contribution" can be judged from the Amazon gift-ordering patent, where a patent was granted on the process of:
1. X choosing a gift from Amazon to send to Y
2. Amazon asking Y where to send the gift to
3. Amazon sending the gift
This apparently is a "technical contribution" to the state of the art, and therefore patentable.
Finally, the protests in Brussels are are a laugh: against multimillion dollar turnover businesses using patents and contributing to the wellbeing of the EU economy, you have a bunch of jokers with "terrorism is corporate suicide" and other fairly poor and non-objective slogans doing pantomines. Unless the arguments show facts and figures, and more substantive evidence, this is entirely dismissable.
Actually, as a photo-opportunity it was quite successful. And as a chance to get people concerned about software patents together in a festive environment, it was very successful.
But you may be interested to know that it was followed by a four hour conference, attended by leading MEPs and addressed by leading economists, with representatives from the Commission and the EPO also on the panels.
If you're a subscriber to LWN, there's a report about the gathering by Tom Chance on the latest weekly front page. If you're not a subscriber, the page becomes freely available next Thursday.
What if I invent a new and technically useful process for effecting interprocess memory transfer? That's not a mathematical process. It is a software process. I should be able to obtain a patent to protect it.
The question is, why is it in the public interest to grant you 20-year monopoly rights on this process ?
For a new pharmaceutical, development cost maybe $400m, development time maybe 15 years, quite probably it might be.
But will the public gain by letting you erect a tollbooth around your software process, orders of magnitude cheaper and quicker to develop ?
The observable fact is that inventions tend to be "of their time". So given the prevailing conditions, especially in software, it is quite likely that other people may quite independently also come up with the same idea, in the same time-frame.
Secondly, do you *need* a software patent to develop your invention?
Because of the opaqueness of the law around software patents (which this directive would do nothing to clarify, only enshrine), and the extreme difficulty in doing due diligence (prior art, obviousness etc), patent insurers (at least in Europe) refuse to insure patents on software, nor against patent violation -- the premiums would just be too uneconomic, I have been told. So four your software patent to have any meaningfulness, you would need credibly to be able to spend about $2m without flinching to push it in court. And because of the uncertainty factors above, if you are up against a good legal team from one of the big players, even if you can spend the money you are likely to have a very hard fight to win, no matter how good your patent.
A much better decision could be to trust in the innate difficulty in reverse-engineering software products, and to go the trade secret route instead (like eg one-time stock market darling Autonomy).
So, why should society mess up the lives of independent developers everywhere by adding another mine to their patent minefield ? What's the benefit ?
> If data structures are patentable does this make it possible to prevent interoperability?
No! The courts have continually prevented patent owners from enforcing their rights in circumstances that would restrict competition: and interoperability is one of them.
In the UK, the famous British Leyland case found this. It is codified in UK copyright and designs law.
The British Leyland case (British Leyland vs Armstrong, 1986) was where the House of Lords ruled that with the sale of a car by a manufacturer, the owner acquired an implied licence to repair it; even more strongly, that the owner *must* be allowed to repair it; and that other manufacturers must be free to sell cloned spare parts (exhaust pipes in this case) to allow this.
This decision was subsequently slammed as constituationally unacceptable lawmaking from the bench, and has been almost entirely reversed (eg Lord Hoffmann, for the Privy Council, denouncing and rejecting it for the judges having on their own account created a new head of public policy in order to modify express statutory rights of property, Canon vs Green Cartridge, 1997; see also Creative Technology vs Aztech Systems, 1997; Mars UK vs Teknowledge, 2000).
In contrast, consider the very strong support of an IP holder's ability to defend their monopoly in the IMS "brick" case recently in the European court again in Luxembourg, especially in the interim judgments, despite strong evidence (and an EU Competition ruling) that this was making competition impossible in a wider market. This Court is where the Microsoft antitrust decision is next headed, and MS seem confident that they will be able to tie the process down in legal knots for years.
There is a procedure in UK patent law for seeking a compulsory licence. But it has to be applied for separately for each single patent; has several stages which take years (and cannot even be begun until 3 years after the grant of the patent); only grants UK rights to use the patent, to serve the UK market; requires extensive lawyering; and is therefore very expensive and almost never used.
We now know the Danish delegation wanted a provision to require Member States to be able to impose RAND licenses on patents needed for the sole purpose of interoperability, and to institute fast-track procedures for making such decisions. But this was rejected.
The only provision in the new text to protect interoperability is a new recital 17 that the draft text should be "without prejudice" to the workings of the competition law Articles 81 and 82 of the EU Treaty. But any attempt to use these would require the full machinery of an EU Competition investigation, followed by the prospect of years more legal battling in the Court
in Luxembourg.
Occasionally somebody the size of Microsoft or Oracle might get taken on by the EU Commission, over mass-availability products. But nobody is going to go down this route to unlock patented data lock-in monopolies in smaller niche markets.
European Patent Application EP1338960
US Patent Application US2003156138
Alongside your existing directory tree, an automatically maintained virtual directory tree of the form [Calendar Date]/[Subject]/...
Claim 1:
1. Calendar-based interface software in a computer readable medium, comprising:
a system activity monitor that monitors user activity on a user-operated computer;
a calendar system database of metadata relating to objects stored or activities carried out on a computer;
a similarity system that determines a degree of similarity between at least a pair of computer files or objects;
a chunking system that uses the degrees of similarity determined by the similarity system for a set of objects and groups them into a number of sub-groups; and
a calendar user interface system that utilizes information stored in the calendar system database by the system activity monitor, and information determined by the similarity system and the chunking system to display in a calendar format time-based information relating to objects stored or activities carried out on the computer.
Note that MS is claiming the idea of arranging the files by date and then by automatically analysed
subject-area -- the patent is directed entirely to the user-interface concept, not any particular algorithms to deliver it.
Slashdot Mirror
-- Even more unconvincing if you look at some of the patents the UK PO has granted.
As per the other post on this thread, it would be good if you could send me the full details, to j.heald (at) ffii.org.uk -- or if you could put them up on a website somewhere. Thanks!
If you could scan or transcribe the letter, and send it to j.heald (at) ffii.org.uk, it would be really useful to identify exactly whose briefings Chris Huhne's assistant is basing her reply on.
Thank you a lot if you could do this.
A lot of people in the States seem to think that European-style opposition procedures, where private companies can challenge patent applications before the PTO grants them, are some sort of magic solution to the patent mess.
They aren't, at least not by themseves. Opposition procedures can help, but Europe still grants its share of daft patents.
More worryingly, the number of oppositions at the EPO has been steadily falling over the last ten years, though there is no evidence that EPO quality is improving. Instead, companies seem to be deciding that it's simply not cost-effective to put in the resources to do the EPO's job for it. If you're the size of Canon(Europe) for example (who I've heard this argument from), you've got a pretty good arsenal of your own patents you can hope to counter-sue or cross-licence with, and if the bad patent does come to court, you have the resources to fight it at that stage.
The people the worst patents really impact are SMEs, who have to settle, because they can't afford to fight them.
At the first reading debate, in all three of the largest parties (Conservative bloc, Socialists and Liberals) split on the directive; with the only the smaller ones (Greens, Nordic Greens/Communists, and Euroskeptics) solidly anti-swpat. In fact, the "official" line of all three biggest parties was substantially in favour of the Commission proposal (ie pro-swpat). The vote only went the anti-swpat way it did because so many individual national party groups inside the big blocs, and so many single MEPs came to understand, individual by individual, just why the Commission proposal was so deceptive, and is such a trojan horse for all-out software patenting.
This time, both the Socialists and the Conservatives have chosen MEPs as their official experts who were "swpat-skeptical" last time. But, again, every national delegation goes its own way; and then there are individual MEPs who do their own thing: the major political blocs tend not to vote solidly along party lines. So every single MEP contact matters.
And the pro-swpat people have stepped up their lobbying like you wouldn't believe. Latest story we have heard is that there are now a couple of Microsoft lobbyists working the entire parliament, trying to speak to every single MEP or assistant, a bit like Jehovah Witnesses...
The parliament is our best chance, to put over a coherent swpat-skeptical position. But it is critical for everybody to contact their local MEPs, to make sure that they understand the issues, and haven't been confused by the other side.
See also FFII's Breaking News wiki
The Council of Ministers' first reading text had been scheduled for fast-track approval before the end of the year, probably by Agriculture and Fisheries ministers.
On the contrary, it's a most valuable view, and very helpful for seeing why unitarity and/or determinism is fundamental to the Second Law, not in opposition to it.
It reminds us always to remember that the entropy is not a property of the universe itself, but rather it is a property of the description of the universe -- coarse-grained and inevitably simplified -- that we have chosen to adopt.
So, in the simplest terms, we think of the universe evolving from one of a set of initial microstates M1 through a complicated black-box operation to one of a set of subsequent microstates M2. Because of determinism, each initial state in M1 evolves to exactly one subsequent state in M2. But our description of the initial state -- in terms of macroscopic variables &c -- is not sufficient to identify the microstate. Our description is missing some of the information, and this is the entropy S1.
If we could perfectly map our whole initial distribution of possible states through the black box, microstate by microstate, then our final entropy would still be exactly S1, reflecting the deterministic evolution of that initial distribution of states. But inevitably we can't follow all of the shuffling in the black box in that detail, so some of our initial information ceases to be useful -- with the result that at the end of the process there is more information we are missing, so S2 >= S1.
So the Second Law inequality rests on two things: the total amount of information there is to know remains the same (because of the determinism); but the amount of useful information we actually have has fallen (because we couldn't follow the shuffling) -- and that is why the difference between the two, the entropy, the information we don't have, has increased (or at best remained the same). The second law does not conflict with the assumption of determinism: it depends on it.
This carries over directly to quantum mechanics, where the meaning of unitarity is essentially a guarantee that volumes in the phase space are preserved -- a grid of microstates maps forward to another grid of microstates the same size. Again, this does not conflict with the second law; it guarantees it.
In terms of the accounting, it's very important that the microstate of the Hawking radiation does represent information about the state of the universe, but information that we don't have.
Maximal entropy = maximum number of corresponding microstates. The universe is in just one of those microstates, not any of the others, so in selecting that microstate the Hawking radiation does actually represent an real flow of information.
If this is enough to guarantee that the Second Law of thermodynamics is obeyed, as the previous poster suggested, ie that
then there's no really fundamental reason why the whole thing shouldn't be compatible with a more fine-detailed, deterministic quantum description for the whole process.Can anyone here confirm that second-law inequality ?
What the Commission has offered is to suspend the sanctions only for a few months until the Court rules on whether they are acceptable as "interim measures".
The European Court of Justice is expected to take several years to decide on the Microsoft appeal as a whole.
But the ECJ ruling on appropriate "interim measures" is expected much sooner, literally within months.
The talk will then be followed by round table comments and discussion by Euro-candidates from all of the political parties. Come along, and tell them what you think.
The event is free, and all are welcome.It's a 300 seater lecture theatre, and this has had to be organised at the last minute, so help us get the word out. Let's show the candidates that swpat is something we really care about.
I've highlighted the diffs against the Presidency text proposed by Ireland.
The only diff left by the Commission was the phrase "new and"
The Commission "compromise" on Art 2(b) was to delete everything from "whereby" onwards, leaving just the addition "new and".
But apparently this was enough to win over Germany, Poland and Latvia, which was enough to prevent the whole Irish draft being sent back for the re-write it deserved.
The article mentions an interesting theory, that instead of an external meteorite triggering mass eruptions, it might be the volcanic eruptions that came first. The eruptions were powerful enough to fire a great gob of rock into space, and each big crater is where it re-impacted. On this view the eruptions would be the prime cause of the mass extinctions - at Permian, Cretaceous and Triassic - and the impact craters just a side effect.
No.
IBM Europe make it quite clear in its official response to consultations on the patentability of software in the EU, that it supports the patentability of software, and wants to see the software patents upheld, that have been controversially granted by the European Patent Office:
But IBM Europe thinks that things have gone too far in the United States, particularly as regards the patenting of business methods:
Below is Claim 1 of patent 5,206,951 filed August 31, 1992, a continuation of an application filed Aug. 21, 1987.
As you can see, it's pretty generalised stuff, not about low-level details of the implementation at all.
Did a draft of the CORBA standard, before that date, mandate the ability for object managers to be able to carry such an operation ?
If so, then it's prior art.
Well, strictly speaking the CLAIMS are what matter in a patent.
The author has to give a description of one way what is claimed could be implemented, that is good enough that somebody with normal skill in the art could create a working version of the invention ("sufficient disclosure").
But what the patent grants is a monopoly not just over this specific mode of implementation, but over any implementation covered by one of the claims. In contrast to the specifics of the decription, the claim is drafted to be as widely drawn as it can be without running into prior art.
So it's more than a little misleading to think of a patent as only covering one specific implementation.
In practice actual patent claims tend to be a lot broader and wider and more abstract than this -- much closer in fact to the original poster's notion of patenting /all/ implementations of the particular idea (if the claim is well drafted).
There's also a further problem with Articles 81 and 82.
Property law is assigned to member states, and the European courts have decided that by analogy this applies to Intellectual property too.
But articles 81 and 82 are European law.
There are big restrictions on how European-level procedures can affect areas still reserved for Member States' law.
This means the EU antitrust authorities have to take a lot of care in what they do concerning intellectual property, unless they can show direct impacts in another area where they are allowed to take action.
This was the reason the EU's antitrust remedies were blocked in the IMS "bricks" case, and now why the EU's competition directorate is being *very* careful in the Microsoft case.
Though it's good that the Australian Computer Society (ACS) seem at last to have woken up to the dangers.
(Report , 4 May).
But that's actually a very UK-centric view. The UK patent establishment has been pro swpat for a very long time (the patent office even granted some in the late 1960s, before they were apparently forbidden by the European Patent Convention in 1973); and the UK was pretty much the first state to buy into the EPO's new definition of when a computer program isn't actually a computer program.
But elsewhere in Europe the position is a lot less clear, with many courts, including the Patent Court in Germany (BPatG) very skeptical about what software should and should not be patentable; other states where the doctrine hasn't been tested at all; and even some patent offices, eg Finland, were still refusing all patents where the innovation was entirely in software until as recently as last year.
Note, too, that even many patent lawyers in favour of software patentability believe that the Commission version of this directive is the wrong way to go about things -- entrenching a doublespeak where an algorithm as such is absolutely not patentable, but the operation of an algorithm may be patentable, as may any machine readable expression of the algorithm (program claim) which would put the algorithm into effect; furthermore, through these mists of doublespeak the Counci/Commission version does absolutely nothing to clarify the very confused (and even intentionally ambiguous) line between what is and what is not "technical".
But to get to the heart of your post, if we've had software patents in the UK for so long, and the sky hasn't fallen, what's the problem ?
Well, actually I think we've been incredibly lucky so far, but I think things are changing:
1. The number of software patents being applied for is exploding.
2. They are starting to be used much more pro-actively.
It may be almost 20 years since Vicom, but it's in the last few years that software patent applications have really started rocketing. We estimate that well over 70% of all swpat applications ever have been filed since 1998, and the numbers are still going up exponentially, with a doubling time of about 4 years. In fact, most of the patent applications are so recent that probably a third are still waiting to be decided.
One of the biggest drivers for this explosion is fear. There are now so many other swpt applications being filed, that companies feel they have to file their own "defensive" applications just to establish their right to breathe. (Small companies can't do this of course). And the net effect actually seems to be a reduction in innovation -- a transfer of resources away from R&D and into legal costs.
The second issue is that the patents are starting to be used much more aggressively.
There seems to be a real change in atmosphere. Management of patents is becoming much more active to generate more income from them. Companies hit by such demands in turn try to monetise their own patent assets. And so the whole climate gets much more aggressive.
The main thing that has limited trouble so far is that most commercial software companies have able to go happily along ignoring patents, because they were very seldom prosecuted.
But this is starting to change. And as a Microsoft rep recently told me, if it does, there's going to be a bloodbath.
So that's why I think we're living on borrowed time with the software patents already issued, and I only hope we can be successful in defusing this ticking minefield.
As for Open Source, what has changed is that I think it has now become important enough that a number of commercial companies are being made genuinely uncomfortable by the competition from it. And I think the SCO case has shown that you actually can find people to sue. Now it's very easy to see what algorithms an open-source program is using, and that's why I'd say the prospects of a major patent case against one of the leading projects in the near future is probably a dead cert.
(But maybe not while the EU Directive is still not finally decided ??)
Secondly, the EU directive _actually reduces the degree of software patentability_, because currently after IBM, claims to "programs on a carrier" are allowed, but the directive removes the ability to claim this.
No, article 5.2 of the new Council draft overturns the parliament text, and explicitly permits program claims.
Fifthly, the EU always maintains more stricter examination than the US: business methods per se are _not_ patentable in the EU, and equally, flakey software patents have a harder time getting through. Stop transposing the failures of US into the EU.
Patents directed at improving methods of doing business have previously been disallowed by UK case law. This will be overturned by the directive, which will bring the UK into line with EPO practice, allowing patents for improved business methods which contain a "technical contribution".
The EPO's standards of what constitutes a "technical contribution" can be judged from the Amazon gift-ordering patent, where a patent was granted on the process of:
1. X choosing a gift from Amazon to send to Y
2. Amazon asking Y where to send the gift to
3. Amazon sending the gift
This apparently is a "technical contribution" to the state of the art, and therefore patentable.
Finally, the protests in Brussels are are a laugh: against multimillion dollar turnover businesses using patents and contributing to the wellbeing of the EU economy, you have a bunch of jokers with "terrorism is corporate suicide" and other fairly poor and non-objective slogans doing pantomines. Unless the arguments show facts and figures, and more substantive evidence, this is entirely dismissable.
Actually, as a photo-opportunity it was quite successful. And as a chance to get people concerned about software patents together in a festive environment, it was very successful.
But you may be interested to know that it was followed by a four hour conference, attended by leading MEPs and addressed by leading economists, with representatives from the Commission and the EPO also on the panels.
If you're a subscriber to LWN, there's a report about the gathering by Tom Chance on the latest weekly front page. If you're not a subscriber, the page becomes freely available next Thursday.
(more points to follow)
The question is, why is it in the public interest to grant you 20-year monopoly rights on this process ?
For a new pharmaceutical, development cost maybe $400m, development time maybe 15 years, quite probably it might be.
But will the public gain by letting you erect a tollbooth around your software process, orders of magnitude cheaper and quicker to develop ?
The observable fact is that inventions tend to be "of their time". So given the prevailing conditions, especially in software, it is quite likely that other people may quite independently also come up with the same idea, in the same time-frame.
Secondly, do you *need* a software patent to develop your invention?
Because of the opaqueness of the law around software patents (which this directive would do nothing to clarify, only enshrine), and the extreme difficulty in doing due diligence (prior art, obviousness etc), patent insurers (at least in Europe) refuse to insure patents on software, nor against patent violation -- the premiums would just be too uneconomic, I have been told. So four your software patent to have any meaningfulness, you would need credibly to be able to spend about $2m without flinching to push it in court. And because of the uncertainty factors above, if you are up against a good legal team from one of the big players, even if you can spend the money you are likely to have a very hard fight to win, no matter how good your patent.
A much better decision could be to trust in the innate difficulty in reverse-engineering software products, and to go the trade secret route instead (like eg one-time stock market darling Autonomy).
So, why should society mess up the lives of independent developers everywhere by adding another mine to their patent minefield ? What's the benefit ?
No! The courts have continually prevented patent owners from enforcing their rights in circumstances that would restrict competition: and interoperability is one of them.
In the UK, the famous British Leyland case found this. It is codified in UK copyright and designs law.
The British Leyland case (British Leyland vs Armstrong, 1986) was where the House of Lords ruled that with the sale of a car by a manufacturer, the owner acquired an implied licence to repair it; even more strongly, that the owner *must* be allowed to repair it; and that other manufacturers must be free to sell cloned spare parts (exhaust pipes in this case) to allow this.
This decision was subsequently slammed as constituationally unacceptable lawmaking from the bench, and has been almost entirely reversed (eg Lord Hoffmann, for the Privy Council, denouncing and rejecting it for the judges having on their own account created a new head of public policy in order to modify express statutory rights of property, Canon vs Green Cartridge, 1997; see also Creative Technology vs Aztech Systems, 1997; Mars UK vs Teknowledge, 2000).
In contrast, consider the very strong support of an IP holder's ability to defend their monopoly in the IMS "brick" case recently in the European court again in Luxembourg, especially in the interim judgments, despite strong evidence (and an EU Competition ruling) that this was making competition impossible in a wider market. This Court is where the Microsoft antitrust decision is next headed, and MS seem confident that they will be able to tie the process down in legal knots for years.
There is a procedure in UK patent law for seeking a compulsory licence. But it has to be applied for separately for each single patent; has several stages which take years (and cannot even be begun until 3 years after the grant of the patent); only grants UK rights to use the patent, to serve the UK market; requires extensive lawyering; and is therefore very expensive and almost never used.
We now know the Danish delegation wanted a provision to require Member States to be able to impose RAND licenses on patents needed for the sole purpose of interoperability, and to institute fast-track procedures for making such decisions. But this was rejected.
The only provision in the new text to protect interoperability is a new recital 17 that the draft text should be "without prejudice" to the workings of the competition law Articles 81 and 82 of the EU Treaty. But any attempt to use these would require the full machinery of an EU Competition investigation, followed by the prospect of years more legal battling in the Court in Luxembourg.
Occasionally somebody the size of Microsoft or Oracle might get taken on by the EU Commission, over mass-availability products. But nobody is going to go down this route to unlock patented data lock-in monopolies in smaller niche markets.
European Patent Application EP1338960
US Patent Application US2003156138
Alongside your existing directory tree, an automatically maintained virtual directory tree of the form [Calendar Date]/[Subject]/...
Claim 1:
Note that MS is claiming the idea of arranging the files by date and then by automatically analysed subject-area -- the patent is directed entirely to the user-interface concept, not any particular algorithms to deliver it.