We are developing the social individualist meta-context for the future. From the very serious to the extremely frivolous... lets see what is on the mind of the Samizdata people.

Samizdata, derived from Samizdat /n. - a system of clandestine publication of banned literature in the USSR [Russ.,= self-publishing house]

Government Lemmings Department

Ever since the US implemented software patents, the EU has been determined not to be left behind. According to an item in Debian News:

“More on European Software Patents. An article at ZDNet UK says that the EU bureaucrats aren’t even considering the numerous anti-software patenting opinions out there. According to a well-connected lobbyist group, they have determined there will be patents, and the only question is what kind.”

Software patents are a very bad idea as many have discovered in the US. They put a fence around parts of mathematics. Even worse, the government guardians of the fencemaking are not familiar with the field and its’ literature and are vastly understaffed to boot. This has led to some gawdawfully silly (and economically destructive) patents being granted.

Within the last couple years BT tried to claim a patent in the US on hyperlinks1. This was shot down after a massive search by the open source community. It ended with the web publication of a late 1960’s lecture discussing the idea.

Virtually everything in well done programming is “obvious” when you get “there”, down to the core of the problem at hand. Two top programmers given a reasonable time to attack a complex problem will very likely find large sections of their work very similar if not almost identical.

A program in a well formed language is equivalent to a mathematical expression. Such expressions are in most cases transformable into each other.

The Debian News cited article can be found here

1 = 01/07/2000 NewScientist p017 “The Net Strikes Back: BT tries To Patent links”

14 comments to Government Lemmings Department

  • Russ Goble

    Not to be hyperbolic or anything, but I’m pretty sure that U.S. enhancement and interpretation of copywrite & patent laws is the worst government “innovation” of the last 30 years.

  • And perhaps related to the US being ahead of many countries in computing?

    It wouldn’t be the first time a government did something mercantilist in motivation to help a domestic industry, but ended up hurting its own pet industry as well as others’ and trade in general.

    But we should notice that the US lead in security encryption is related to this too, surely? Lots of once ‘pure’ number theory (which 1930s English mathematician Hardy mistakenly predicted would never have military uses, would never “hurt anybody” in his words) now gets classified as military information. This must affect decisions to grant patent coverage.

    If it can be a weapon, it can be an invention? That kind of inference?

  • Dale Amon

    The US software patents are quite a recent “innovation” and if they have caused any economic advantage it has been to lawyers rather than the general public.

  • David Mercer

    On a personal note, I have had at least one serious software venture shot down in the crib by a software patent (they wouldn’t even license for 10% of gross!!), and have had numerous other ideas stall out when I discovered that portions were already patented.

    The damage from them is real. I wouldn’t be so opposed if they had a statutory compulsory licensing rate, like some types of copyright in the US, but I think I’ll see the devil snowboarding first.

  • Della

    Software patents are not a bad idea in themselves, the problem with the American system is that some people have patented stuff that is pretty obvious, and it is very difficult to overturn the patents. This is a problem with the patent office and not patents themselves. I have read quite a lot of US software patents, and quite a lot of them are very high quality research, but others are very slight variations on patents that have gone before. The language in those repeat patents seems to have been made deliberately unclear so the fact that it is the same thing as an earlier patent is not obvious to the layman, although to be fair when one invents something new one also has to invent terminology to describe it, and different people may just have chosen different terminology. One also has to be consider that the inventors may be under some pressure to produce patents, and this might lead them to taking an easy way out, patenting something already existing. In my own case I must produce patents or I get fired, I also know that if I do produce a few patents then I will then I will probably be able to triple my income.

    To say that virtually everything done in programming is obvious is quite simply nonsense. I will admit that most programmers are pretty much useless, people who can hardly program much less think of an original idea, but that is not a problem with software per-se. You can create genuinely inventive stuff in software, stuff that is not obvious, is not simple, and has commercial value. Picking out a few bad apples in the field of software patents is not going to convince me that this sort of property should not exist, any more than a drive by shooting is going to convince me that cars should be banned.

  • S. Weasel

    I suspect it’s down to an unfortunate confluence of judicial ignorance and American greedy litigiousness. Computer-illiterate judges and legislators have tried to give us all sorts of unworkable nonsense, like the Child Online Protection Act. Unfortunately, patent is something they understand, if algorithms are not.

    Errrr…sorry ’bout that.

  • Patents have traditionally been required to cover things that are non-obvious, and that haven’t been invented before. In recent years lots of patents have been applied to things that are obvious and/or where there is prior art, and then the holders of such patents (if they are large companies) can blackmail smaller companies and individuals simply on the basis that the legal fight will be too expensive to fight. We are so overwhelmed with frivolousness that the question of whether some software patents are non-frivolous is barely the point.

    I have no problem at all with patents on things that genuinely are non-obvious and where there is no prior art (a situation that still applies quite well in the case or pharaceuticals, for instance) but the situation in practice with software patents (and with business method patents and the like) appears to be a complete perversion of what patents were designed for.

  • Phil Bradley

    I have submitted a US software patent and read quite a lot of them. They vary from the trivial, through generic (very wide applicability) to pretty much incomprehensible, but I yet to see one, where I said ‘Wow! What a cool idea’. My reaction to suprising number of them is that, in part or in whole, I have seen what they are trying to patent before (I’ve been in the SW biz a long time).

    I diagree with Dale on one point – Two top programmers given a reasonable time to attack a complex problem will very likely find large sections of their work very similar if not almost identical. There are large numbers of ways to solve the same problem, and the fact that any solution is transformable into any other solution, of the same problem, is irrelevant to SW patents. If it wasn’t almost all patents would be invalidated.

    Software patents are a very bad idea, because they do not protect the whole of a work or the code itself. Actual code is protected by copyright. They in fact protect ideas, and as should be, and until recently was, the case ideas are not patentable.

    For most ideas, expressing the idea through technology is a considerable hurdle, for software its trivial, which leads to another problem, which is that a lot of ideas find there first (or nth) expression by a hacker in their bedroom.

    Yet another problem is that even a moderately complex piece of software incorporates many hundreds of ideas. I maintain that it is impossible to write a piece of software of even moderate complexity without infringing on the protection claimed by an existing, and probably many, patents. Of course it doesn’t matter, unless and until you are making money out of and then you get sued.

    US software patents are a major barrier to innovation and most considered the EU’s much more restrictive approach to software patents a better way. The USA is pushing for US type patents in the EU supported of course by lawyers. The open source crowd is pushing for no patents. And of course the legislators don’t know what to do.

    If the EU wants to steer a different path to the USA then I can think of a single better issue to do it on than software patents. severely restrict or eliminate software patents altogether, and if I am right and I am sure that I am, then a vigorous European SW industry will result with considerably more inovation and less legal overheads than their American competitors.

  • Doug Collins

    From my not very authoritative position, it seems over the last decade – or more accurately during the Clinton administration – US patent procedures changed to make obtaining patents very much easier than it had been. This was a boon to lawyers and, as Micheal Jennings has already pointed out, to large companies with large legal staffs. Not too surprising, considering where a lot of his (Clinton’s, not Mr Jenning’s) political support came from.

    Perhaps someone can correct me if I am wrong, but my understanding is that traditionally the US philosophy has been to make patents difficult to obtain, with strict conditions of non obviousness, new art and so on. A US patent was, among other things, a government certificate saying that in all probability you would prevail in a legal patent fight. I understand that German patents, and I assume most other european patents are less expensive ad time consuming to obtain, are granted with less detailed inspection and essentially certify a date of priority, but make little claim of legal rights. That must be settled in litigation if the question arises.

    Under the old US philosophy, software patents would probably be an impossibility. I wonder what the situation was with hardware? In many technical problems, either software or hardware approaches will work. Software is preferred often because of price. When someone came up with a new arrangement of logic chips – say the first flip-flop gate (which is a clever arrangement of OR and AND gates) – was that patentable? I’d guess it was not.

  • Phil Bradley

    Here is a good site on US software patents – http://www.bitlaw.com/software-patent/index.html

    The US patent office originally refused to allow any patents on software, then the lawyers got in the act and forced them to (via the Supreme Court), such that effectively anything is patentable (the no prior art restriction, is useless at the patent granting stage).

    Yet another problem with patents is that they are written in a mixture of legalese and computer terminology. Most of the computer terminology you would find it hard to find any significant number of knowledgeable people who agree on what the terms mean.

    Resulting in recent high profile cases using terms that are invented and have no relation to any actual usage, e.g. ‘middleware’ for a browser.

    I’ll avoid a major rant on the legalisation of problems, and suffice it to say technologists are in the business of solving problems, whereas lawyers are not. They do not make a good mixture.

  • Alfred E. Neuman

    Two top programmers given a reasonable time to attack a complex problem will very likely find large sections of their work very similar if not almost identical.

    Dale, this is not true. It is sort of true as you go more lower-level, but as you get into the higher languages there are always multiple ways to solve problems, sometimes quite a number of them. Part of this is because of the flexibility and scope of the higher languages, and partly from OOP (object-oriented programming) providing other people’s solutions to you.

    For instance if I am dealing with XML, do I write my own XML parser as an object model? Do I use use Microsoft’s MSXML 4.0 object model? Do I do everything relating to XML in string form, and eschew all objects? Do I use Microsoft’s parser to load the XML but then work on it as a string? And there are different reasons why you might want to do any one (or more) of these things depending on your project.

    There are many, many ways to solve problems in programming, with more options the higher up you go. Programmers will choose their method based on what will work best for them combined with personal pereference and personal experience with components and programming styles.

  • Alfred E. Neuman

    Whoops, forgot to add my conclusion.

    The point is that programming is a very murky area when it comes to patenting. Two programmers can achieve the same results but use two totally different methods to get there. And one of those methods could be significantly superior to the other (as in IE vs. Netscape).

    As a programmer, I strive for elegance, simplicity, and robustness in my code. I have personally designed systems that I regard, internally (code-wise), not so much as code but as a work of art. Should I be able to patent that? That’s the $60,000 question.

  • Russ Goble

    I think Phil Bradley has the best summation of the problems in U.S. patent law. I also agree that software solutions are not always obvious and that 2 people won’t necessarily come up with the same solution. When you are dealing with software that in many cases has a million lines of code, you are going to have an awful lot of variation.

    Also, while lawyers are easy to blame, don’t discount the software companies themselves. The larger they became the better lawyers they could hire and they could hire lobbyists to produce pieces of crap like the Digital Millineum Copywrite Act. And because of the current percentage of our economy software accounts for, these companies will yield greater influence.

    Mark’s comment that patent law is why so many U.S. software makers are successful doesn’t really work for me. Possibly in some of the encryption stuff, but I’m not sure patents and issues of national security are synonymous.

    Finally, Virginia Postrel wrote a pretty good article for Reason Magazine a couple of years back regarding the constitutionality of copywrites & patents. The U.S. Constitution actually does have a lot to say regarding these things and it’s pretty clear that consumer benefit and perpetuating innovation is of the utmost importance. My beef is that the modern interpretations of patents and copywrites has taken consumer benefit and innovation almost entirely out of the picture. Patents are now used to prevent innovation and copywrites are often used to punish customers. And apparently with the Supreme Court’s blessing.

    As Phil Bradley pointed out, this will hurt the U.S. software industry over the long haul if other governments are smart enough not to copy us. There’s got to be a middle ground somewhere where non-obvious solutions are patentable but obvious solutions are not (which, granted is a subjective thing). And business method patents deserve the very strictest of scrutiny.

  • Phil Bradley

    Russ

    The Digital Millineum Copywrite Act is a product of the entertainment industry, although Microsoft and Intel are getting in on the act. Doubtless as a way of extending their de facto monopoly.