Not crypto, but this list has moved out onto IPR issues before,so..

[rrw]

On Friday 4 May 2001, Owen Lewis
<oml@eloka.demon.co.uk> wrote:

 [ Owen's mail program appears to have made the entertaining 
 decision not to bother quoting Julian's message, so I'm going 
 to try and do it myself with the $ character: apologies in advance
 for my inevitable mistakes ... ]

>
>-----Original Message-----
>From: ukcrypto-admin@chiark.greenend.org.uk
>[mailto:ukcrypto-admin@chiark.greenend.org.uk]On Behalf Of Julian T. J.
>Midgley
>Sent: 03 May 2001 15:19
>To: ukcrypto@chiark.greenend.org.uk
>Subject: RE: Not crypto, but this list has moved out onto IPR issues
>before,so..
>
>On Thu, 3 May 2001, Owen Lewis wrote:
>
>>> The lesson to be learnt from the USA would appear to be that
>>> software patents are prone to abuse, are rarely adequately
>>> understood/researched by those granting them, are consequently
>>> granted even when they ought to have been declared invalid and
>>> serve /only/ to redistribute wealth in the direction of the
>>> lawyers.  They stifle innovation, place small companies at the
>>> mercy of corporate behemoths, and, so far as I can tell, have no
>>> useful place in society.
>
>> Yes, patenting is an expensive process and not for the faint hearted. It
>is
>> also the surest way for an inventor reap a reward for his risk,
>> inventiveness and time. Be very sure that if one, solely or corporately,
>> spends years creating and proving a unique and valuable process most
>others
>> will implement that process for their own benefit and without thought of
>> rewarding the inventor.
>
$>This is only true if we trust the patent system (and therefore the patent
$>office) to be careful when granting patents.  Now, in the USA, it is
$>standard practice to accept practically any application for a software
$>patent unless it is obviously invalid.  It would appear that it is rare
$>for checks for prior art, etc., to be completed before the patent is
$>granted.
>
>That is not my experience of the US process. Granted that one swallow does
>not make a summer, it is also the common experience of the patent attorneys
>I deal with. With my application (a document of only some 70 pages and 36
>claims), proof of inventiveness had to be clear and not assumed. The prior
>art search was global and brought to light documents in more than one
>country. The process resulted in the loss of one claim from the application
>and the redrafting of several others. I am not sure what your information
>source may be but, with coalface experience, my view is that it is not the
>pushover you describe it to be.

 I suspect that it depends critically on who you are: small inventors
often find it quite difficult to get patents on even quite obviously
patentable inventions, because they don't know how to work the system.
Large companies find it much easier in most jurisdictions: they can
work the system, and at least in some places (mentioning no EPOs), 
the patent office is eager to please them.

 It also depends on how easy it is to search the literature: computing
is particularly bad here, because there is no literature to speak of,
and the patent office can't by and large afford good computer
scientists.

>
$>The result is that the decision about a patent's validity only
$>occurs in the court room, at the expense of the protoganists (and to the
$>satisfaction of the lawyers).  It is often cheaper (and more feasible) for
$>a small company to licence an invalid patent that it is for it challenge
$>it in the courts.  This is a long way from being ideal.

 Worse, in software, the marginal cost of licencing even one patent
from an IPR shell company may be many times the cost of production of
the product.

 This is one of the side-effects of assuming that industrial steel
production and web service development are basically identical
businesses.

>
>That's not quite the way I understand things to be. Firstly, it is for a
>patentee to apply to the courts for a judgement upholding his patent and not
>for some other party to apply to the courts to strike the patent
>down.

 .. but the assumption is that if the patent has been granted, it is 
valid. The defence must make out a claim that the patent is invalid -
it's not for the plaintiff to show that it is.

 I believe that the presumption is stronger in the US than it is in 
the EU.

>Secondly, whilst you are right that it takes a court decision to validate
>some or all of the claims in a patent, what is wrong with that? It is
>essentially similar to the manner in which all property disputes may be
>finally resolved.

 I suspect Julian was rather hoping to find a way to resolve a dispute
over a product worth $100k over three years by some means cheaper than
a court case costing $50-$75k and taking five to seven years.

 Note also that (in software especially), it won't be just one patent:
it'll usually be a family that you're infringing - somewhere between
4-12 for each company claiming. That's an awful lot of evidence you 
have to sift through. In some senses, it's really quite good that 
companies are forced out of business, or stopped from starting, by 
spurious patent claims - if they all litigated, the legal system would
grind to a complete halt.

$>Now, if careful checks are conducted on every patent prior to its being
$>issued, then this problem diminishes significantly.  The problem is that
$>the potential volume of software patents is great, so it seems likely
$>either that the patent applications will be improperly checked, or that
$>patents will take far too long to be granted.  I am not even certain that
$>the European Patent Offices will be able to employ a sufficient number of
$>adequately knowledgeable people to accurately determine the validity of a
$>patent prior to its being granted.

 I'm not sure that would help: the EPO has a nasty habit of granting
patents that break its own guidelines (e.g. on biotech), and then
hoping that state patent offices will go along with them.

[snip]
>d. If relevant prior art is discovered by the applicant or any patent
>office, then the applicant must revise and/or diminish his claims (or
>withdraw hurt). Moreover he is under an obligation to inform each and every
>office examining his application of any and all discovery. Failure to do so
>would be fatal.

 .. which is one reason why, having paid your fee, it's a good idea
not to look very hard.

[snip]
>> In regard to your last sentence, it is an inversion of both the purpose of
>> patent law and, in my experience as a patentee, of the truth. Patent law
>> gives the small man/business a protection that enables him compete with
>> giants. I was about to write here that it is as was David's sling. More
>> closely. it's akin to Sam'l Colt's Equaliser.
>
$>The problem the small company has is that the large company can afford to
$>file more patents, and more quickly.
>
>True. But without the patent process, the small business would have no
>effective protection at all.

 Agreed. I think that an effective patent system is, on balance, a
 good thing. The trouble with the one we've got is that, although in
 1886, pretty much all manufacturing business was the same, in a
 modern context it just isn't sensible to treat minor variations on
 simple algorithms in exactly the same way as a new industrial process
 for producing ammonia or the invention of a gene therapy for
 Huntingdon's chorea.

 Worse, various lobbies seem to believe that the only thing wrong with
 an overextended patent system is that it's not overextended enough,
 and keep pushing to extend it into places it was never meant to go -
 for example in protecting naturally occurring gene sequences, despite
 the fact that they're not novel, not invented, and that no-one can
 think of any industrial application for them. One day soon, someone
 will suggest patenting trade secrets ...

[snip]
 >Essentially, business is about risk management. One makes decisions and then
 >must live with their consequences.
 >Yes, the patenting system is not perfect and it is expensive. It is for each
 >inventor to assess whether or not iit is in his overall interest to use it.

 Speaking as a random programmer, I'm less bothered about the
 obligations placed on inventors because interaction between inventors
 and the patent office, than the obligations placed on me by that
 interaction.
 
[snip]
$> It is well known that large
$>companies in the states patent everything they possibly can, not so that
$>they can extract royalties for their use, but so that they can threaten
$>other companies with their patent portfolio if the other companies are so
$>foolish as to demand royalties from the big guys.  "Ah, so you would like
$>to claim royalties on your procedure foo?  Well, it seems that you are
$>infringing our patents on methods bar, widget, bar2, foobar7 and fnord.
$>Perhaps it would be better if we all just forgot about this little matter
$>of patents?"

 Well, yes.

>
>You take the popular swipe at 'big business'. May I say that my experience
>of trading with major corporates is that they are - or try to be -
>punctilious in their dealings with small business. The same cannot be said
>for trading with small businesses, where standards are not so uniformly
>high.

 In general, I agree, but if big business starts feeling threatened, 
they can turn awfully nasty...

[snip]
>> Ob Crypto:
>>       Has patent law inhibited either invention or exploitation of
>cryptographic
>> processes? I don't think so.
>
>What about the RSA patent - practically all open source software that uses
>the patented algorithm was written in Europe, since the US Patent Laws
>don't apply over here.  Now if we enforced Software Patents as well...

 It's not a software patent thing: RSA is perfectly patentable under
EU software patent law, at least as applied to means of communication.

 The problem with RSA was that the authors unwisely published about a
year before applying for the patent - although this is allowed under
US law, it counts as prior publication under EU (and UK) law, and
therefore a patent couldn't have been issued.

[snip]
>You do not address the question nor make a clear point. RSA has been
>exploited commercially on a global scale. It is also in substantial
>circulation in freeware applications.  The patent did not prevent its
>widespread adoption and use.

 It (and the related IDEA patent) significantly retarded the use of
ssh, SSL, and PGP -  which has contributed to a significant slowdown
in the speed with which the internet is becoming practicably secure.

 [ but then, that's what patents do: they retard innovation in order
 to reward innovators. ]

> The patentees (also the DH patentees) were
>largely crippled in establishing an effective monopoly on their inventions
>by the US export controls - but that's another story and nowt to do with
>value of patenting.
 
 Agreed.


Richard.