[Patents] Letter to UK Patent Office

[Ed Avis epa98@doc.ic.ac.uk]

After lurking on this list for some time, I felt I ought to write to
an official in my country, the UK (which is not mentioned in
<http://www.freepatents.org/callme.html>) and explain why software
patents are a bad idea.  I found the address of somebody responsible
for 'modernizing the intellectual property rights system' and wrote to
him via email.  The 'letter' is below.

--------
To: xxxx
Subject: Software patents
From: Ed Avis <epa98@doc.ic.ac.uk>

Dear Sir,

I am concerned at moves to allow patents on computer programs and
algorithms in Europe.  I think that this would be misguided, hamper
innovation, and disadvantage British consumers and businesses.  The
patent system is not appropriate for software.

First, I don't think that the question of whether a computer program
is an 'invention' is relevant.  Just because patents are useful for
some kinds of invention doesn't mean that we should blindly apply the
same policy to software.  Speaking as a software developer myself, I
would say that a program is more like a literary work, but either way, 
we should consider a patent system on its merits, and not by just
carrying over a system from some other area.

We should consider whether software patents would promote innovation,
whether they would encourage disclosure of new techniques, and whether
they would benefit the writers or users of software.  In doing this we
have an excellent example to consider, the software patent system in
the US.

Software patents in America have been a disaster.  Software developers
constantly face the threat of lawsuits from companies which hold
hundreds, thousands of patents on ideas which any software engineer -
or in some cases, even any layman - would consider trivial or obvious.
The only way to defend against this is to get your own collection of
patents, preferably worded as vaguely as possible so that it will be
impossible to write a program without infringing.  Then you can
countersue if anyone claims you are infringing on their patent, and
probably reach a cross-licensing arrangement.

The people who lose out are the small to medium size software
developers, who cannot afford a large enough legal department and a
big enough patent portfolio for defence.  Some companies (such as
Oracle, the leading database company) have openly admitted that the
only reason they apply for patents is to defend themselves against
spurious lawsuits from other patent holders.  Small developers are the
ones who lose out, and it is small developers who make most of the
breakthrough innovations in software.  In any case, the hard work is
covered by copyright (see below) and patents are obtained only as a
legal weapon.  So patents do not help innovation.

Do patents encourage disclosure?  This is not true either.  Disclosing
the human-readable source code to a program opens up a developer to
patent infringement lawsuits, which are less likely if the program's
workings are kept secret.  The fact that patents once granted are made
public is not very significant here, since any important and
non-trivial algorithm would need to be made public in any case, in
order to become an accepted standard.

Do patents on software benefit the consumer?  No.  There is a grave
threat to competitiveness in many markets from patented business
models, which are quite easy to achieve if you are allowed to get
patents on particular software ideas.  For example, a patent was
recently granted in Norway which, according to its owner, covers all
e-commerce in Europe.  In America, the bookseller Barnes & Noble was
sued by rival Amazon.com over placing a link to 'buy now' on its web
pages.  If Europe starts granting patents on computer programs, it
will be possible to get a monopoly on a particular business model
simply by patenting the idea of a software program that implements
that model.  In the software market, consumers also lose out.  Patents
make it easy to stamp out competing or compatible products, by adding
some trivial 'wrinkle' to data formats used and then patenting it.
The software market is naturally very prone to creating monopolies; it
doesn't need any extra help from the patent system.

Developers are expected to check every line of their code against
thousands of existing patents.  Consider that the USPTO has allowed
the same algorithm (LZW compression) to be patented twice, by Unisys
and by IBM.  If even the patent office cannot check an application
against previous patents, what hope is there for the developer
checking a 500,000 line program?

I don't think that software patents in America have benefited anybody
except a handful of very large corporations and patent lawyers.

Some people have advocated introducing software patents in Europe,
claiming that it will help small European software firms compete
against large American ones.  I think I have explained why quite the
opposite is true; the best help we can give to software firms is an
open and competitive marketplace.  Others claim that software patents
will help in the fight against piracy, which is a complete non
sequitur.  Software piracy is a copyright violation and has nothing to
do with patents.  Neither should we harmonize our laws with America
just for the sake of it; it's unfortunate that the US system is so
harmful, but at least we are free of it here.  (In any case, patents
on software happened by accident; Congress didn't pass any law, but
rather judges ruled that software was an invention and hence should be
patentable.  They did not, however, rule that since it was an
invention it should not be copyrightable.)

On a different level, software is unlike physical inventions.
Innovation in software development is not a big event but an everyday
occurrence; every program a developer writes will contain some new
technique or a new way of combining existing techniques.  Software
development is fundamentally about combining a large number of
existing ideas in new ways, and this work is covered by copyright.
Copyright works well for software; it covers the hard part of
development, which is actually writing the code, testing, and
documentating.  Patents serve only to hamper other programs which have
been developed independently.

Don't take my word for it that patents are a danger, hear what Tim
Berners-Lee, the creator of the World Wide Web has to say:

  "I appreciate the reasons why the patent system was set up, but
   there is a really big problem here...  The bar for innovation
   seems too low. You are able to take an existing social
   practice and write software to do it and get a patent."

If you wish, I can refer you to dozens of other sources who agree with 
what I am saying.  It is difficult to find anybody in favour of
software patents, except for statements from large companies such as
IBM.

Sorry that this is a long letter, but this is a subject of great
importance, not just for the software industry, but for the whole
world in the information age.  In summary:

- Consider what will promote innovation, rather than trying to define
what is an 'invention'.

- In practice, software patents are used as a legal weapon against
competitors, rather than to protect genuine innovations.

- Patents discourage rather than help innovation and disclosure.

- Patents are a serious obstacle to competition both in the software
market and online.

and most importantly:

- Software is already covered by copyright.

I hope that the Patent Office will argue against granting patents on
software in Europe, whether by changing article 52.2 or by any other
method.
--------

In fact, I did not write to exactly the right person (that's why I
haven't given the address above).  A few days later (yesterday) I got
a reply from Jeff Watson at the Patent Office, who it turns out is a
subscriber to this list!  Here is his reply:

---------
Subject: Software patents -Reply
Date: Thu, 04 Nov 1999 14:01:21 +0000
From: Jeff Watson <jwatson@patent.gov.uk>
To: epa98@doc.ic.ac.uk

Dear Mr Avis

Thank you for your email of 1 November. I am responding as I
co-ordinate policy matters for software patents.

Reading through your comments them seem to be quite consistent with
the conclusions reached at our conference held in March 1998 (see
http://www.patent.gov.uk/softpat/index.html).

As an introduction to that event we pointed out the role of copyright
as well as other intellectual property rights in providing protection
for software.

I think you will also agree with the conclusion drawn by Paul Hartnack
(then our Comptroller-General) that any proposals for change need to
be judged on whether they provide " a system for protecting new ideas,
on software, which enables the modern dynamic and inventive people,
who make up this industry, to prosper, to the benefit of Europe in the
twenty-first century".

As I expect you are aware, we are awaiting a proposal from the
European Commission for Software Patents Direcitve. Our approach will
be consider this proposal on the basis outlined above encouraging
comments from UK interests.

I hope the above is helpful.

Jeff Watson
Intellectual Property Policy Directorate
The Patent Office
Tel- +44 1633 813650
Fax +44 1633 814922
--------

Mr Watson has said he'd prefer not to be swamped with further
messages, so it's probably polite _not_ to add his details to the
'write to me' web page.  Unless the UK Patent Office changes its
policy, of course :-)

-- 
Ed Avis
epa98@doc.ic.ac.uk