Fraid I just couldn't get my old subscription to this list to work,
so I've tried resubscribing. I hope this works, but bear in mind I
may not receive any replies unless directed straight at my address.
Note also that I have missed all posts to this list since January
though I've flicked through the archives.
Below is the text of a letter I have sent every Irish and English
MEP. I would have also sent one to every MEP from a country likely to
speak English (eg; Germany) but since I've been unemployed for
eighteen months now and had to move back to live with my parents, the
sixty odd euro I've already spent is as much as I can spare.
I hope the ideas contained within the letter will be useful to
others. What is important to remember is that MEP's don't understand
technological or ideological argument - they want to know (a) what's
wrong with a proposed directive and (b) how to fix it. With these two
bits they can make arguments without having to research the matter
for themselves.
Hence in the letter I have (a) explained in simple terms what's wrong
with US-style software patents and (b) an alternative form of patent.
You must remember that TRIPS mandates software patents, by treaty the
EU must adopt some form so just saying "no to patents" is no good. I
have spent much of the last few months honing my arguments with my
local MEP's so I am confident the letter below will resonate with all
MEP's.
Lastly when sending your correspondance, remember that most MEP's are
on holiday till late August. The plenary session is in early
september, therefore you should probably post to the European
Parliament directly and mark the top left corner of your envelope
with "ATTN: For September plenary session" or something similar. This
will ensure it is read before the plenary session and not lost in the
backlog that builds up during the summer.
Best of luck!
Cheers,
Niall
--- cut ---
Dear MEP,
I am writing to you again concerning the directive on the
patentability of computer-implemented inventions 2002/0047 COD which
is due for plenary debate in early September. Thanks to
correspondence with MEP Brian Crowley I now understand my original
letter to you was unclear and overly-technical and so here I shall
break down the issue:
1. Why patenting software in the same way as traditional patents has
the opposite effect to patenting manufactured goods:
Since we have had patents for so long, we automatically associate
them with generic protection of invention & fostering a good
environment for innovation. Very rarely do people actually think how
patents protect inventions and while I will now give a quick summary,
you can consult Thomas Jefferson's work (yes, the same founding
father of the US) on the 1793 Patent Act.
Patents traditionally protect the inventor, not the invention. They
are intended to provide a temporary monopoly allowing the small
inventor to bypass conventional economic rules so that he/she may
become big enough to properly compete when the temporary monopoly
expires. Conventional economic rules means that an established
manufacturer has scale of economy for raw materials and established
manufacturing capacity such that they could bring any new invention
to market much quicker and cheaper than any smaller operation - this
is what is being suspended. Because of this patent law historically
permitted a competitor to substantially enhance a patented item and
patent that without having to pay royalties for the parent work.
What's most important to understand is that patents protect the
inventor, not the invention.
Now we as a society have forgotten this especially in the last twenty
years - I should also add that patents were controversial in their
own day with some saying that gradual slippage into a tool of
economic warfare would happen in the way many would argue they have.
However back to the point - software differs from manufactured goods
in two major ways: cost of raw materials is nil and cost of
manufacturing is almost nil. Because of this I with a CD burner can
manufacture Microsoft's software nearly as cheaply as they can and if
I distribute it over the internet, then even cheaper again. We're
talking fractions of a centimo here.
This means that computer software's cost is exclusively research and
development costs. The very fact that patents were invented in the
first place was because small inventors could outdo an R&D department
of a large company spending millions i.e. R&D costs are
extraordinarily variable - or even you could say that genius is
incredibly expensive if you don't already have it.
Having established that software is very unlike a manufactured good,
what are the consequences? Well effectively you are solely patenting
the research and development process i.e. preventing other people
performing similar research and development to the one you patented
for fear of infringing that patent (especially as the proposed patent
directive is not clear about how much more innovative you must be for
a new patent i.e. it's shelved onto the judiciary, and thus the
innovators in legal costs) plus there is a much diminished return if
you must pay insurers more if you do innovative work (higher risk).
Can you see clearly now why there would be a chilling effect on
innovation? Large companies can afford to purchase many patents and
if you look at IBM's example, it patents hundreds of promising areas
of software development every year as a "patent land-grab" yet does
further research into only a few of them. Nothing in Arlene
McCarthy's recent revisions will stop that curse hitting us here.
2. Arlene McCarthy's report (C5-0082/02) is fundamentally incorrect &
misrepresentative on many levels:
The report (a) shows a complete misunderstanding of how patents work
in general (b) misrepresents that software patents would be good for
SME's (Small & Medium Enterprises) and (c) incorrectly states that
the EU needs to have US-style software patents because of TRIPS
mandate.
On point (a), I merely need to refer to how many times the report
says things like "Patents play an important role in ensuring the
protection of technical inventions in general". Patents are not for
protecting inventions, they are for protecting inventors!
On point (b), I refer you to page 30 of the report: "While opposition
to patents relating to software was expressed by a large majority of
the individual responses to the consultation, the collective
responses on behalf of the regional and sectoral bodies, representing
companies of all sizes across the whole of European industry, were
unanimous in expressing support for rapid action by the Commission
more or less along the lines suggested in the discussion paper".
Firstly, that misrepresents the grave misgivings expressed by
European companies about the idea (yes it urgently needs to be
tackled, but we don't like US-style software patents). Secondly the
phrase "opposition to patents relating to software was expressed by a
large majority of the individual responses to the consultation" is
extremely important because it shows that software engineers
themselves are overwhelmingly opposed to US-style patents. Where
SME's and large companies were consulted, invariably support for US-
style patents was expressed by management without consulting their
engineers - and quite frankly, the management of most software
companies just don't understand enough to be able to say (and any
software engineer will back me up on this).
On point (c), just because the United States of America has shoddy
software patents doesn't mean we have to repeat their mistake.
Instead we should look at this as an opportunity to improve the
regulation of a sector long insufficiently supported by law for the
benefit of the industry, consumers & society. The key to the new &
much improved form of software patent which remains compatible with
TRIPS mandate is to make the computer software itself the patentable
item - not its concept. Here is how it should look:
1. All computer software must come with sources (i.e. the files which
create the computer program). The reason for this is fourfold:
a. To permit study by those learning computer programming. In the
1980's, most computer software for home computers came with source
and it led to a generation of talented programmers who taught
themselves in their own home (such as myself). The fact this has been
removed has inexorably led to declines in competency and thus overall
quality of the profession - markedly obvious in the large &
increasing number of failed or over-budget IT projects.
b. To permit inter-operability with other systems (especially those
the vendor has a vested interest in not being compatible with). While
legal rights already exist in Europe for reverse engineering in this
regard, the process is made much more difficult and less successful
than it would be with source.
c. To permit users to fix bugs in third-party software. I cannot tell
you how frustrating it is to report a bug to a vendor, to even know
where it is and how to fix it but the vendor refuses to fix it
(because doing so would cost profit). The result is lowered quality
of engineering and much waste of time for hundreds of millions of
people globally (think of every time your software has gone wrong and
the work lost).
d. After seven years from the date of release, all source must become
available for unrestricted public use for any reason at all. This is
probably the most contentious part of my idea - computer software
operates on a three to four year cycle so seven years is
approximately two evolutionary cycles of software - thus only a
minute amount computer software is still being sold in its original
form after seven years. Making all code freely available after seven
years means that the very best code gets amalgamated into a common
pool available for the benefit of all software programming everywhere
- thus preventing to some extent the massive industry-wide
duplication of identical functionality which currently happens - but
still allowing for plenty of potential profit. Might I add that
already some industry leaders release the sources for products after
seven years (e.g. id Software).
2. An implementation of an algorithm (hereon in called a "library")
should have strong Intellectual Property protection. The owner of
such a library can licence its use to others for a royalty.
Competitors can also develop competing implementations of the same
algorithm(s) and thus a state of competition is created, where
implementers compete with each other for customers - thus driving up
quality, lowering price and preventing the current massive waste of
production. The existing system (which would get worse with software
patents) has so many companies duplicating the same functionality -
under this new system, they'd be strongly encouraged to sell off
reusable functionality as libraries to supplant development costs.
The industry would benefit as a whole.
3. Computer software should be sold, not licensed. The purchaser
should get the same rights as when buying anything else, including
the right to a free replacement of a defective product or their money
back and the right to modify their purchase as they see fit. This
includes customisation, fixing bugs and indeed applying other
people's bug fixes. Naturally though, they must not redistribute or
sell a modified product without previously agreeing a royalty payment
- after all, they are reusing the work of someone else - however,
they can freely distribute the information to others on how to
similarly modify their purchase.
4. Computer software should not be sold as separate versions for each
computer e.g. one for Apple's, one for Windows etc. nor as one per
computer. If you buy a program, you buy the right to use it on all
computers you personally use (within reasonable limits). Companies
will complain that each "port" to a new platform is expensive -
actually, technologically this has become much cheaper in recent
years and will soon sink to zero overhead cost so in fact vendors are
making unfair profits from users using (a) more than one machine and
(b) using more than one type of machine (however I would permit a one-
off separate charge for a brand new port). If you think that this
might reduce the incentive for vendors to produce versions for
alternative systems, point 3 above will ensure that if enough users
want it, they'll make it happen on their own - and thus the vendor
loses out on profits. Plus, most computer software license breaches
are currently made by private individuals installing multiple copies
of the software - and I can see no point in criminalising vast
segments of the computer using population when nothing can be done to
prevent it anyway.
5. Following on from the last point, not-for-profit distribution of
computer software among private individuals who personally know each
other should no longer be criminally sanctioned. The near-zero
duplication cost of computer software makes any less tolerant law
unenforceable and again, one is criminalising large proportions of
users for no gain and much loss. Furthermore, distribution among
friends is invariably done because of
recommendation so in fact this process helps vendors sell more
software. Of
course, support and upgrades and other legal rights listed above can
not be
granted on anyone who does not own a copy - this provides an
incentive in
addition to the moral one to pay those who work to make your life
better.
Hence I urge you next September to reject any and all forms of US-
style
computer software patents as currently proposed by the directive on
the
patentability of computer-implemented inventions 2002/0047 COD. It
would
substantially chill technology venture start-ups across Europe (as it
has done
in the US) and leave them open to technology "bear raids" by US
multinationals - to the detriment of software quality, innovation, IT
employment & Europe's ability to set the technology pace in today's
global environment.
Yours sincerely,
Niall Douglas