Richard Stallman on Software Idea Patents

Espoo/Otaniemi, 23.07.2004

Confusing term: "Intellectual Property"
Intellectual Property is an issue that doesn't exist
Copyright Patent
specific work, detail of expression exclusive monopoly
automatic issued by patent office, expensive.
New, useful, non obvious (IQ 80)
lasts forever 20 years
copy only absolute monopoly, no possibilty of double invention

Patents are a lottery: the focus is always put on the unlikely possibility that you win.
The other point of view is shown by the danger of being pointed to by someone who says: "give me your money".

Secret developments: compress, using the LZW algorithm, published in a paper. gzip: happy end?

Problems with patents not issued yet, but also pb with patents published.
Huge stack. Search for relevance: unreliable
Example: natural order recalculation in spreadsheets, using topological sort - known as that since the 1960s.
Tortuous legal language
Australia: no use for patents for Australia, but pressure
Failing purpose to get inventions published
Read something you are not allowed to do?
Secret settlement Heckel - Apple

"I didn't know the scope of my protection"

Chance in lawsuits

Options: 1 avoid, 2 license, 3 challenge

1. other algorithm? example fft.
Efficiency, but also standards! LZW: mandated by Postscript (only uncompress in fact).
gif - png (png's not gif): hard to get people to switch.
jpg currently on court.
Actually 2 patents on LZW, although not intended.
Patent offices try to please their customers, fast.
mpeg2 video: 42 standards (?)
Feature patented - downgrade.
Motto: Make programming safe.
Some patents so broad as unavoidable - e.g. public key encryption.

2. Pay a lot for it. N.O.R: 5% price of every spreadsheet (what if 20 such patents?)
Cartels of companies sharing in fact their portfolios.
Avoiding problems with others = cross patenting.
IBM estimate: 10 times as much as the gain by licensing.
This is a measurement of the harm per use ratio!
Defensive patents.

Patents do not protect the isolated developer. "Co-protection" with IBM: IBM may compete with him.
"Parasites" who don't want to coprotect (to develop).
Free Software: possible to produce software without money, but not to pay licenses without money.

3. the dice were rolled years ago.
Qualcom defended and lost 13 million dollars.
Risk to lose before to get to the point where evaluation occurs.

Practically speaking, people don't try to know (extra penalty for knowing).
Walking in mined lands (mine fields)
Other fields but SW:

"Everybody got cancer, why not you?"

Fallacy: one patent for every idea.
Pharmaceuticals vs software.
SW is easier.
"Perversity of matter".
Push to the limit: software crisis.
One program combines more ideas than items in other fields.
All are points of vulnerability.
SIPs intolerable.

Stupid arguments:

Marco Schulze NightLabs GmbH

Software is not protected - investments are (or are not).
24.09.2003 clear EU Parliament decision against software patents.
30000 SW patents already issued in Europe (illegal, non enforceable), out of which 70% owned by US ad Japanese companies.

Kristian Engström

Important time frame. New vote in Parliament on revised council proposal during the autumn 2004. Draft during the Irish presidency.
Whom to talk to?
Politicians accessible
Nokia small with respect to Microsoft or IBM
Amount of lawsuits in the US (Slashdot)
200000 patents
Acacia case: EPO566662 - won in court against some porno sites (send movies over the web).

Electronic Frontier in Finland (EFF?)

Transcript, audio, (also in mpeg-ps format: 1, and 2),
Lectures ToC
Marc Girod
Last modified: Sun Jan 1 14:27:48 EET 2006