Invalidating harmful patents
By invalidating the most harmful software patents, we can reduce some harm being done by software patents.
Limits to effectiveness
There are too many
We can only invalidate a small number of patents. Example: MPEG-LA represents dozens of organisations which hold 900 patents which they say are required to implement the h.264 video format. Freeing this format by invalidating every one of them is completely impractical.
In the words of Richard Stallman:
[...] eliminating software patents one by one is as useless as trying to eliminate malaria by swatting mosquitos. Every software patent is a threat to the field of software, and we need to get rid of them all at once.
Some can't be invalidated
Patents can only be invalidated if they are invalid to begin with. Take the RSA patent for example.
We don't know when we've got them all
Some software patent pests have a large collection of annoying patents they could use instead if their favourite patents were invalidated. For example, Microsoft accused TomTom of violating eight patents. The patents look invalid, but if they are invalidated there's still the worry that Microsoft will choose some others from its portfolio of 10,000 patents and continue the menace.
Another example is the MPEG-LA situation. If, by a miracle, all of those 900 patents were invalidated, there is still the possibility that another patent holder (who did not submit his patents to MPEG-LA) could announce that he has patents required for MPEG video formats.
- See also: Incompatible delays and durations
This can be difficult, expensive, and can take years. The Amazon's one-click shopping patent is an example where the review in the USA took 5 years and the patent wasn't even invalidated, only narrowed. In Canada, Amazon applied for an equivalent patent in 1998, and in 2010 the court cases are still ongoing.
An example is Germany is this case of a patent on spam whitelists, the review of which took five years: German original: T-Mobile SPAM Patent ruled invalid (English translations[?]: Google, bing translator)
Sometimes the good guys give up too soon
In the case Trend Micro v. Barracuda (2008, USA), the Groklaw community helped gather prior art to help Barracuda fight Trend Micro's patent, but when it seemed that Barracuda would win, Trend Micro offered an out of court settlement and the invalidation never happened.
Prior art or subject matter
This is almost always done by presenting prior art, thus claiming that the idea is not new. Sometimes it can be done by claiming the patented idea was too obvious. These invalidations don't create any useful precedent for reducing the scope for the patent office to grant software patents in future. To do that, we would need to invalidate patents based on subject matter. If it was argued that the patent was granted for something that's not patentable (software, business methods, math), then the precedent would be very significant because it will narrow the scope for the granting and enforcement of software patents.
Related pages on ESP Wiki
- Duds and non-solutions
- Prior art database - this is one method that can be targeted at invalidating specific patents
- Amazon's one-click shopping patent - the review took five years and only narrowed the patent
- Buying harmful patents
- Invalidate patents based on subject matter
- How to read patents and gather prior art
- State of the art - what is "prior art"?
- EFF's project: "patent busting"
- If We Can't End Software Patents Tomorrow, What Should We Do In the Meantime?, by Bradley M. Kuhn
- Post-Issue Peer To Patent
- USPTO's 1-Click Indecisiveness Enters 5th Year
- Transcript: Andrew Tridgell on Patent Defence, 2010
- Stallman Richard, End Software Patents[archived], stallman.org, 2012-03-19.