Fair points although I didn't say Lodsys wasn't a patent holder ('generally' meant 'mostly' not 'universally') I think you overstate the density of patents in general software (at least of valid ones without prior art).
I agree expense of the US legal system that makes settling vexatious (Lodsys etc.) suits financially attractive is a real problem.
The 1.36% of revenue is for an aspect of the product that Google were not happy to remove or change sufficiently following a loss in court and includes an increase for willfulness. If I understand correctly it includes the key revenue maximisation process. It is not just for a minor UI feature and that is presumably why Google didn't just work round it. Yes the stacking of percentage royalties could be a real problem but that would mostly happen on standards if FRAND wasn't enforced and everyone stayed outside patent pools and claimed 2.4% of final retail price (Motorola/Google).
Regarding your article that does sound troubling although a large part of the problem seems to have been a dependence on winning some big deals. I do suspect that speech recognition is one of the patent heavy areas along with codecs as I mentioned, probably autonomous cars, audio and video tagging, watermarking and recognition. In these areas I would definitely want to be careful and perhaps apply for patents before proceeding too far (this would both mean patent searches were conducted and give something to defend attacks with). Alternatively I would be studying the literature and material that was 20 years old and documenting the process of obvious steps when combing approaches. Most software is not in these patent thick areas.
A proper list of dangerous areas for patents would be really quite useful...maybe an Ask HN topic?
I agree expense of the US legal system that makes settling vexatious (Lodsys etc.) suits financially attractive is a real problem.
The 1.36% of revenue is for an aspect of the product that Google were not happy to remove or change sufficiently following a loss in court and includes an increase for willfulness. If I understand correctly it includes the key revenue maximisation process. It is not just for a minor UI feature and that is presumably why Google didn't just work round it. Yes the stacking of percentage royalties could be a real problem but that would mostly happen on standards if FRAND wasn't enforced and everyone stayed outside patent pools and claimed 2.4% of final retail price (Motorola/Google).
Regarding your article that does sound troubling although a large part of the problem seems to have been a dependence on winning some big deals. I do suspect that speech recognition is one of the patent heavy areas along with codecs as I mentioned, probably autonomous cars, audio and video tagging, watermarking and recognition. In these areas I would definitely want to be careful and perhaps apply for patents before proceeding too far (this would both mean patent searches were conducted and give something to defend attacks with). Alternatively I would be studying the literature and material that was 20 years old and documenting the process of obvious steps when combing approaches. Most software is not in these patent thick areas.
A proper list of dangerous areas for patents would be really quite useful...maybe an Ask HN topic?