Here’s a very interesting read for those of you interested in or affected by software patents (the latter group includes anyone who uses a computer, a MMO, or just Google Earth): The Technology Liberation Front: Software Patent of the Week: Google Earth vs. Skyline (via OgleEarth).
It gets technical very quickly, so I’ll try stick to the high level issues. And since I helped write the software that became Google Earth, I’m obviously not going to comment directly on any pending legal claims that Keyhole/Google copied, intentionally or otherwise, Skyline’s secret formula, except to reiterate what I’ve said before: I think Skyline is mistaken at best.
However, I will offer an opinion that it’ll be interesting to see Skyline’s new whole-earth product, now in beta. My reading of their 1999 patent is that, for one thing, it is so vague that there is simply not enough information in it to adequately solve the problems they claim. It raises some curious speculation in my mind: Did Skyline know what they’d actually need to solve the problems efficiently, beyond the basic common wisdom of "use hierarchy, client/server, and Level-of-Detail" techniques? Could they do so for many thousands of simultaneous users and for an entire planet worth of terrain? Were they just ‘holding back’ in their patent, to keep their real technology secret?
It makes me wonder if they maybe didn’t know enough, back in 1999, and were simply planting a flag on the beach of a "new" continent ("new" as in already inhabited) that they’d take another seven years to try and conquer. It makes me wonder, in all that time, with Keyhole/Google’s application publicly available for five or six years, who exactly borrowed what from whom?
I wish I could just give you the answers. You’ll have to decide for yourselves. And for the record, I have no interest in Google other than my residual stock and, perhaps, a modicum of pride in the growth of GE, which I can’t imagine would be significantly affected by this lawsuit either way. For me, it’s more about right and wrong, technically and otherwise.
Now, I am against software patents for the most part, but not absolutely. I can think of at least a half-dozen inventions we (I and various companies I’ve worked with) considered patenting for defensive purposes. But we decided otherwise, largely for the following reasons. There are three or four basic tests any software patent should pass to be valid, IMO (and this, I believe, is reflected in the law, though perhaps not practice).
Novelty and Non-obviousness — Is the invention based on common computer science principles, "prior art" or cool ideas from the literature? Was it developed independently and solved similarly by two or more groups? Would any skilled person in the field solve it the same way? Does the technology simply apply one existing technology (say, networking) to another existing technology (say, anything we did before the Internet–substituting "Internet" for "data storage" or "browser" for "GUI" etc…)? If so, no patent for you!
Disclosure — Does the patent include the necessary information for a skilled person to recreate the invention? Remember, patents were introduced to trade limited monopoly rights for inventors in return for making their designs publicly available for all to see and even copy, with a reasonable license fee. Does the gibberish that passes for patent text even come close to giving enough information to copy a thing, let alone tell you if you may be inadvertently infringing? If you want to hide your real IP and still get protection,no patent for you!
Actuality — Does the patent describe a specific invention that was reduced to practice (i.e., built and tested)? Does it actually do what the claims imply? Or does it describe an entire theoretical class of inventions, for which the "inventors" can only guess a solution? Applying for a patent (not a provisional, but an actual patent) on something you have not actually built should be a crime, IMO, fraud and waste of taxpayer resources. Say it with me: no patent for you!
For the record, my personal opinion is that the skyline patent fails most those tests. But you can decide for yourselves. However, no matter what you think of it, it exists, as does the lawsuit. And how much did our opinion of the "crapberry" patents matter in the end?
I’m hoping that the Peer to Patent project will go some way to removing bad patents. It serves no one but the lawyers when tech companies get dragged into years of endless litigation over concepts that are already in the public domain. It’s even worse when you see a SCO-like claim by a company that has lost out in the marketplace because their technology, sales, or usability is inferior try to seek money or derail the products that kicked their collective ass. It used to be called sour grapes. Today, it’s big business. Tomorrow, it’s called ‘no more innovation, except by billion dollar companies, when they feel like it.’
If Peer to Patent doesn’t work, then the only alternative I see is to universally void all software patents and throw in business-method and natural-DNA patents too. Give inventors 12 months to re-file under a much stricter, much clearer set of rules, perhaps with the benefit of another 17-20 years protection for any that pass those tests and subsequent community review. Very few should. And if the PTO is swamped by the new work, let’s triage patent applications by first evaluating their value to society and then their validity. Patent protection is, after all, a special economic privilege we grant to those inventors who agree to provide some additional social and economic benefits for the rest of us. If it doesn’t carry those benefits, why bother with the privilege?
Previously on Brownian Emotion:
http://www.brownianemotion.org/2006/05/27/brainmachine-interfaces-for-the-masses/ (see the bit about the Sony mind-reading blue-sky patent)
[edited 6.18.06 for wording, edited 6.26 to clarify novelty and non-obviousness and make it concise]