It’s not often that I get up on my soapbox about ‘IP’ issues, though I did write about a copyright infringement problem I had a few years ago — that one turned out well, maybe this one will too.
It’s also not often that I will write a negative review — not because I like everything that I read, just that I'd rather follow "Thumper's Rule" and just not write anything. Every once in a while though, something comes up that makes me go against these personal policies. Yesterday it happened again.
Lately, I’ve been getting interested in extracting information from unstructured data. (I blame some of my recent reading like: Visualizing Data and Programming Collective Intelligence.) So, I was really excited when IBM Press sent me a review copy of “Mining the Talk”.
I sat down and started reading, I’d nearly finished the introduction when I hit a problem. Right there on page xix, the last paragraph reads:
On the other hand, if you want to take our methods and create your own software solution to sell as a product, while we applaud your initiative and enthusiasm, you really should first discuss this with suitable representatives from IBM business development. IBM has sole ownership of all the intellectual property described in this book, all of which is protected by U.S. patents, both granted and pending. All rights reserved, etc., etc.
Wow! How broken is this? IBM has put two senior staff members through the effort of writing a good looking book; put untold others to work editing and marketing said book (at $45 a pop); then put the whole thing under lock and key by claiming that they hold patents over the whole of it. They might as well have written the book in Klingon -- ehh, no I guess not, too many geeks can read Klingon.
I don’t know about you, but I read books so that I can learn how to do something. If a book carries a warning label which states that I’m putting myself at risk of patent litigation by implementing the ideas in the book, I’ve got a problem with that. A big enough problem that I'm not going to reading it.
I brought this to the attention of the authors, and I got the following response from Scott Spangler:
You raise a very important point, and it is one that we wrestled with in writing the book and in making the software that implements its methods available for free on ibm.alphaworks.com. It seems to me there are competing interests here. There's the interest I have as an author to publish my work so that interesting and thoughtful people will read it and comment on it and use it for productive purposes, thus giving me valuable feedback and new avenues for application of my ideas. There's the interest IBM has as a publicly traded company to produce value for its shareholders and make a return on the investment it has made in its research assets. There's the interest of the reader has to get a return on her investment in purchasing the book and spending the time to read it. I believe all these interests are fairly balanced by this book and by the underlying patents that support IBM's ownership of the intellectual property described. ...
Is this a perfect solution? Perhaps not. It would be better if the reader could just read the book and do whatever they wanted to based on the ideas within it without fear of lawyers. And then somehow, magically, if the reader used the ideas to make significant money, IBM's fair share of that profit would be transferred to compensate the company for its foresight in investing and supporting our work over the years. But this is the world we have. And in that imperfect world, I don't regret writing this imperfect book (and I hope Pearson [the publisher behind IBM Press] doesn't regret publishing it!).
Scott closes by saying "I don't know if I have persuaded you to keep reading, but I sincerely hope so." Sadly, while I understand his point of view, I'm still not willing to cross the line they've drawn in the text. Hopefully, I'll hear back from their business development person with a clear statement that Free/Open Source software developers can implement the ideas contained in the book without fear of patent litigation.
I hope that IBM press, and any other parties involved in this book will see the light and specifically allow individuals or companies to use the information that they have worked so hard to publish. Until that happens, I can only recommend that you stay away from this book and look for other alternatives.
23 comments:
Apparently IBM has patented certain data mining algorithms? Or they think they have...
At any rate, that's radioactive, I wouldn't want to continue reading it either. It would seem that the only viable audience for a book like this would be IBM employees.
Yeah, that's pretty bogus. The result is that it's no longer a useful book -- instead, it's effectively just a sales brochure for a rather dubious business (software patents, that is).
Ah, good to hear the Big Blue hasn't entirely changed over the path of lightness and good. ;) If nothing else, it can become a bad PR move.
Gah, I need to say more, maybe I'll blog it, also.
Why did they even write the book then? it sounds like they expect someone to implement their ideas further for them, and share the profits? Honest to god are they that crazy? Do they really think what they have is so groundbreaking that they need to grasp onto it tightly?
Screw "not reading it"; are you going to return it?
I think IBMs hope is people (companies) will license the patents. What they are saying is, "In this book you can learn about these techniques. Try them out and if you find there is commercial potential, call and ask for licensing information." No?
"And then somehow, magically, if the reader used the ideas to make significant money, IBM's fair share of that profit would be transferred to compensate the company for its foresight in investing and supporting our work over the years."
This already happens in the form of revenue from book sales. The 99.9% that buy the book and never make significant profits based on the ideas presented make up for the fair share of the .01% that build a stellar for-profit search app.
Wow. Just wow.
If it were me I'd stop reading it. I look forward to hearing what happens next.
I suggest that the public completely ignore their assinine claims and make whatever use of the data in that book they want. If they want to maintain control over the data, they should not publish it. Period.
In short, ignore the disclaimer.
Good job chaosmotor, because it's not like this country is, you know, litigation happy or anything
I'd go on reading, and even use the knowledge. Algorithms are not patentable here in .eu.
Hell, it would even be good that IBM have patented it, lessening my competition :D
IF they published the book any ideas in it that were not already patented would have been made UNPATENTABLE by their public disclosure of them.
(sometimes they have 1 yr after public disclosure)
http://www.vivzizi.com
That's very odd. One would think it would be illegal or at the least unactionable to explain something in great detail in a book that you purchased like that and then say "But you can't use this" in a very discrete place in the introduction.
Of course, if you found a lucrative enough application for the technology, you could host your website in a country that doesn't honor the patents...
i too, would not read it. i am embarrassed to work for ibm.
Honestly, I doubt it matters if you read it or not. If you implement their algorithm, they can sue you for infringement.
According to the USPTO website, any unauthorized making or selling of a patented invention is infringement. Ignorance of the patent doesn't appear to be a factor. I've also never heard of ignorance successfully being used as a defense in court.
So read it, don't read it, whatever. If they can look at your source code and see that you did something they've patented, they can still sue you for it.
Patents aren't the same as trade secrets (particularly once the patent has been issued). Not reading the book once you've bought it is kind of a mistake; the patents (if issued and if valid and if etc) still apply, even if you don't know about them. You can still be sued for infringement if you happen to reinvent the same ideas. Sticking your head in the ground and hoping it'll go away is rather useless. By reading the book you bought, you might learn something new and you will learn certain areas where you'll need to tread carefully.
As matt said, this is "just a sales brochure for a rather dubious business," the business being licensing of their ideas.
I hope they didn't patent the idea that the whole book is covered by patents. Otherwise, even discussing that idea could get you in trouble.
I'm very glad you posted this blog, it might save several people $45, which we should instead donate to the EFF.
Sounds alot like what Unisys did with GIF and the patent that they had on LZW compression.
Though iirc Unisys waited till GIF format was being used mainstream and THEN extorted everyone.
Guess we'll have to wait till the patent expires for the book to be useful.
Anonymous, anyone could sue you for using a patented idea whether the book you found it in contained a disclaimer or not. Actually, even if you came up with the idea on your own, they could still sue you.
Yes, if you violate a patent you're unaware of, it's still a violation. However:
* Willful violation (violating a patent when you know it's a violation) is generally "rewarded" with stiffer penalties. As a developer, you want to stay as far the hell away from reading patents as possible.
* This book in question is (apparently) an introduction to patented ideas and algorithms -- why would you read it if you know you can't use any of it? If you don't read it, you're much less likely to use the specific patented IP that the books contain.
Oh for god's sake people, just read whatever you like, eliminate any evidence you did so, and if asked about it - LIE!
It's not as if these people deserve respect or honesty.
Patents, by their very nature, are a form a public documentation of an invention. You have to admire their inventiveness in packaging their patent portfolio as a reference book. You could get the same information for free by reading the patents, apparently.
Thanks for the warning.
Post a Comment