Wolfram|Alpha's Surprising Terms of Service

eldavojohn notes that Groklaw is highlighting the unexpected Wolfram|Alpha ToS — unexpected, that is, for those of us accustomed to Google's "just don't use it to break the law, please" terms. Nothing wrong with Wolfram setting any terms they like, of course. Just be aware. "We've seen people comparing Wolfram's Alpha to Google's Search from a technical standpoint but Groklaw outlined the legal differences in a post yesterday. Wolfram|Alpha's terms of use are completely different in that it is not a search engine; it's a computational service. The legalese says that they claim copyright on the each results page and require attribution. So for you academics out there, be careful. Groklaw notes this is interesting considering some of its results quote 2001: A Space Odyssey or Douglas Adams. Claiming copyright on that material may be a bold move. There's more: if you build a service that uses their service or deep-links to it, you may be facilitating your users to break their terms of use, and you may be held liable."

Comments

That's pretty standard

[hedwards]

They aren't claiming ownership of the bits of data they provide, they're claiming copyright over the whole page. Sort of like how an encyclopedia will copyright the book even if it includes quotes from people. Basically over the presentation of the data.

Additionally much of what they would be claiming copyright over isn't subject to copyright protections. Things such as birth dates and astronomical data aren't subjected to copyright protection.

Re:That's pretty standard

[PCM2]

Then I guess you should have read the actual terms before you posted, hmm?

Attribution and Licensing

        As Wolfram|Alpha is an authoritative source of information, maintaining the integrity of its data and the computations we do with that data is vital to the success of our project. We generate information ourselves, and we also gather, compare, contrast, and confirm data from multiple external sources. Where we have used external sources of data we list the source or sources we relied on, but in most cases the assemblages of data you get from Wolfram|Alpha do not come directly from any one external source. In many cases the data you are shown never existed before in exactly that way until you asked for it, so its provenance traces back both to underlying data sources and to the algorithms and knowledge built into the Wolfram|Alpha computational system. As such, the results you get from Wolfram|Alpha are correctly attributed to Wolfram|Alpha itself.

        If you make results from Wolfram|Alpha available to anyone else, or incorporate those results into your own documents or presentations, you must include attribution indicating that the results AND/OR [emphasis mine] the presentation of the results came from Wolfram|Alpha. Some Wolfram|Alpha results include copyright statements or attributions linking the results to us or to third-party data providers, and you may not remove or obscure those attributions or copyright statements. Whenever possible, such attribution should take the form of a link to Wolfram|Alpha, either to the front page of the website or, better yet, to the specific query that generated the results you used. (This is also the most useful form of attribution for your readers, and they will appreciate your using links whenever possible.)

        A list of suggested citation styles and icons is available here.

      Failure to properly attribute results from Wolfram|Alpha is not only a violation of these terms, but may also constitute academic plagiarism OR [emphasis mine] a violation of copyright law. Attribution is something we expect you to give us in exchange for us having provided you with a high-quality free service.

        The specific images, such as plots, typeset formulas, and tables, as well as the general page layouts, are all copyrighted by Wolfram|Alpha at the time Wolfram|Alpha generates them. A great deal of scholarship and innovation is included in the results generated and displayed by Wolfram|Alpha, including the presentations, collections, and juxtapositions of data, and the choices involved in formulating and composing mathematical results; these are also protected by copyright.

        You may use any results, including copyrighted results, from Wolfram|Alpha for personal use and in academic or non-commercial publications, provided you comply with these terms.

        If you want to use copyrighted results returned by Wolfram|Alpha in a commercial or for-profit publication we will usually be happy to grant you a low- or no-cost license to do so. To request a commercial-use license, go to this form and provide the input for which you want to use the corresponding output along with information concerning the nature of your proposed use. Your request will be reviewed and answered as quickly as practical.

DISCLAIMER: These results are the property of Wolfram|Alpha.

Re:That's pretty standard

[Bakkster]

Don't see what the big deal is, here. Since Google doesn't host any of the actual information, you don't need to cite them as a source. You do need to cite the page you get to from Google, though. Think of W|A like a procedurally generated encyclopedia/textbook/almanac. Just like any of those other sources, you should cite it as a reference.

The sooner people stop associating Google and Alpha in their heads, the better.

Re:That's pretty standard

[DragonWriter]

From the terms of service:

In many cases the data you are shown never existed before in exactly that way until you asked for it, so its provenance traces back both to underlying data sources and to the algorithms and knowledge built into the Wolfram|Alpha computational system. As such, the results you get from Wolfram|Alpha are correctly attributed to Wolfram|Alpha itself.

If it didn't exist before I asked for it, and my asking for it was the only human action that caused it to come into existence, if there is an "author" for copyright purposes, it's me. The only way Wolfram could, therefore, claim copyright on it is if it was a work for hire, but since I'm not a Wolfram employee acting within the scope of my employment, and since there is no agreement signed by both parties designating it a work for hire, that doesn't work either.

Consequently, I'd say their own terms of service defeat their claim to copyright.

Nothing to worry about for academics

[jw3]

All they ask is that you attribute them when publishing results derived from their service. Example:

Methods: "The comparative population studies were derived from the Wolphram Alpha service (Wolphram, 2009)"

Regular thing for academics. I cite NCBI blast service, I cite PFAM, I cite dozens of other services out there. Most of these tools require or ask for an attribution; and in most cases, this is anyways necessary in a scientific procedure.

j.

Re:Nothing to worry about for academics

[forand]

How is it reasonable to ask for attribution for having a computer perform a calculation on someone else's data? Wolfram Alpha has do nothing except code a turing machine, I do not cite HP when I do a calculation on my calculator and I see no reason why more complex but equally wrote calculations should be. I ask the computer a question and it gives an answer, is the question or code used to find the answer the insightful/citable part of the idea?

Of course.

[PCM2]

Wolfram Alpha doesn't just provide you with knowledge. It provides you with a new kind of knowledge. Any knowledge you gain from it must be attributed to Stephen Wolfram ... because he invented it. It is actually safer to attribute all citations to Stephen Wolfram, in fact, because he is smarter than you.

Re:Wolfram|Alpha just killed their business

[ScentCone]

Hope they are not expecting to make any money by selling out their Customers at the drop of a hat.

How are people who show up to use a free service "customers?" Google's customers, for example, are their advertisers, not the people who use the free stuff.

Does it cut both ways?

[Bill Dimm]

If attribution is required because (according to the TOS):

In many cases the data you are shown never existed before in exactly that way until you asked for it, so its provenance traces back both to underlying data sources and to the algorithms and knowledge built into the Wolfram|Alpha computational system. As such, the results you get from Wolfram|Alpha are correctly attributed to Wolfram|Alpha itself.

Does that mean that Wolfram|Alpha can be sued for slander if its algorithm generates a false statement about some individual or corporation by "misunderstanding" the data it is digesting? In other words, if the result is something uniquely generated by Wolfram|Alpha, deserving of attribution in the same way that an author of a book deserves attribution, do they also deserve to be held liable if the content they are generating is incorrect or slanderous?

Oops! The Vulcan Academy cheer is now copyrighted

[Otis P. Foont]

I innocently entered "Secant Tangent Cosine Sine 3.14159" into WolframAlpha. The result, 74.69263, now belongs to Wolfram. Sorry about that.

deep links

[zarathud]

I note that Wolfram|Alpha happily deep-links to Google Maps.

Re:Hah!

[Xest]

Well it claims to make information computable. I accept it's not meant to find results like Google but the issue with it is it doesn't even seem to gather basic data in a computable form.

I mean, you try things like "On what date did the Falklands war commence?", "How many species of Melocactus are there?", "On what date was Adolf Hitler born" and it outright fails.

Okay, so I figured maybe I'm asking questions that are out of the intended realm of knowledge it supports and the assumption is that you'd never want to compute with this information. So I tried something Mathematical - I mean, that is Wolfram's speciality right?

"How many non-isomorphic labelled trees are there with 4 vertices"

Fail.

I've tried a few other relevant, factual questions and it just falls flat over, not even able to try and answer them.

I'm sure it does do a great job of making information computable, the problem is it's unable to gather the information in the first place.

Ironically, Google, that doesn't claim to make information computable manage to provide answers for all these questions within it's first page, often as the first hit. Sure it may not be presented in a standardised format, but data that needs to be parsed is certainly more computable than data that simply can't be provided at all.

I can see what Wolfram was trying to do, but why did he have to couple it with immense hype that it's as important as Google? Why has he been going on and on about it to the media when it struggles to even do what it's supposed to absolutely excel at? I think they could've at least saved face if they'd stopped being so cocky about it and released it with a little less hype and fanfair and let it improve and become more useful and hence more greatly adopted over time. One has to ask when there was so much hype about it and with a ToS like this whether it was all just about Wolfram gathering data for himself or something than providing a tool useful to everyone else. Either that or he simply beleives his own hype and believes the tool is better than it really is. Perhaps in developing and using it himself he was blinded in making and seeing it work well for applications specific to what he wanted without ever truly seeing how well it performs in other problem domains?

Re:Nothing new for Wolfram

[The_Wilschon]

Well, it's hardly surprising, as Stephen Wolfram is a well known egomaniac who refuses to admit that anyone other than himself can possibly achieve anything. http://www.cscs.umich.edu/~crshalizi/reviews/wolfram/

Re:Hah!

[B1oodAnge1]

More importantly, it completely fails at this question: http://www26.wolframalpha.com/input/?i=historical+popularity+trends+of+shaved+genitalia+in+pornography

I was looking forward to the graph too :-\

Re:Nothing new for Wolfram

[registrar]

Could you give some examples? Not that I'm doubting you, I'm just curious.

I've been left without access to mathematica licenses on multiple occasions due to misunderstandings between Wolfram and my institution. Because Mathematica was my primary platform at the time, that meant days that I was unable to do or access my work.

The first time that happened, I decided to learn an open platform; the second time, I migrated. In my projects, I now absolutely avoid writing core functionality in Mathematica.

Another complaint: you can't discover how defaults work in some cases. As far as I can tell, setting things to "Automatic" means "proprietary and undescribed." I've asked Wolfram for details in one case, only to get a "we can't tell you" response.

Oh, and being told off for filing bug reports is pretty unimpressive. I separately reported different manifestations of the same bug, separated by some time. I'd actually forgotten about the first report, but if they'd fixed the bug, the situation wouldn't have arisen. When I've submitted a bug report to open source projects, they have usually been along the lines of "this line is wrong, and this seems to be an acceptable fix."

I think the arguments for open, modifiable, redistributable source code (that is guaranteed to retain those properties) are extremely strong. I.e. the GPL, probably v3. Once you know it well, Mathematica is a stunning programming language and library set, but I now don't care: as a whole, the platform has been unreliable for me.

Re:slashdot

[Hal_Porter]

http://www.wolframalpha.com/input/?i=(sinc+(x)+*+sinc+(y))+

I'm sure you could do better if you had more time.