I found your skepticism about search engines for science interesting. You might be interested in the extensive interview I conducted last year with Michael Alvers of GoPubMed http://www.nextgenerationscience.com/science-resources/gopubmed-interview-with-michael-alvers/.

And I think that the work that Deep Web Technologies http://www.deepwebtech.com/index.html does with the Office of Scientific and Technical Information (OSTI) of the US Department of Energy like the search engines WorldWideScience.org http://worldwidescience.org/, Scitopia http://www.scitopia.org/scitopia/ and the E-print Network http://www.osti.gov/eprints/ are steps in the right direction.

I agree with you, though, that we will need a new paradigm for science search. That is why I have tried to get the search community interested in Open Science—search needs to get into that space both to advance science and to find new markets for itself. I hope that some search people will attend Science Commons Symposium – Pacific Northwest http://sciencecommons.org/events/salon.

I explained why I think they should in the post, “A Plug for Science Commons Symposium – Pacific Northwest” http://www.altsearchengines.com/2010/01/31/a-plug-for-science-commons-symposium-%E2%80%93-pacific-northwest/ and they should definitely attend ScienceOnline http://www.scienceonline2010.com/.

I had an interesting note from Jean-Claude Bradley the other day as we both mulled over the many comments on your post, “Science and Web 2.0: Talking About Science vs. Doing Science” http://scholarlykitchen.sspnet.org/2010/02/08/science-and-web-2-0-talking-about-science-versus-doing-science/.

Jean-Claude and I both found it odd that some of the readers of that post argued that Open Science does not appeal to scientists of a competitive temperament and that Open Science is too squishily collaborative in nature to make it into the mainstream research realm. Jean-Claude made the point to me that in order for scientists to excel in science these days and to make their reputations, the first thing they often look for is brilliant potential collaborators. That is where Open Science and the new science search paradigm you mention would come in—we need platforms that would enable scientists to quickly find collaborators. Research Scorecard http://www.researchscorecard.com/ and Elsevier’s grant finding platform SciVal Funding http://www.info.funding.scival.com/ are on the right track on that.

Many of the commentators in that forum argued that Open Science is not the way for ambitious scientists to go. I would argue that those who are career-minded should do Open Science for hard-headed, calculated, career-building reasons. In an Open Science world, the results of one’s brilliance are on display 24/7 to a worldwide online audience and one’s name and work become recognized far more quickly and among a broader community than that that the specialized journals cater to. After all, how can you get ahead in science if your peers can’t get copies of your articles because the libraries at their institutions have dropped them and they themselves can’t afford personal subscriptions? Get your name and your science out there. Get cited. Get ahead. Do Open Science. Idealistic? Do Open Science. Mercenary and me-minded? Do Open Science. It works for everybody! What a deal!

The publishers seem to get the need for facilitating the search for collaborators. It is just that they are having spam problems (as The Scholarly Kitchen has noted of Elsevier’s 2Collab http://scholarlykitchen.sspnet.org/2010/02/10/are-publisher-linking-networks-choking-to-death-on-spam/

As you say, Nextbio was paid a substantial sum by Elsevier. I hope other search services will look into the Open Science space. The in-house teams could use help from such firms, apparently. It will be fascinating to see what happens with DeepDyve.

There have been many, many “search engines for science” that have cropped up over the last few years, and none have gained much traction. I’m not convinced that such tools are likely to do so, unless they offer some sort of new paradigm we haven’t yet seen. Though I will put in a plug for GoPubMed, which I think is superbly helpful:
http://www.gopubmed.org

NextBio do interesting things as well, but as I understand it, Elsevier paid them a substantial sum of money for the design and setup they use, rather than taking them on as a partner for shared revenue. One could easily see DeepDyve licensing their technology or even being bought out as you suggest, which would give a publisher a shortcut into this market.

As far as wanting them shut down, publishers want to make money. If this is a lucrative market, then they’ll want to exploit it, not make it go away.

I find your comment here quite interesting, “…each journal presents unique content” very interesting because we seem to moving to an era in which it is the article and the subject matter that matters and not the journal. That is why the instant gratification for users of the DeepDyve 99-cents model via the CiteULike tie-in (although DeepDyve is not limiting it to the alliance with CiteULike) is so interesting because users would be doing a search on a specific subject and would not care what the journal was.

For instance, I am interested in amyotrophic lateral sclerosis (ALS). You are right that journals offer unique content. The leading journal on the topic is Amyotrophic Lateral Sclerosis:

http://informahealthcare.com/aml

but there are articles about ALS in many journals such as those on neuroscience and neurology, muscle and nerve, palliative care, nutrition, etc. It is the article that matters, not the journal. The publishers still think in terms of marketing the journal and are not very good at marketing the individual article or subject matter. (Even RSS feeds on subject matter are poor for various reasons—not enough specificity, such as those from Highwire, for instance.)

Peter Binfield’s work on article metrics is fascinating in this respect—I heard him talk about that at ScienceOnline2010:

http://www.scienceonline2010.com/index.php/wiki/Article-level_metrics/

I agree with you that Google and PubMed are the gateway to the content of most publishers’ sites, but so far most the publishers have not tried the 99-cent model (although they could do so with DeepDyve even without being tied to CiteULike). I think it would be great if the publishers tried the 99-cents model on their sites so that I could indeed simply be greeted with that option no matter how I found their sites: via Google, PubMed, etc.

I agree with you that the traffic from reference managers is probably miniscule. Perhaps if the publishers got their acts together via rendering their stuff more search engine friendly and making it easier to search once you get there (like Elsevier’s employment of NextBio for ScienceDirect—and that is a case of a big publisher calling in a middleman) they probably would need not need DeepDyve or any other middleman. But I still think DeepDyve deserves credit for coming up with the 99-cent model. That at least has generated some discussion about the incredibly high cost of individual articles.

And of journals. I had a nice exchange with a journals retention executive at Informa last week about the journal, Amyotrophic Lateral Sclerosis. I wrote him:

“I am an individual subscriber. I was told last year by one of your colleagues that there is now no distinction between individuals and institutions with this journal. I watched the journal increase in price from around $350 a year to around $850 to the now unaffordable $1,100. Please do tell your executives that the market simply will not bear that kind of inflation (which is basically price gouging) no matter how much we love the people we know with ALS. You are pricing yourself out of sustainability of subscribers.

It is these kinds of rate increases that add fuel to the demand for Open Access publishing—which is not in your interest.”

He said such comments are taken seriously—let us hope so!

You make a good point here, “most big publishers have a dedicated IT staff who could build in similar functionality, and most smaller publishers outsource such things to platforms like Atypon or Highwire who could do the same and amortize costs over many different journals.” But all of those are clunky given where we are with Web 2.0 slickness and, again, don’t offer the innovative 99-cent option. But as you say, “The question is whether there’s enough of a market here to cover those costs and to replace the potential losses in pay-per-view revenues that are likely to occur.” For instance, email alerts based on RSS feeds from PubMed are great and I might routinely purchase articles from the publishers linked to in those alerts if the costs were reasonable. MIT and the University of California Press are trying DeepDyve. I just wish the medical publishers would—that is where the need is greatest vis-à-vis quality of life for real people.

Maybe Springer or Elsevier could simply buy DeepDyve, thereby acquiring the services of smart thinkers in pricing and search expertise. (Or shutting them up by shutting them down, which they might prefer.)

Regardless, most big publishers have a dedicated IT staff who could build in similar functionality, and most smaller publishers outsource such things to platforms like Atypon or Highwire who could do the same and amortize costs over many different journals. The question is whether there’s enough of a market here to cover those costs and to replace the potential losses in pay-per-view revenues that are likely to occur.

There are tons and tons of third party services out there looking to “partner” with academic publishers to provide services and take a cut of revenue. Most don’t offer very compelling services and those that do may face this question of why a middleman is needed.

I agree with you, though, that it would have made sense for the publishers to do what DeepDyve is already offering to do for them. But DeepDyve is quicker off the mark incoming up with innovative revenue models for the scholarly publishing industry and the publishers may simply be holding back to see if the 99-cent model works before simply bypassing DeepDyve and doing what you suggest: handling everything themselves. We shall see what we shall see. Thank you for making me ponder these matters!