Science Online '09: Searching the Scientific Literature
Our session wasn't quite as planned - I had hoped to demonstrate some things. The slides are
online.
First, John and I described what a librarian in a university and one in a research lab does. Most people have some idea of what the main jobs were for their elementary school librarian. Some people who have children know that public librarians do story times sometimes (this is 1/100th of what public librarians do)... but unfortunately, most people really have no idea what librarians in corporations, research labs, and even in universities do. So we talked about that.
Then we wanted to talk about some big themes like:
- making connections between things - getting from a need to an answer or from a citation to the full text or from a collection of citations to a publishable paper
- how librarians can be your best resources in educating your peers about open access, in getting support for open science,
- in universities we can help you get your classes going or consult with you individually on your own research
- even if you have no university affiliation, you can walk in and use public university computers (public might not be the right word in your neck of the woods, but state funded like State U or U of State) and most of their subscription resources. AND you should always use your local public library and get assistance from the librarians there. They also have research databases
We also talked about some spiffy research databases, ebooks, and some free tools online.
I mentioned Inspec and Compendex, of course, because that's the way I roll (engineering and applied physics and all) - my point with Compendex was more about the interface, though.
John talked about Safari and books24x7 - ebooks
I talked about CRC Handbook of Chemistry & Physics - the online version with tables you can sort and now substructure searching. This is another of my major themes: helping you mobilize and *use* information, even information that's found in books. To this end we mentioned searching in google books or on amazon to find the content, and then using "find in a library" on google books or a special bookmarklet or browser plugin to find the book that's listed on Amazon. Your library might even provide you with an electronic copy, right there at your desk, even if your desk is at home (so long as you have proper credentials), and if not, might deliver the print copy to a location that's easy to get to (our staff get books at their mail stop).
If I remember other things, I'll add them.
Labels: scio09
Science Online '09: Sunday AM
Reputation, authority and incentives. Or: How to get rid of the Impact Factor — moderated by Peter Binfield and Bjoern Brembs
I was frustrated by the premises of this session: that the IF is inherently evil, nobody likes or uses it, and everyone agrees that it should go away. I was happy to hear someone piping up that they do indeed use it (if only as a first cut) and it does have some value. Like any other mathematical formula, it isn't inherently evil. It's only as good as the data that go into it, and it can only provide so much information. The problem is that it's abused and misused. It can be a useful tool if used as part of a much larger set of metrics. For example, for collection development decisions, when combined with local citation measures, subject matter expert opinion, cost (per page, part of package) measures, topical relevance, local usage, appearance on syllabi/course reserves, if a squeaky wheel is on the editorial board....
What was even more frustrating is the idea that one measure could be easily replaced with another simple measure, that we could brainstorm in one hour. Sigh.
On the positive side, I think it was Gee of Nature who suggested compling multiple inputs into single pages -- ideas of adding in commentary from the blogosphere, pre-prints, and other things seems useful. From this we somehow got around to uniquely identifying authors. We discussed on friendfeed the merits of developing a new tool or adapting openID or like to scientific purposes. Along these same lines, the value of having a consortium of publishers manage this (like DOI which is very successful) or using some other open model.
----
Providing public health and medical information to all — moderated by Martin Fenner
Martin prepared a very useful framework and posted it to the
wiki page.
Doctors post for doctors - as a filtering mechanism. These filtering mechanisms are quite common in clinical medicine, like Cochrane Reviews. Doctors also post for the public, to comment on new research.
--
then there was our session, sigh. It needs its own poast
and I went home.
Labels: scio09
Science Online '09: Saturday PM
Ok, here's the weird part. I distinctly remember being at two different sessions that are on the schedule for the same time slot. Huh? So there might be some time travel involved here or maybe I'm imagining something...
Web and the History of Science – moderated by GG, Brian Switek and Scicurious
very cool session - but there were some deeper issues here that the audience did press that were not totally addressed. There's the myth of science - really probably mostly post-WW I - that paints this picture of science as clean and linear. Many popularizations reinforce this with 20-20 hindsight. Most popularizations also only hit the "cool" and the weird (there's a neat Fahnestock article on this - she's at UMCP, but I've never met her). Some of the bloggers go more deep and talk about controversies and the complexities of good scientists working hard whose science was later found to be faulty (or to not adequately describe and predict once better instruments became available). Seems like some bloggers just go: ha-ha look at those crazy old guys! One of the moderators was unable or unwilling to go deeper and that sort of set the tone, unfortunately. If/when this session is done again - bring in someone from rhetoric who looks at the language of popularizations or doing science history to ask tougher questions of the bloggers who do posts on this.
Social networking for scientists – moderated by Cameron Neylon and Deepak Singh
(how I could have attended this when it was at the same time as the above? beats me). Interestingly, these scientists came up with Rogers' innovation adoption decision process for communication technologies :) This was a very worthwhile and useful session, though. Is there value for having separate science networks, or should scientists just use the general purpose ones? Seems like if the network is built around *things*, and these things need special treatment, then science or scholarly networks might be in order. One example is myExperiment - this is built around sharing of workflow pieces - you can't do that well on facebook because you need special metadata, searching, and attribution. Likewise with citeulike or connotea - they are better for scholarly articles than delicious because they understand what metadata is required to describe scholarly work. Otherwise, sites that are just like linkedin, but have a smaller user base really don't offer anything over the general space, and might offer less if they can't get to critical mass.
Anonymity, Pseudonymity – building reputation online — moderated by PalMD and Abel Pharmboy
This is a perennial favorite. I think people who follow blogs at all get this: that you get authority and trust over the course of the blog through your posts. That your pseudonym becomes your brand and is meaningful and trustworthy based on your history of posts. This might be better than relying on your institution or journal IF for authority (IF, ha!). People who don't know blogs, and don't read them, really don't seem to get this. And there are legitimate reasons to not use your real name, even if people know it anyway. Also assume that you can be found out, no matter how hard you try to stay anon.
Labels: scio09
Science Online '09: Saturday AM
My very much delayed notes from the sessions I attended Saturday.
Open Access
Moderated by Bill Hooker and Bjoern Brembs
http://www.scienceonline09.com/index.php/wiki/Open_Access_publishing/
What’s open access – green vs. gold models…
for university and disciplinary repositories, they can be listed in directories and have machine harvestable/federated searchable content
citation advantage and acceleration of the research cycle.
also allows for text mining
example: iHOP, information hyperlinked over proteins – works amazingly well, but can only work on the abstracts, and needs to be able to work on the full text
serials crisis – costs rising >> consumer price index
Can’t access own article
member of audience – databases in NAR database issue, exponential growth (published in Scitable open access) – duplicating efforts – how to you prevent duplication
6 major publishers control most of scholarly publishing
my point – if journals are dead, why proliferating? Societies as well as (or perhaps more so than) commercial publishers
journals go on your cv – where you’re published
- journal quality as a proxy for researcher quality or research quality
- is using journals for assessment such a good idea
- do you want to delegate that filter to someone?
- idea separate the research that we’re doing – dissemination – vs. making judgments of people for promotion. separate those things out
- this was already done in particle physics – pre-publish, even before the net by hand dissemination through the mail, then through e-mail, ArXiv… everything in particle physics is open access, but this didn’t change the market in scientific publishing – didn’t have anything to do with making the science available, but for the point of reputation, credit, etc. Market wasn’t created by evil companies, it was created by the people working in universities – everyone has to find a place to published – even if the work is very poor. Overlap btwn ArXiv and the journals is 100%, but need the certification, etc., comes from the published edition.
- criticism of making money – business model – vs. spirit of open access.
business model
- is open access really more expensive in total because management decentralized vs. centralized within the library, who has more or less figured out how to manage licenses based on subscriptions
- shifting the cost from the library to the individuals is sometimes resisted, but sometimes ok – depends on the field and if they’re accustomed to paying page charges
So, this session covered a bit that was really appropriate to the impact factor discussion. Unsurprisingly, there were still those in the audience who wouldn't publish OA because their advisor bought the line about OA != quality. This is just like the advisors saying ejournals != quality when in fact nearly all journals are in electronic format and the format has nothing to do with the quality.
Another time this week someone who works at a big evil society tried to get me to make some absolute statement about OA and I just wouldn't take the bait. I think for certain communities, it works quite well (for example, HEP and biomed) while for others there are a lot of very real barriers. I don't think libraries are necessarily the place to manage the article payments, but some libraries have taken on this task and have been able to support their researchers this way. There's a question about scalability.
There was a very interesting contribution to LIB-LICENSE this week from someone who actually went to Kenya and talked to some users of the literature. He found that they were getting just about all they needed from HINARI. So the LDC or developing countries argument might not be the best one - the argument might have more weight for smaller institutions in developed countries -- they seem to be the ones who can't get what they need.
-----
Not just text – image, sound and video in peer-reviewed literature
http://www.scienceonline09.com/index.php/wiki/Not_just_text/
This session was very helpful in distinguishing for me the difference between these two things and how they might be valuable.
Two versions: You Tube, and Journal Like
Moshe P, from JOVE
journal articles are very inefficient at transferring tacit or craftsman like knowledge, such as how to do certain experimental techniques
“golden hands”
he had to fly to the original lab to get expertise and then bring it back.
need video publication – show me
like cooking – small things omitted from the recipe that you can get through watching someone prepare it or by watching a cooking show
questions:
- what incentives are there to publish science on videos
- what format would be most useful?
- what equipment is required
incentive – make it a publication, a scientific journal, peer review, indexed in PubMed
tools – they do it for the scientists. Distributed network. Specialized outsourced video production companies with expertise and equipment to capture.
each article begins with a schematic representation, then an introduction of the scientist
(questions – Java – can I embed – have to do it by e-mail them, bcs, they had people ripping off the entire content of their site)
are the authors given help in speaking to the camera?
(real question – about widening audience – but this isn’t about this at all, this is for an expert audience which should indeed be full of expert language, somewhat incomprehensible outside of the exact area)
Scivee.tb
- science video sharing web site – synchronize video, literature, slides – within the browser
- more discussing a paper, vs. showing an experiment
- profiles and community to connect scientists to each other
- changed – people wanted to upload stand alone science videos (lions on the savannah or conference presentation)
- poster casts, slide casts
questions to scivee:
- requirement that poster already be accepted/published elsewhere first?
- they contact conference organizers to have a private community where these are available to conference attendees by password for some period of time and then open later
questions to JOVE
- do they get the text with what the equipment is and where the consumables came from and such answer is yes
- is this new methods – or is this methods that already exist in the literature – could be both
- how does peer review work – goes to 2 or three reviewers who give time stamped comments
- is this multiple or duplicate publication? not really because you’d have a methods paper vs. the results paper
funding models:
- sci vee- trying to go to conference organizers
- jove: advertising, author fees ($1k) when they produce for you
question about animal research
- they review carefully special board
- nevertheless face associated with animal research – firebombings, attacks etc.
- they’re concerned about self-censure and the science being put out.
(question that occurs to me now - people judge other people by their names and institutions and stuff, of course, but also about personal attributes - when you know CK Pikas is a white woman in her 30s, does that make your think any differently about her work? so here's the question: is there value or what difference/impact whatever, does it make to see and hear the scientist with the protocol? Ideally using Mertonian norms of universalism, it makes no difference whatsoever... but people are really funny.... if you had the same or equivalent protocols done by an older white man and a younger minority woman, would the reaction and the use be the same? I really hope so! Maybe it's horrible that I even speculate? The choice to have a little intro of the scientist before the protocol is an interesting one - presumably to help the viewer trust the protocol.)
----
Semantic web in science: how to build it, how to use it
John Willbanks
semantic web rdf
triple – subject, relationship (directed arc, with label), object
literals vs reification (“has category”)
bootstrap using grddl (“griddle”) – extract rdf, using parser, from sql database or whatever
owl stuff- types of relationships (symmetric,
sparql (queerly language)
unlike xml, pull statement out and it still makes sense – can remix
licenses are really nasty – even if intended to be open, can prevent some remixing bcs eventual user might be corporate?
need a public domain- in US databases can be public domain (not so in Europe)
cc zero – zone of certainty for semantic web – certify types of reuse that are ok.
viral licensing actually causing problems… need this new way to make things usable in a semantic web
neurocommons – their proof of concept
- e pluribus unim
- this domain bcs pubmed/nlm non-copyrighted data
uses:
dns for life sciences
api to the public domain of data
enhanced document markup
activity center analysis
working with Microsoft to like spell check articles as they are written so all of the connections to the data are in tact
(all life sciences all the time - actually, I think astro and earth/planetary sciences have some goodies like this, but the people who work on these sorts of things apparently don't show up at this meeting --- it would be great from someone from ADS to show up and talk next year)
Labels: scio09