uncool uris
On December 18th I was asked to shut off lcsh.info by the Library of Congress. As an LC employee I really did not have much choice other than to comply.
The lcsh.info domain was registered by me in order to demonstrate how the Library of Congress Subject Headings could be represented as a Semantic Web application using SKOS . In particular I was eager to get feedback on how the data was being published with respect to Linked Data best practices. I got lots of great feedback, wrote a paper which I presented at DC2008, and learned that other institutions like the W3C and the Royal Library of Sweden were beginning to use URIs for concepts from lcsh.info in their metadata.
It was always my intention for concept URIs at lcsh.info to be cool. I advertised the service as ‘experimental’ and indicated it was going to hopefully inform the development of a similar continually updated service at LC. I had the good fortune to have a shared server with Kevin Clarke, and others from the code4lib community where I could spend $5.00/month on making the service available.
My thought was I could leave the service running until there was something similar at LC that I could redirect the concept URIs to. After a year or two when people had rewritten their data to point at loc.gov I could retire lcsh.info. I never imagined I would be asked by LC to take it down. Some people who have been around the block a few more times than me saw this coming (you know who you are) and I apologize for not taking your concerns more seriously.
So here’s this blog. I put it here so you could leave your comments and thoughts. Feel free to comment on this post, or start a new post by registering with the site. It should accept your OpenID if you have one. Your input and criticism are most welcome.
LC is still considering running a service like lcsh.info at loc.gov, but it’s not there for me to link to yet. Please accept my apologies, and leave your comments (however brief) here.
Ed, this is a huge shame. Getting lcsh.info going was a huge step, and I and the guys I know in the SemWeb community were really excited. This must feel like a total kick in the teeth. I feel for you.
Is there any way we can show LC what step backwards this is? Let me know.
mauvedeity
That is clearly a loss: the data is of great general interest around the world. Can you talk about the reasoning behind the decision? Were there serious issues of intellectual property rights being infringed, or other legal matter? Or was it a case of an organization not having that rare skill of being able to recognize and nurture innovation and the champion of forward ways of thinking? I ask partly to see whether other outside LoC could help, partly to understand whether an external group could acquire or extract the same thing and keep it running independently, maybe just as a frozen snapshot. What is the IPR on the data? As US government data is it presumably in the public domain?
A great shame, but you can also know that you have already done a great service to the world just by firing this up, showing the potential, and getting people interested and eager to work with linked library data. The traction you helped start will continue with other forward-thinking ideas (like the recent release of University of Huddersfield’s Library circulation data). Eventually, your work will be recognized and this project will get new life.
Good luck,
Patrick Murray-John
While it’s very sad to see this site go offline, I believe it is only a delay before the inevitable. I believe LCSH will inevitably become part of the Web. I just hope we’re looking at days or weeks rather than months, years or decades.
Ed did a great job here showing a very practical path for resources like LCSH to find a new life in the Web. The site made great use of linked data techniques, of W3C standards like RDFa and SKOS, and it was starting to see some use for experimental cross-referencing between LCSH concepts and other systems, from Wikipedia/dbpedia to FAO’s Agrovoc. It will be missed.
Perhaps lcsh.info has served it’s purpose, and in 2009 we’ll see the Library of Congress build on this experiment in a more ‘official’ way. If Ed has bent any rules in making this site, I hope he can be forgiven. His enthusiasm here has definitely been infectious, and the work was much discussed in conferences and workshops I’ve attended lately. It would be a natural role for the Library of Congress to be championing the use of such techniques amongst all bodies who maintain large public sector descriptive vocabularies, and to take over and use lcsh.info as a showcase and testbed for exploring the interlinking of these datasets through Semantic Web techniques.
This is a great pity, we were just getting excited about using this fantastic service.
(in the Delft University of Technology Library Digital Product Development Department, and in the Dutch cultural heritage community)
The lcsh.info service served as an extremely useful example on how to do things. As a matter of fact I was in the process of setting up a similar service on http://iconclass.info/ copying (for now, quite literally) the way things were done. I was hoping to create some tools that could use lcsh.info and iconclass.info together, and then inspire other large classification system to do the same and gather momentum for getting more Linked Data out there.
Willing to help in any way in which I can assist in making a convincing case for an officially sanctioned lcsh.info service by the Library of Congress.
Exposing LCSH data in this kind of way is incredibly important to the continued value of LCSH as a controlled vocabulary in the contemporary environment. Not having it available in this way seriously limits it’s use in the contemporary environment. I share the hope that LC provides this again in weeks, not years.
In the current environment, the belief that you can make money by controlling and selling important parts of the library metadata infrastructure—it’s a business model doomed to failure. In part because it dooms your ‘market’ (that is, libraries) to eventual irrelevance. And the horizon for ‘eventual’ is fast-approaching.
[...] is therefore with great disappointment that I read this on the lcsh.info site the other day: On December 18th I was asked to shut off lcsh.info by the [...]
I originally posted this on the Thingology blog but I thought it might be appropriate and useful here, too:
My library’s subscription to ClassWeb (paid access to LCSH) expires every year in October. Every year in October we automatically send payment. Every year they shut off our access, saying that our subscription has expired. It takes us MONTHS to get it back–weeks just to get a live response from anyone at LOC and then only to be told that it will take several weeks to renew our subscription.
This year, over the months that I have been without ClassWeb, lcsh.info has saved me numerous times, basically by being available and accessible to look up subject headings and especially their relationships when ClassWeb was not. I do try using authorities.loc.gov, but lcsh.info gave me more of the information I needed, with a much easier and faster search interface.
To me it wasn’t even about the experimental techniques, though I grant I was interested in those as well. it was a simple case of access and availability–lcsh.info made subject headings available to my library and allowed me to accomplish my job when a service from LOC (which my library pays for) could not.
@timbl and @danbri you are asking the right questions. Unfortunately I am not in a position to provide definitive answers…but I will do my best.
All I was told was that Library Services (the largest organizational unit at the Library of Congress) found lcsh.info to be “confusing” and that it needed to be turned off. In large part I think your questions, and my inability to answer them, highlight something I did wrong with lcsh.info. Despite the support of various management level people at LC I couldn’t see how to get the service through the normal process at the library in a timely enough manner for me to get feedback on the implementation…so I sidestepped the process.
This process would have involved ironing out the licensing issues which you are asking about, figuring out which unit in the library was going to “own” the service, getting software through a security audit, convincing everyone that SKOS was useful. In short, it would have required organizational commitment, and a strategic goal. I avoided all this by getting the informal OK from Barbara Tillett (Cataloging Policy & Support Chief) to use the master LCSH data in my conversion process, and by hosting the service on my own server.
I have operated under the assumption that (like all the data output of the Library of Congress) the LCSH is in the public domain. I know that work is afoot (to my knowledge mainly by Barbara Tillett) to update the 1902 statute that requires (actually it’s “authorizes”, thanks cjmconnors) the Library of Congress to do cost-recovery for the distribution of the data (at that time catalog cards). It is my understanding that this statute is preventing the Library from simply giving away the public domain data, like LCSH.
Since the SKOS data at lcsh.info was a derivative of the LCSH MARC data that is being sold perhaps some people thought that lcsh.info would fly under the radar. As my colleague Dan Chudnov pointed out:
lcsh.info was my attempt to implement the ideas of Barbara Tillett and Corey Harper about the deployment of place, name and subject authority data (library assets) on the web using semantic web technologies. I got the OK to work on this as a prototype, and took on the public presentation myself to get the valuable feedback. I’m personally hopeful that the fact that I could do what I did with 200 or so lines of python, and a shared colo server for $5/month has opened up some eyes to just how out of date the 1902 statute is.
I don’t feel like I’m answering the important questions you raised. I would like to think that the engagement of people like you and Dan Brickley will help LC express what its strategic goals can be for its data assets. I don’t think there are any legal reasons why someone couldn’t buy the data, and use my minimal code or something similar to provide the service. But I think LC wants to get to a place where it is providing this service. We just need help in getting there.
Hi,
I’m not incredibly well informed in this area, but reading through makes me ask two questions:
1) is it impossible to host the resource at another domain? - if so I can offer and would be prepared to fund one.
2) is this an Intellectual Property issue? - if so we need the equivalent of an OpenStreetMap project.
rgds,
Richard
richard@caliban.org.uk
—-
Richard Rothwell
http://www.cockspiracy.com/
@Richard the legal issues are kind of blurry for me (I’m a programmer not a lawyer) so I’m not really the best person to ask. I *believe* that there’s not much stopping someone from crawling (as Simon has done) the data from the authorities.loc.gov ; or from purchasing the data and distributing it. The OpenLibrary is well positioned as the OpenStreetMap equivalent for the library world.
Ed, I am really sorry about this. You and lcsh.info have been a great inspiration for us at the National/Royal Library of Sweden, and you are still welcome to visit us to talk linked data anytime you want.
I am not surprised somone at LC thought lcsh was “confusing”. Still, I think you did the right thing by sidestepping the “getting-the-project-formally-authorized” process. Lcsh.info have been important and helpful at least for us in Sweden. Linked data is, I think, the obvious direction for libraries. Time will prove you right!
I am not sure about the legalities, but judging from personal experiences I think its very probable that “confusion”, not intellectual property issues, was the reason to shut down lcsh.info. Since LCSH is in the public domain (at least in the US), I don’t think there would be much problem if someone else put up the data elsewhere. This would probably be even more “confusing”, but if it was done by someone not employed by the LC, I don’t think LC could order that someone to shut the service down. I’d love th see the OpenLibrary do this! It might be good to talk to a lawyer first, though…
Ed, thanks for the response. I hope that the “do it externally” route proves unecessary. The service would receive much more use (at least in the library scene) if it is somehow eventually blessed by LC, I’m sure.
Regarding the cost recovery statute, … I’m no lawyer and haven’t read the relevant texts yet, but I hope there is a reading of “cost recovery” that could take into account some wider notion of cost-recovery.
* money saved on software development costs
* money saved on QA and user testing through collaborative development
* money saved through external creation of mappings and adjunct data
* money saved on marketing / documenting LCSH data services
* money saved by providing a platform that makes it easier for outside bodies to propose amendments or revisions to LCSH - eg see my experiment at http://wiki.foaf-project.org/SanfordBergmanLCSHScorecard
If LC open up this dataset as a linked information hub, I expect we’ll see sufficient collaboratively-produced software, services, and cross-referenced data, all of which can be of direct value to the LC in fulfilling its mission, and which can help offset the costs associated with LCSH. In it’s few months of existence, lcsh.info I think has already provided fair evidence of this.
lcsh.info was a wonderful service, and will be much missed. I’m happy to do anything I can to help and I hope that LC will reconsider their direction and start encouraging their work be put on the Internet instead of working to keep it off.
First: Ed, you did a truly great thing. You inspired lots of people so your work is not completely undone by shutting down the site.
Second: Wow. The fact that LCSH.info was useful from the second it was launched should cancel out any “confusion”. I am, however, too familiar with this specific type of confusion. For a few months LC played a leading role in the real world outside “library technology”. I hope that the project is still seen as something positive and wonderful within LC, anything less would be kinda scary.
Also, people over here at the National Library of Sweden were getting pretty excited about things like automated updates of *parts of* authority records through linked data. They will be writing letters.
See also: Epic fail — Not just any failure - not just an “oopsy” - but a full stop, game over failure.
I can only reiterate the disappointment that others have expressed at this decision. Ed’s work - realised, as he notes, with a relatively modest amount of technical effort - illustrated the great potential of Semantic Web technologies for enhancing access to, and use of, valuable data - both within and beyond the library community.
I only hope that it does indeed turn out to be the case that LoC builds on this exciting initiative and (promptly) offer a “non-confusing”(!) Web source for this data, using the technologies & conventions already so effectively demonstrated by Ed in the lcsh.info project.
[...] with shock [1, 2], anger [3], and disappointment [4, 5, 6] at the news that LC has asked (as in “ordered”) Ed Summers to take down his wonderful experimental LCSH.info service. It’s not entirely clear [...]
[...] on the rest of the semantic web. It was elegant, had a SPARQL endpoint and a squiggly animation. Ed was asked to shut it down on December 18th by LC. Apparently because of “confusion”. Which is ironic since [...]
[...] the Library of Congress has asked Ed Summers to take down LCSH.info. Apparently because it was “confusing”. This is a shame since the Swedish [...]
I will be frank and admit I’m a simple enough soul that I don’t entirely understand what it means to represent LCHS “as a Semantic Web application using SKOS.” But I take it that was part of the point of this experiment.
There have always been very powerful opposing forces at work in our profession — some pulling us open, some pulling us closed. Right now we are feeling very forceful tugs in both directions. Two major institutions (OCLC and LC) are yanking the rope one way. On the other end are you and the people who support what you’re doing. Just know you are on the right side of history.
On page 2 of the Future of Bibliographic Control final report,
http://www.loc.gov/bibliographic-future/news/lcwg-ontherecord-jan08-final.pdf
in the context of the future role of the Library of Congress, it says:
It simply is neither feasible nor necessarily appropriate for the Library to continue to perform all its assumed roles —- particularly when considering its own demanding legislative mandate for managing its vast and complex internal collections, services, and programs.
Five areas of recommendations were made in the report, including these three:
* Increase the efficiency of bibliographic production for all libraries through increased cooperation and increased sharing of bibliographic records, and by maximizing the use of data produced through the entire “supply chain” for information resources.
* Position our technology for the future by recognizing that the World Wide Web is both our technology platform and the appropriate platform for the delivery of our standards. Recognize that people are not the only users of the data we produce in the name of bibliographic control, but so too are machine applications that interact with those data in a variety of ways.
* Position our community for the future by facilitating the incorporation of evaluative and other user-supplied information into our resource descriptions.
In every way, lcsh.info was a brilliant example of how to achieve these goals. Thank you for showing us how to do it right: presenting bibliographic information in an elegant semantic way.
In light of the WoGroFuBiCo’s report, I simply can’t imagine what LoC’s thinking was in asking you to take the service down.
[...] lcsh is no more… how sad — there were so many lovely URLs [lcsh.info] “On December 18th I was asked to shut off lcsh.info by the Library of Congress. As an LC employee I really did not have much choice other than to comply.” I really hope we will see an official version returning very soon. [...]