Monday, June 05, 2006

Library Mashup Competition

The library vendor Talis just announced a library mashup competition.

It's a pretty wide-open thing. You can use any source you want--Google, Amazon, OCLC, Z39.50*--and do anything you want, so long as it's nifty. You don't even need to work in a library (although the necessity of saying this is troubling!). About the only hard rule, is that you need to release it under some sort of copyleft license**. The winner gets £1,000, the runner-up £500. The contest ends August 18.

Best of all, I'm going to be one of the judges. This has a down side—I can't enter myself. But it will be very fun. The other judges are a pretty august group.

Although I can't enter it, I WILL do some mashups. More importantly, I'm going to start releasing APIs that others can use to build their mashups. As many of you know, I'm constrained by the Amazon API. Offering an API to the full LibraryThing data set would inevitably involve releasing Amazon API data. So I'm going to have to stick to ISBNs, LibraryThing codes (like the "work" number) and user-generated data, like tags and such. That shouldn't be a problem, since user data is what LibraryThing is all about.

Talis has set up a discussion area for the contest itself, another for ideas and another for entries. But feel free to talk over here too, particularly as regards what data would be fun to extract from LibraryThing.

*I pushed for them to include that in the suggested materials list. Z39.50 is a TREMENDOUS resource, almost completely ignored because it's a little wiggly to work with.
**The small print says "winning entrants will need to satisfy the judges as to the spirit and rationale behind their licensing decisions, prior to prize money being made available." As someone who recently signed a financing deal, I'm getting very wary of small print***. I can tell you that I'm going to push hard to allow any copyleft license to apply.
***Abe managed to slip in an "Andorra" clause, a yearly tribute of wine, four hams and forty loaves of bread. That's okay, my employment agreement mandates a water cooler filled with Guiness, in invisible ink on the back.


Anonymous Paul Miller said...


I, too, look forward to you being a judge... :-)

I also look forward to seeing some great entries.

On your specific point about the licensing of any winning entry, this is not intended to be restrictive. Rather, it is intended to ensure that the whole community can benefit from the ideas and innovation showcased during the competition. We want everyone to be able to learn from the efforts of the best.

As such, I'm fully in favour of your "push[ing] hard to allow any copyleft license to apply."

6/05/2006 11:14 AM  
Anonymous Anonymous said...

Um, for those of us less technologically incline, could you please define mashup? Because it means nothing to me.

Maybe I've got a great idea, and don't even know it.

6/05/2006 6:06 PM  
Blogger RJO said...

I don't have the programming skills to produce anything, alas. (And I'm still mad at Steve Jobs about HyperTalk, which could have put us ten years ahead of where we are with web programming, but don't get me started on that....) But here's an idea for someone to try perhaps:

With a Z39.50 collection from a major research library it would be fun to create a few small animations of "The Spread of Printing in ___" (Europe, America, Africa, etc.). Just pull the oldest printed works from each country or region, sort them by date, pull the place of publication for each and plot them one by one. If it has to be a static map, you could do it with a deep color for the starting places, and then progressively faded colors as time goes on.

6/05/2006 6:38 PM  
Blogger Tim said...

Sorry. I should have explained the term. A mashup is when you take data from two different places and do something interesting with it. For example, taking CraigsList and Google Maps, and showing where appartments on Craigs List are on a map of Boston.

The idea here is to use some library-related (or maybe book-related) data, and mash it up with something.

6/05/2006 10:17 PM  
Anonymous Ankur said...

Hi Tim ... I'm an Indian student and I intend to participate in the contest ... I am using isbndb and amazon and yahoo api as of now ...

My question is in regards to

"More importantly, I'm going to start releasing APIs that others can use to build their mashups"

When is this going to happen? ... Is the API going to be in REST/XML-RPC/SOAP format ... I hope it's in Rest format there by bringing down the entry level barrier ... I am looking fwd to ur API

6/06/2006 12:23 AM  
Blogger Tim said...

Well, I want to hear what APIs people want. What I can't give is the book data itself, which is mostly Amazon's data. But I can give ISBNs and user data, like tag percentages.

6/06/2006 2:59 PM  
Anonymous Jenny Spadafora said...

Mashups are really gaining ground in library land. "Compelling Content Combinations" is the theme for the Internet Librarian conference this year

6/06/2006 3:18 PM  
Anonymous Woodrow Jarvis "asim" Hill said...

What I can't give is the book data itself, which is mostly Amazon's data.

Which is what I really need to do citations off LibraryThing...
...OK, sour note off. THANK YOU. I suspect I can find plenty of potential uses for this data (*Evil grin*);

6/06/2006 7:20 PM  
Anonymous Ankur said...

user data along with ISBN would be nice too ... one can implement a one way syncronization,importing one's LibraryThings book cataloge ...

6/07/2006 12:25 AM  
Anonymous Magnus Enger said...

OAI-PMH could also have been mentioned as an interesting source of (meta)data, perhaps?

6/15/2006 11:28 AM  

Post a Comment

<< Home