Code4lib 2012

Code4lib 2012, Seattle

I was the lucky winner of the Index Data lottery (no actual lottery took place) to go to Code4lib 2012. I was a (Code4lib) Newbie, so I didn’t really know what to expect, but reading Jakub’s blog about his experiences, it sounded like great fun.

It was also my first time in Seattle, so I did take some extra days on both ends to do some exploring. Arriving on Saturday to sunny and warm weather (15 degrees Celsius warmer than Copenhagen, nice!), Seattle did its best to welcome me.


I am currently working on our Harvester and Local Unified Index (LUI) products, the latter implemented using Solr, so I was really excited about the half-day pre-conference on new features in Solr. Sadly Eric Hatcher wasn’t able to make it this year, but we had a good Q&A session. A lot of wisdom came from Tom Burton-West with his experience with the Hathi Trust Indexing.

I have a personal interest in maps, so I went to the GEO session in the afternoon. Mike Graves and Tim Shearer, University of North Carolina, Chapel Hill held a very good introduction to Geo APIs and how to do Geo Hashing, e.g. turning two-dimensional coordinates like latitude, longitude into a single number that can easily be indexed and searched. They also had some great demos of the stuff. Slides


The conference started with a deeply personal keynote by Dan Chudnov, George Washington University with an introduction about being a cancer survivor, but the theme was really about how code4lib hadn’t scaled up to accommodate everyone that wanted to come this year. It was done as a one-on-one talk between a couple with: “We need to have a talk”, “I love you (the audience)”. Quite funny.

There was a lot of very good sessions, but I will limit this to the ones that stuck out the most to me.

Jennifer Bowen, University of Rochester, had an interesting talk (Slides, Download) about the road map for Linked Data in XC. On a more “Do linked data now” approach, Jason Ronallo, North Carolina State University, introduced us to HTML5 Microdata format and schemas on Slides

Tom Burton-West did a great talk about Solr indexing of Haiti Trust, too short when you are interested in the stuff, but his slides are available here (download).

The (other) Danes, Jørn Thøgersen and Michael Poltorak, Statsbibliotek / State and University LIbrary, Århus, held the presentation “Kill the Search button II” about alternatives way of navigating in a mobile application (tilting, rotating). I think we are going to look a little funny when we start using our whole body for interacting with our mobile devices ;-)

In the “Design for Developers” (Download) by Lisa Kurt, University of Nevada, we developers got some advice about doing good design (do not use a “Rainbow shitstorm” of colors and other funny expressions, I sadly didn’t get down). If the video gets up on the code4lib site, I would recommend watching it before “throwing” together some UI.

On Newbies Dinner Tuesday, I had a great evening with Tim Shearer (UNC, Chapel Hill), Andrew Nagy (Serial Solutions), Cory Low (NCSU Libraries) and Tim Clarke (Muhlenberg College-Trexler Libraries) at a nice Japanese restaurant Momiji.

Wednesday started out with “Discovering Digital Library User Behavior with Google Analytics” by Kirk Hess, University of Illinois Urbana-Champaign (download) and [“How People Search the Library from a single search box” by Cory Low]( both on what can discovered about user behaviors from statistics.

Stack View was a neat graphical way of showing search results, displaying as books with different colors and sizes depending on page count and on popularity.

“In-browser date storages” (Download) by Jason Casden, North Carolina State University was a great summary about Web storage APIs, if you are going to handle off-line results. But as with most browser techniques the “Only One” cannot be called out yet.

The Wednesday evening event was the “Bring Your Local Beer”. I sadly hadn’t packed any Danish microbrews. Mikkeller Beer Geek Brunch Weasel would have been perfect for this occasion. But I donated some Pike Brewery beers, and a growler of Hofbraü München beer from Tap House Grill (160 beers on tap). Multiple rooms were “in play”, some warnings about noise were given, but no rooms had to close down. It was a great evening. Don’t really recall my bedtime. I think this could be the cause

Thursday started with a keynote from Bethany Nowviskie about “Lazy Consensus” or how to manage/live with Committee (Design) decisions at big institutions, followed by an interesting talk about combining standard relevance ranking with static ranking by Mike Schultz Download.

Lightning talks

The lightning talks sign-up started when the conference began. I picked a Thursday spot within the first half hour, which was good because it was sold out pretty quick, and I didn’t have a presentation yet. In retrospect I should have picked Wednesday, so I could have had the chance to talk with people about it. (And also not following the “Bring Your Local beer” evening.) Most interesting/funny:

  • Andrew Nagy, Serial Solutions: “Vendors suck”. Why vendor doesn’t always suck, introducing him now on the other side as new Product Manager at Serial Solutions.
  • akorphan - Heat maps… not just for input analysis
  • Eric Larson – Finding images in book page images PDF.
  • Kåre Fiedler Christiansen (@kaarefc) – Chucking all the software components in a library together to present recorded radio and TV. Funny illustrated with Lego persons. PDF

And my own:

  • Turbo MARC in YAZ: I was afraid of running out of time but ended too soon. I should have used the last minutes for questions. Wasn’t good enough to catch this and use the time. But the word on TurboMARC has been spread. Naomi Dushay, Standford U. has emailed about doing a Ruby Gem. Last evening at the bar I talked with David Bucknum, LoC, and he would mention this to his colleague working on this.

YAZ (and Index Data) got some good karma on the IRC channel.

It was a really great experience, learning a lot of new stuff, meeting up with grateful users/customers, getting good street cred by other vendors/developers, but I really maxed out with information from talks and flooding of messages from the IRC channel. Some kind of memory expansion plugin for my brain would be nice ;-).

As a side note, I mentioned in the talk, I have done some work on creating a similar TurboMARC writer in Java. Based on MARC4J by Bas Peters, I have added a TurboMarcXmlWriter. I haven’t seen the same 4-fold improvement here, but I am looking into this. Until it has been integrated upstream, the source can be found at git://

In the beginning, I did feel a little like a stranger as vendor along with all those Universities, but now after it: “I’d love to see you all again”.

cheers, :-Dennis