ALA2007

My writeups have been a bit delayed, mostly due to traveling complications. But, back to the grindstone:

I missed the start of Peter Morville’s talk on “Ambient Findability: Librarians, Libraries, and the Internet of Things”, but here’s what I did get. His whole presentation is online here:

Important questions to ask when designing a website:

1. Can users find our website?
2. Can users find their way around our website?
3. Can users find information despite our website?

Credibility and findability are interlinked as concepts – whatever links show up as the top Google results, users tend to trust as being authoritative. I’d never thought about this before, but it makes a lot of sense – I know that mindset is how I react when doing casual searches.

Then, there are also long tail searches. Peter once worked on a redesign of Cancer.gov. Most of their traffic came from search engines linking in to their central page that directs users to info on different types of cancer. After the redesign, all of those individually linked pages show up more prominently in search results – getting the user to the page they want faster, even if that distributes the stats among lots of pages. Our goal is to make information accessible, not just make a website. Search engine optimization is absolutely crucial.

The CSA search interface was greatly simplified in the recent past. This was done to ease choices for students and professors, and not intimidate them so much. But, what if the student or professor doesn’t know the database exists in the first place? Sites such as AccessMyLibrary.com have put the metadata for a number of journal articles online, in the hopes that they’ll show up in Google searches and then be able to link a user to the article via their local library. A good idea! Unfortunately, the Google results for this metadata are in “Google Hell” – so low on the list that they might as well not exist.

Web designers need to remember that we’re designing the legacy systems of tomorrow – learn from the past, design for the future.

The definition of ambient findability is being able to find anything at any time. This is unattainable, but an ideal we should strive for.

“A wealth of information creates a poverty of attention”. There are simply too many options out there for readers to assimilate them all. So, we need to be more findable – a pull model instead of pushing content out.

Meanwhile, there are small steps into making information ambient via alternate interfaces. Peter mentioned the “ambient pinwheel” – it sits on your desk and spins faster as more e-mails accumulate in your inbox.

“Bigger needles need to be put in the haystacks” – I like this metaphor a lot.

The internet turned everyone into a librarian – at least as far as paying attention to metadata creation. Tagging is a prime example. We should bring the new and the old together.

Peter closed with a story about three stonecutters working on the same project. When asked what they were doing, one replied: “Making a living.” The second said “The best job I can”. The third said “I’m building a cathedral”. Try to keep that broad, grand view when designing any kind of information architecture.

I am utterly exhausted tonight – somehow I forgot how much walking a big conference like this can involve. So, here’s my raw notes with very little elaboration. A very worthwhile session though! I know I missed writing down a lot of good points, but LITA should be posting the full session as a podcast somewhere soon. The speakers were:

John Blyberg

Karen Coombs

Roy Tennant

Marshall Breeding

Walt Crawford

Joan Frye Williams

I’ll label who said what with their initials whenever possible.

MB: Library automation is an important area.
Consolidation, investment by venture capital, etc. all bring major changes and heightened distrust by customers.
Open Source is another trend – there are now legitimately considerable ILS options that are open source. Meanwhile, where do new companies that provide support for open source products fit in?
There is a new focus on updating front end interfaces to match user expectations.

JB: Back ends need to be shored up to support new front ends, due to a ripple effect of those changes causing more stress on the existing structure.
RFID sorters and storage options’ privacy issues are illusory – they’re really just barcodes, nobody can resolve what book it refers to without direct back end database access.
There’s a new desire to uncouple the OPAC from the ILS and make everything more modular

WC: RFID can have privacy concerns if patrons’ cards are chipped too
JFW: This is trying to use logic on a political issue. We’re the only institution with an application for RFID chips that won’t sue protesters, so we are their prime targets when they’re really fighting against Walmart and that sort.
MB: What about backwards compatibility issues with future generations of RFID chips and readers?

KC: The end user as content contributor is a growing trend – this has implications for the archives of local history, etc.
-Where are today’s equivalent of WWII letters written to families at home? It’s all electronic and ephemeral.
-Libraries should try to capture some of this kind of thing
-Ex) Australia’s Picture Australia Project – partnership with Flickr. Users submit pictures to a Flickr group, some are chosen for inclusion in the archive project.
Audio and Video are what users now want, and we need a way to deliver access digitally.
The line between desktop and web apps is being obliterated – where will future software reside?

JB: There’s no push to the semantic web (or web 3.0) yet.
KC: Bad HTML inhibits this
RT: If the semantic web really gets going, we’ll know hell has frozen over
MB: 3.0 can mean true information apps built from the ground up, not today’s wraps around legacy systems. This is a long way off.
WC: Users mostly don’t want to do the XML and such that the semantic web requires

RT: Trend of demise of the catalog – new tools unify access to a wide variety of information. Kill off the term OPAC.
Software as a service – get it out of local server rooms and onto central storage with the vendor – this eliminates upgrade issues.
Intense marketplace uncertainty aids a push towards open source systems.
Where do indexers fit in when someone like Google goes directly to the publishers and full text?
Eventually, an ILS will be used mostly for back room maintenance, not front end

WC: Privacy still matters – do patrons want us to be Amazon, do they understand the potential for government data mining of user records?
-The Slow library movement: Locality is vital, think before acting, and use open source only where open source really genuinely works (NOTE: Please see Walt’s comment below about this bit)
-The role of the public library in telling the community’s stories is changing due to the availability of publishing online.

MB: ILSes need to handle more formats than just MARC data
WC: For all the complaining about MARC, a session about it at 8am the other day was overflowing out the door. There’s obviously still interest.

JFW: End user focused technology is being used as a replacement, not a real change (I think I missed writing something down here, this note doesn’t make much sense to me now)
-Where we are currently tactical, we need to be strategic – don’t congratulate ourselves too early.
-We’re holding ourselves back by being afraid of irrelevance, which is self-fulfilling
-We have a reluctance to be involved more directly in the development cycle

JB: Direct visits to library websites will drop as mashups of library data are on the rise and used directly instead, but the site will still be necessary.

KC: Users’ interaction with information is changing, and we are responding. This is where much of the current environment of change comes from.

This talk was broken up into three pieces – one each by Matthew Bejune (Assistant Professor of Library Science at Purdue University Libraries), Meredith Farkas (‘Queen of Wikis’ and Distance learning librarian at Norwich University), and Tim Spalding (founder of LibraryThing).

Matthew opened with a summary of his research into wiki use in libraries. The results will appear more formally in the 9/07 issue of ITAL. For his purposes, library wikis can be classified in one of four groups:

1. Collaboration between libraries
2. Collaboration between library staff (internal)
3. Collaboration between staff and patrons
4. Collaboration between patrons

Groups one and two combined make up about 76% of all library wiki use.

Some highlighted examples included the SJCPL subject guide wiki, the USC Aiken Gregg-Graniteville Library’s site (using a wiki as a full CMS), OCLC’s Wikid, and The Biz Wiki from Ohio University.

As we move forward with Wikis, Matthew highlighted some questions to keep in mind:

Why aren’t we more in category 3 or 4?

How might we enable patrons to build or modify a library’s information?

How will libraries next use wikis?

His own wiki (http://librarywikis.pbwiki.com/) has many more details.

Meredith took the stage next. Her presentation is available online at meredithfarkas.wetpaint.com

She stressed the need for knowledge management. All organizations need it to collect and maintain just how to do a task and share areas of expertise. Many libraries’ systems for this purpose are too informal right now. For example, scraps of paper with notes left at the reference desk can very easily disappear. A blog might work, but the reverse chronological listing makes it hard to locate an item in the long run.

Meredith also talked about user tagging as used in Ann Arbor District Library’s catalog – these tags are more familiar to users than the LCSH system. But, tagging needs a critical mass of tags applied to items to be of much use in making recommendations. Hennepin County’s catalog lets users create lists like on Amazon, and leave comments in the catalog. Worldcat is going to allow these lists too.

A great example of a community-focused wiki that Meredith showed is the RocWiki – a big guide to the city of Rochester, NY. (near where I grew up, incidentally) Users have asked and answered each others’ questions in discussion areas of the wiki. While this one isn’t run by a library, why couldn’t a library build something like this as a community service? It would enhance reference service too.

The Biz Wiki was brought up again, as an example of drawing on faculty expertise and what students have picked up.

But a Wiki is not an easy solution. It takes time to get buy-in, run training for staff, and then integrate it into an existing workflow. Also, I got a big kick out of this library wiki tech troubleshooting page.

Later, in Q&A Meredith mentioned that some moderation by authority is necessary for online communities. But it has to be done carefully to maintain a grassroots feel.

Tim Spalding’s presentation was jaw dropping. I’ve played with LibraryThing before, but only a little bit. I had no idea of how deep its current functionality goes. LibraryThing is now the second largest library catalog type system in the world.

It relies on “social cataloging” and the idea that knowledge is a conversation. A catalog itself is not a conversation, but is a tool to get you to it. LCSH helps start this process, but then users need to ascend.

The users do an amazing amount of cataloging work for librarything, probably without even realizing it. They combine multiple editions into one authority listing, write extensive author pages (making sure to clear copyright for any pictures of them), etc. By letting users apply tags of their own design, the resulting system is a bit more natural than LCSH – ‘cooking’ instead of ‘cookery’, for example. There are also tags for genres like chick lit and cyberpunk, which LCSH doesn’t cover. And even if LCSH were to add headings for these areas, there’s no guarantee that past written books would be classified with it retroactively.

Of course, tagging isn’t perfect. What if someone labeled The Diary of Anne Frank with an antisemitic tag? Well, it hasn’t really happened. But relatively useless and infrequently used tags for the book like “in belgium” or “historyish” can be ignored when only popular tags are shown. This relies on a critical mass of tags though, enough that the oddities can be declared mere statistical blips and ignored. Any new tagging system has this hurdle to overcome.

LibraryThing has some cool new functionality coming soon, including the ability to combine and subtract tags. For example, users will be able to search for books with the tags WWII and France, but without the tag Fiction. Also, the mass of tags that LibraryThing has accumulated will soon be made available to libraries (for a fee) for use in their catalog. Tim says it’ll work with any OPAC! Danbury Library is the first to take advantage of it. He’d also like to apply LibraryThing’s methodologies to articles soon. I think I’m going to give LibraryThing a more in depth look when I get home…

I only caught a little of this session due to a conflict with other events I wanted to be at. Google’s Adam Smith opened with a walkthrough of the http://books.google.com/ website. He noted that Google has agreements with roughly 40,000 publishers to scan their books that are still under copyright. I hadn’t heard that anywhere before. Records with metadata have been added for non-scanned books too, like Harry Potter.

Representatives of each of the first five schools to join the scanning project were also on hand. I didn’t get to hear them all, but most noted that their efforts started with off site collections that in storage to minimize impact on the main collection. Harvard is including all bound volumes, not just books. Michigan scans fragile items in house and lets the outside scanners handle the rest. Interestingly, they keep their own copies of the scan in addition to what they pass on to Google. They plan to build their own interface to search the local copies – I’d be really interested to see the final result and compare the functionality of each.