The Ultimate Debate: Who Controls the Future of Search
sponosred by LITA’s Internet Resources Interest Group
Saturday June 24, Convention Center
Roy Tennant, California Digital Library
Stephen Abram, Sirsi/Dynix (also president-elect of SLA)
Joe Janes, Univ. of Washington
Be it resolved that the future of search will occur without library influence.
A series of sample questions, paraphrased as:
Will large commercial search services like Google replace libraries?
Should libraries cooperate with large commercial services by sharing metadata?
What would Google Scholar or similar need to have before libraries abandon their own metasearch? Or are they there now?
Will libraries incorporate the better search tools into their own catalogs?
Will user familiarity with the new tools make library catalogs obsolete?
What is your worst nightmare and finest vision for the future of library search services. How likely do you think either is?
*NB by Jason: the “more” is a near-transcript of the session…Chris did an amazing job of capturing the content. Be warned that a massive amount of text follows…Thanks Chris!
170 chairs, about 2/3 full as session began. Maybe more.
Intro by Laura Cohen, chair of interest group. Open meeting 10:30-12 at Marriott NO. Fleur de Lys room.
Holly Yu vice-chair
Panel needs no intro (repeat of brochure material)
Presidential debate format. Opinions do not nec reflect those of the speaker or his organization. Six topics for discussion.
Moderator to intro issues and ask questions in alternating fashion.
3 minutes to respond, 2 minutes for rebuttal. Flexible. Governed by tin bell.
Audience invited to produce questions.
Thanks for making the five-mile walk.
Big companies are challenging us in disturbing ways. This is intended to discuss and provoke.
By coin flip Janes to answer first.
Will Google replace libraries?
Market decision. If libraries let it happen, yes. If Google satisfies and is available and is tailored to what people want, yes. Libraries couldn’t handle Google’s volume. At all. We can’t do all of what Google does. Adult Services is not a good Google search. Possible growth opportunity. Users will vote with their feet. This isn’t necessarily all bad.
Who cares? Libraries weren’t ever about search. They were about community and learning. The top websites aren’t about search either. Networking stuff like MySpace is the new big thing, much more like what we do. FB & MS have 90% of college students. 40% of Internet traffic is MySpace. Search is not the issue. Are we creating that for our users? MySpace is less than a year old–it’s an experience.
Search is powerful. Community is a good point. Yahoo bought Delicious and Flickr, so a big deal. “Algorithmic search” is not everything. Question answering services acknowledge our point about what users need, but no large market for this. Search was A reason people came to libraries, a tool. But rarely the point. People for whom search is the point become librarians (mild para).
Search services write algorithms to meet marketing needs, not users needs. Lots of folks think sponsored links are higher quality. Localization of Google Maps is a market driven phenomenon. GS and MSLive are driven by the prime demographic, 18-34. Ads to them are previous.
Popularity doesn’t really equate to quality. Page Rank in the early days was an ingenious solution to not knowing how to do quality. Google results make sense; we haven’t made the case that quality info is worth the extr effort. Google Scholar and Google Book are attempts to index the parts of the web they previously couldn’t see.
Google does what. Libraries do how and why. Commercial services may not be able to. But we are not scalable on what questions. Book search is NOW. 50% of all books in Chinese.
What does GS or MSLive need?
We know boolean logic isn’t teachable. Toilets have electric eyes, even for intelligent conference goers. The interface needs to be smarter. (much laughter). Facts have a half-life of eight years. Water boils at varying temperatures at varying altitudes and pressures and water purities. The comml services are not collaborative. Our important problems will be solved by teams.
Flushing is a good thing. No fund disagreement. Comml services are working on the social network (blogs, Flickr, MySpace, etc). For the young this stuff is very powerful. Microsoft invited a whole bunch of librarians over to explain why their search sucked. Fish in a barrel. People in their teens and twenties live online–tagging and sharing are really important. These tools could be used to do business, also. The world is ahead of us and comml services.
Should libraries abandon their metasearch tools?
The comml services are like old TV and are desperately trying to break out of the single user orientation. People solve problems in small groups, not demographics. Libraries can work with groups of 4-6. Like gardening books for folks in Toronto–that’s what we do. Information isn’t learning.
Comml services have our users eyes. Libraries can have their bodies. You can have a meeting at the library.
Should libraries cooperate by sharing metadata/other content?
Who cares? let people discover us. They’re not in our market. Our competitor is ignorance (mild para). “I want their hearts, minds and aspirations…Google gives you an article to read.”
We’ve done this. Own vs. license for journal content. How do you explain WorldCat? What if it were open source? (ref to the Fortress of Metadata in Dublin). The Internet community could add a lot to what we’ve built over 40 years, an “unimaginable amount”–nothing, or everything. Google is planning to make stuff available and see what happens. Our goal, too. We have been contrained as to how much we could make available; this is beyond what we thought was possible. Google per se doesn’t change lives–what you find might. Google can connect the dots much more quickly.
River vs. rock. The river always wins. See Grand Canyon. OpenWorldcat now has everything, with links to local content. People don’t want to find things through metadata, they do through pattern matching. They expect it to work like Google. Librarians and users have no boolean overlap in values. We are arrogantly deciding that users want perfect privacy, rather than the kind of services Amazon provides. Why can’t we let them decide for themselves. Our value system is in the way of a modern search experience. Where is the middle ground?
How do we manage privacy so that we have nothing to turn over to Homeland Security? Because of those concerns we don’t have the data we need about our patrons. Search logs, circ records, other things we could data mine (in a good way). This data has never existed.
What about LibraryThing, which has the 79th largest library in the country? Users have decided to share this. Why are we not letting users decide this?
Younger people more likely to do this. He agrees, but to be contrary: A century of fighting for privacy of circulation records to ensure reading anything. The Attorney General is probably listening. It’s a legacy of how we have done things, for good reasons. But things are changing and we have automated what we are currently doing. The first online card catalogs had little dots where the pole would go. If we’re assuming that the phone company and Microsoft and everybody is going to give us up do libraries need to? Maybe, but there is a cost.
Praise for Library Connection folk in terms of Patriot Act. We need to have a conversation about this, but not in an environment where if you say “terrorist” anything is possible. Searched how to make an atom bomb?
Searches for Islamic beheading websites. Three clicks. Very easy to find. Horrifying. But will having done it cause me problems?
Will libraries add new features?
Aquabrowser, Endeca at NCSU. It’s already happening. How do we teach people how to use it? Users think a list is the epitome of search results. Visual interfaces work best for users–but librarians tend to hate them. We think we’re the alpha user. Tag clouds are a good bridging tool, but librarians don’t like them either. Engineers love it, because that’s how they learn. Providing visual context is great–and Google doesn’t do that. We can. Do you want Mercury the god or mercury the element? Google doesn’t do that. Search engines can’t figure out how to place ads on visual results. We don’t need to do that. Mention of Clusty as doing some of this.
Not soon enough. Summary of blogosphere: What the hell gives? No intent to summarize. The Apologist (Abram: we owe users one): Nothing pathological about the library business. Well, cats. Sensible shoes and bags with cats on them. People who loved searching became librarians. DIALOG is impenetrable (I wrote that stupid book), but it’s librarian porn: power and control over information–precision. Professional knowledge that we have an no one else does. Catalogs were designed from the beginning to represent that. Algorithmic ranked search looks easy–but to experts it really “fries our cheese”. Our skill set appears to be less relevant–people needs us less than we think they do now (which is a load of crap).
No debate yet on how librarians killed Dialog, a monument to obsessive compulsiveness that eventually no one could ever use. At the end of dialog searching you could squeeze a diamond out of your ass. (Janes and how satisfying was that?). The opac is an inventory system. KGS is right: libraries are broken. OPACs are not the problem. Focus on the user and what experience they want. Stop popping out diamonds and ask what experience they want. Put the results in the context the users wants.
I count five diamonds on the floor. Agrees with KGS. The OPAC can’t change because MARC won’t change. There’s only so much you can do: there’s so little there. No text. Very cryptic precoordinated headings. in 1966 it was beautiful. As was OCLC. In their time they were brilliant. Algorithmic search can’t work on them: no text, no hyperlinks, totally different structure.
The point is to tie it to full text somewhere. That’s what the user wants. Users don’t care about MARC records.
Violent agreement. A failure of imagination. We’re stuck in our traditions, and conservative. We have the human record at stake, and we’re trying to be careful–what are the problems? Will JSTOR be around a hundred years.
Studying something to death is not a good idea because death is not our goal.
What is worst nightmare and finest vision, and do you think likely?
Will answer a different question. Vote for me, he’s a jerk. Worked the last two times.
What would perfect search look like? (lousy question). Answer: no search at all. Lightest possible search. Web searching tools are easier. ROI is an interesting calculus, but compare with OPAC.
Worst nightmare: nothing changes. We wither and die. Likelier than he’d like. There may be a lack of will to make scary changes.
Finest vision: put Google out of business. what we do is good enough to make people forget about it. they return in droves of thousands and millions. Confidence: Eh.
Worst nightmare: the profession keeps saying: your side of the boat is sinking. Librarians need to stop blaming each other and their vendors. More people are going to college than ever did before. 30% vs. 12%. A generation that’s very different from what we’re used to.
Finest vision: Better advocacy. If you cut libraries, this is what will happen. If we let trustees fire librarians for collecting in Spanish in our country we should be livid as a profession. Tired of collecting statistics. The very tall guy flew in on an airline designed for the average person by an MBA based on stats. Back to the sixties, when programming and service was the priority. Technology can now be a tool again. Confidence: we’re really smart people. we can stop being so control-oriented. we can let the boomers do what they came for. we can empower the next gen of librarians. Don’t knock library school students–they know what they need to.
Google is not the competor.
Open to the floor for questions
Q: Libraries tied to the publishing industry. Much more being created that isn’t like that.
Abram: Prepare to allow tagging on everything, post-its all over the website, radical trust for users. We need to do more of this. Web 2.0 is about when the user becomes predominant. We manage the ecology that allows this to happen. Let the users decide where to go.
Janes: Look at human history. Cave paintings. Blogs and wikis. Everything in between. People standing up and saying I matter. Helping people do that is our business; helping people find other people’s stories is what we do. Publishing used to be a barrier (barring vacation slides). It is now much easier to speak and be heard. The publishing industry used to be the only game. The internet is what the vertical file used to be. We used to be describable as people of the book; not as much now.
Q:(Rep of Xrefer) With all the Google stuff coming out, what does this mean? Does this increase the demand for online reference? Google may not be able to help users make sense of things.
Abram: supply and demand are not relevant. things in demand become larger on the web; sense-making is not possible as you get more stuff.
Q: Comment on book slips, which used to have names on them; maybe it’s time for them to come back–privacy is recent. Relevance can be done with OCLC holdings symbols. What is a quality result for a typical search and how do we program our systems to do it? Typical example question: Why do adults play?
Abram: what makes a quality result? Ask the user. In corporate libraries the end user makes the judgment. On OCLC popularity: the top 1000 OCLC titles and then top 100 sellers and the top 100 borrowed there is NO overlap. Some overlap between buy and borrow. We don’t buy what users do. User purchases circ seven times more often. Elephan in the room. Walmart analyzes this stuff.
Janes: think of the size of diamond an elephant could produce. Ref librarians answer questions, and pummel people until they give you a question you can answer. The guantanomo theory of reference. IPL experience: either very complicated questions with no answer or questions they were always wondering about. Lots of that goes to Google. The really hard questions you can’t answer, but you can move them forward–here’s another step on your path. Wikipedia is for how and why questions (leaving aside questions about it). Don’t worry about answering the question. Here are the lives we changed, the vacations we planned, the experiences we created.
Abram: any ref librarian will admit that the questions are getting harder. the information cowboys: bring them on, I can answer them alone. how does answering those questions fit into the larger context of the chain of info our users are using. Desk stats are OUR experience, not the users.
Q: Consumer orientation of librarians. We feature what we buy and what we think is cool. We make information overload worse.
Abram: We are focused on the tools we provide. a builder asks how you want to feel in your new addition, doesn’t tell you what hammers he’s using. Libraries should be able to tell a story about how you will feel better when you come. Latch-key moms who dump their kids at story time may be working on a BA and have a great view of us while we sneer. “Here’s our stuff” is not an experience.
Janes: the experience has to be personal. Librarians are frequently nameless: people don’t trust that.
Abram: virtual users don’t allow us to offer the context. A ten year old asking about divorce is writing an essay; a forty year old has a different interest. We always ask.
Janes: understanding is the key to that. it doesn’t automate. Google doesn’t understand. Ref libs understand that people don’t know what they’re looking for; Google can’t help, we can. A ref interview can really create possibilities that a Google search can’t.
Q: What about users who are not physically present? How can you build a web space that works like the physical version?
Abram: On the privacy side, the library card gives you a lot of useful information. IP identification is dead–we need individual users in order to build meaningful tools. If you’re a sophomore bio major I shouldn’t send you the Shakespeare librarian. Mixture of parts of Sakai and Shibboleth and athens. The people in the library and outside it are not the same people. Virtual users can handle tech. Use the data we have, safely and respectfully.
Janes: I don’t know. Look at Ann Arbor district library aadl.org as a thought provocation. The website is about what the library offers. You can’t provide the same services online as in person–forget it. Use the strengths of each and be clear what goes where.
Abram: stop being afraid of success. we can’t tell patrons about ILL they might use it. aadl postings get 500 comments. deal with it.
We can’t do everything. One size fits all fifty years ago worked. We had a niche we served well. We can’t do that anymore. We have competitors who will do things better than us. Ready Reference, which we loved to do and were great at, just doesn’t make that much sense anymore. Better ways to do it. Perhaps libraries have niches that are not as broad. Pick the areas where we can do best. Rethink what we are for. What do we not do? How do we use those resources? That’s hard. We want to be everything to everybody, but we can’t.
Why: improve people’s lives. We need to build a capacity for change based on respect. Get rid of 18th c. models of org. Bring the back room out front. Bring the cowboys at the ref desk into teams. look at what the next gen of librarians can do and need (and don’t smother it). Don’t be afraid of play–we learn more from it. Kids get 50 minutes at a time; only we’re foolish enough to think we can concentrate from 8 to midnight. Take 15 minutes a day and do 43 things at 43 folders. It will help us adapt to change, and in six months we’ll know much much more. You’ll be more fun at parties.
Ilaughed, I cried, I couldn’t think of a better way to spend two hours.
Well, at least we have a definitive answer to who controls the future of search.