CIL2009: A Super Searcher Shares 25 Search Thoughts

Speaker: Mary Ellen Bates

links are at

  • online magazine rack
  • she’s comparing alltop to early yahoo, just add rss!
  • rss aggregator
  • built by “2 guys and a gal”
  • highly selective, well-done

Think about how you can use this in your own organization…

  • visualization and clustering and metasearch…
  • one of those silly swirly visual search thingies (not a fan)
  • claims it’s more immersive feeling
  • You eventually DO get text
  • gives you a choice – viaual, clustered, text, etc… good.

  • adds lexical variants
  • on the fly, you can get:
  • synonym suggestions
  • singular/plural
  • translate terms into other languages
  • does clustering, but only in German
  • interface a bit squirrely

  • a way to skim across web 2.0
  • query example – she did a GTD search…
  • ok. sort of a metasearch for 2.0-ish sites like blogpulse, youtube, twitter, technorati, etc

  • clustering on demand
  • with a choice of sorting algorithms
  • and a choice of search engines
  • cool graphic display
  • looks like the old northern lights search engine! With the folder clustering thing
  • but allows you to choose HOW you want to cluster

  • add prefer:word to query
  • ranks these search results higher
  • a cool way to change the relevance ranking – doesn’t narrow the search

awesome highlighter

  • highlight text on a page
  • saves a copy of the page with a new URL
  • then you can direct others to that page with the highlighted text

textrunner search

  • looks for assertions
  • information mining
  • ex: what kills bacteria in google – lots of stuff. In textrunner, it looks for a sentence with an answer.
  • so it’s looking at the web in a different way. It’s looking at sentence structure instead of focusing on different words

Google Translate

  • translates text into other languages
  • shows text side by side
  • so you drop in search results, it translates your words into words in other languages, then shows the results side by side

Twitter Venn

  • compare frequency of words in twitter
  • generates venn diagram
  • visual way to see this

  • searches all wikipedia articles
  • does a more visual search of it
  • it structures the wikipedia article in a more user-friendly way
  • gives a tag cloud for similar articles
  • lists out recommended articles
  • gives a visual mindmap display of related stuff


  • another visual thing
  • it’s doing clustering

  • federated search (she sped through this one)


  • a tutorial
  • learn to love social media
  • can you:
  • ID the most popular blogs on a topic
  • rank the blog posts
  • eliminate content overload
  • check out the hotness of each post
  • etc
  • Cool – I’ll have to find this and pass it around

How to build a social media cheat sheet for any topic

  • also from readwriteweb

Legal Research Engines

  • cornell law library
  • google custom search engines
  • searches legal stuff


  • aggregated the front pages of newspapers around the world
  • [me – hee. this won’t last much longer]


  • makes a visual tag cloud from text
  • good way to visually see the underlying message or tone of soemthing you read

Google’s search wiki

  • you can comment on search results
  • you can move things around
  • it’s public – your annotations, anyway
  • you can customize your search of google…


  • skipped it


  • it automatically clusters and starts asking you questions


  • looks at wikipedia
  • it’s a sense-making search engine
  • does clustering, looks for sentences similar to your search

  • beta search engine
  • lets you weight your search results
  • looks like a search/tag cloud – you can change the weights visually by changing the weight of the font. Nice.

get conference buzz

  • bloggers live blog, live tweet, etc
  • So check those things out – technorati, google blog search, twitter search, etc

Google audio indexing

  • speech to text indexing

Google Maps Mashups

  • very interesting map mashups!

The ReadWriteWeb needs Sexy Librarians

In December, the awesome blog ReadWriteWeb posted a couple of great articles about how librarians are needed (and even linked to Michael Porter’s flickr photo of Michael and yours truly battling it out on Guitar Hero). That’s all dandy!

But the ReadWriteWeb just posted Deconstructing Real Google Searches: Why Powerset Matters … I’d add “real BAD Google searches” to that title. Sure, the point of the article was to point out the perils of current search engine searches/results, and to show why a semantic-based or a natural language search engine would be better. And ultimately, that really might be the case.

But my librarian self kicked in as I was reading the post, because the author obviously needed the help of us sexy librarians! Here are the search examples given:

  1. what are movie spears made out of?
  2. car hit by bicycle
  3. Famous science fiction writers other than Isaac Asimov

Librarians… I ask you. Are these good Google queries? Hmm… I’m hearing a resounding “not.” :-)

And this is a great example of why we’re still needed. Yes – there’s the web. Yes – there’s Google. And yes – there are extremely smart people that write great blogs like the ReadWriteWeb. But does that mean everyone knows how to search? What happens if the semantic web or true natural language searching kicked in tomorrow – would that negate us? No – we’d still encounter people asking why they get 50 million hits when they type “I need to find stuff on cars” or whatever into search engines.

I’m thinking we can improve the ReadWriteWeb‘s search examples mentioned in the article – let’s have some fun and help them out (not that they’ll notice, but heck – we can try, can’t we?). So – here are my “better” suggestions on structuring the three search queries:

  1. what are movie spears made out of? Why not try zulu extras spears instead?
  2. car hit by bicycle – how about “bicycle accident” “hitting car” or car “hit by bicycle” or even “car damage” bicycle?
  3. Famous science fiction writers other than Isaac Asimov – hmm… why not try “science fiction author” famous -“isaac asimov” instead?

I found better results … but I don’t consider myself to be an expert searcher by any means. What do you think? How can we improve those searches? Librarians, show your awesome search skills! How would YOU do the three searches?

Alternative Search Engines List

This, of course, SHOULD be coming from a librarian… but whatever. The Read/Write Web has a great list of alternative (as in, not Google) search engines. It looks like it’s a monthly feature on the blog.

This month includes some cool stuff, like:

  • FindSounds, an audio search engine
  • PureVideo, and video search engine
  • A variety of clustering search engines

And towards the end of the article, there’s a “Top 100″ list of search engines. Check out the article!

Quintura, a Visual Search Engine

searchI just discovered Quintura, a new visual search engine interface (found via Robert Scoble’s blog). On the surface, it looks similar to Grokker or KartOO (two other better-known visual search engines).

You can play with Quintura’s online demo, but the real deal is downloading their Quintura Search product. It offers a visual map of searches done using an impressive number of search engines, including Google, Amazon, Ask, MSN search, etc.

The biggest downside for me? Sorta silly, really… but I got extremely excited in a geekd-out way when I read their “What is Quintura” description. Why? Because they mentioned LIBRARIES. Here’s what they said:

“Have you ever raked through the paper card index of a big library? You have to find the necessary letter (or their combination), take the correct drawer, and start sorting through the cards. Ring any bells?” (ok – they obviously haven’t been in a library for at least 10 years – I’d be surprised if most of their customers have seen a CARD catalog. But oh well…).

Then they go on… “But on the web, what do you do? You are on your own. Until now. Quintura is the very know-all librarian!” (ahem… Yikes!).

Then, they have an imagined conversation with a “favorite search engine,” which I think is supposed to be the search engine you usually use (ie., Google, Yahoo, etc.). And the conversation is about finding … books on physics. Hmm…

OK – besides not having stepped into a library in awhile, they also don’t get that Google, Yahoo, etc. can’t REALLY find books in your library. Or maybe this thing was written by those college students who actually think you CAN find books in a local library using a search engine.

Anyway… I was really hoping, with all this library/search engines comparison text, that the search software I downloaded would interface with, say, MY LIBRARY’S CATALOG. Now, that’d be really cool. But no luck – it only gives me that list of search engines I mentioned earlier.

Hopefully, they’ll create a version of their product that can interface with localized search engines, OPACs, etc. here’s hoping!


Youtube and Finding Video Tagged IL2006

Anyone else tried hunting for videos from Internet Librarian 2006 in Youtube? It can be done… but it’s not easy!

Here’s the easy way: just click one of the links below (send me your youtube URL if I have left you out!):

Or do a search in Youtube. I did – I entered the tag il2006 in the Youtube search box, pressed enter… and found MUCH MORE than just il2006 tags. Other things I found?

Apparently, when you do a search for a tag in Youtube, it looks in the tag fields… but also looks for a partial match in other fields, too – hence finding many instances of “il” in the From field. It also found “il” in the title field a few times… I wonder if it’s because of the number in the il2006 tag?