(Update: I spoke to the Apra Wisconsin folks at the end of March and had a lovely time.) I have been asked to speak to the folks of Apra Wisconsin at a virtual event sometime in the near future about Web search. I’ll be talking about people search because the audience is prospect researchers, but I also want to discuss a little of my philosophy about searching and why I’m creating my Search Gizmos.
I’ve spent the last couple of days wandering around and thinking about it, and this morning I got up and thought about it a little more. Eventually I realized I’d have to write things down to get them all thought out.
I want to talk about four principles: metadata as contextual boundary, authority as search guide, interest as a directing force in exploratory searching, and overlapping strategies. Each of those is an extensive discussion, so I’ll present it in four parts, starting with…
Metadata as Contextual Boundary
The most common definition I’ve seen for metadata is “data about data.” Everything has metadata attached to it. Metadata for people might include things like birth date, birth location, name, parents, and so on.
Consider Louisa May Alcott and some of her metadata. She was born on November 29, 1832 and died on March 6, 1888. Usually we think of those facts as standalone and singular. You might look up LMA’s birth date for a report, or you might check her death date on a reference site to see if it provides reliable information. You might even use those facts in a Web search to limit your results to reference-type searches.
You might think of a singular bit of metadata as a dot. In this case it’s metadata that’s handy for reference searching and finding reliable information, but it’s still just…. a dot.
But what if you connected two dots of related metadata – like, say, a birth date and a death date – and then used them to define the context of a Web search? Then you have a line. In fact, you have a contextual boundary!
A contextual boundary lets you carve out a particular part of the Web to search. In applying Louisa May Alcott’s birth and death dates to a search, you’re saying “I want to find everything within the boundaries of November 29, 1832, and March 6, 1888.” In other words, everything contemporary to LMA’s life.
Contemporary Biography Builder
Google News and other news search tools have ways to search within date spans but they generally don’t offer those tools in relation to people, so I made one, the Contemporary Biography Builder. It uses Wikipedia data to build lifespan-bounded searches in Google Books, Chronicling America, and other online news sources.
Here’s what a Google Books/Newspaper search for Louisa May Alcott looks like for all dates:
The first result is dated over 60 years after LMA died, and the second one 18 years after that. They’re going to be very different from the articles that were published during her life. I’m not saying these searches aren’t useful, I’m saying the cultural and media assessment of a person changes after their death. It has to. First the person dies, then their cultural context – their generation – dies out. After that any assessment of the person and their work is a matter of research and reassembly, not memory and experience. Of course it’s going to be different!
To contrast, here’s what a Contemporary Biography Builder search of LMA looks like in Google Books/Newspapers:
Instead of “Famous books your children should read,” it’s society gossip and a death in the family. It’s a completely different tone of news because it’s contemporary.
If you’re a Lousia May Alcott fan you might know that in addition to children’s books, she wrote what we would consider thrillers today – suspenseful stories with decidedly adult themes. Her fiction dealt with, among other things, violence, revenge, and drug use. These stories, often referred to as “blood and thunder” tales, were written under pseudonyms and generally unknown until they were rediscovered and popularized by the late researcher Madeline Stern. Stern’s work was covered with a wave of articles in the 1970s, with some additional articles written about them in the late 1990s.
I wasn’t sure if LMA’s suspense writing was known in her lifetime, so I decided to use Contemporary Biography Builder to see if I could find out. I set up a Google Books/Newspapers search for her name and then added the keywords blood and thunder.
And I got a tasty result right off the bat!
I wasn’t sure this search would work since the keyword search is for an entire page at a time and I couldn’t be guaranteed a meaningful proximity of the keywords. This looked promising, though. So I clicked on it, and…
It turns out that almost exactly six months before LMA died, the Boston Evening Transcript published a rumor that she’d written a thriller and also noted a story called “Monica” (which I didn’t see any further mention of, alas.) I can’t find any further news related to this article, but apparently there were some rumors floating around Concord!
I was able to find this search so easily because I used Contemporary Biography Builder and made the years of LMA’s life a boundary to my search. If instead you do an open, undated search for Louisa May Alcott blood thunder, you’ll find out the results are more about Madeline Stern’s scholarship than LMA directly. Here’s an example of what that search looks like:
Birth/death metadata is not obscure; it’s easy for the user to understand and implement into a query. You don’t need to have a clear search goal in mind (though lifespan-bounded searches can be useful for that too.) It’s a simple way to radically change the Web space you’re searching.
But does creating contextual boundaries always have to involve more than one piece of metadata? Not necessarily. You can create useful contextual search boundaries even with one item of metadata; let’s talk about Obit Magnet.
Contextual Boundaries and Obituaries
Searching for obituaries online seems like a simple task: enter a name and maybe a death date in a search engine. Often you can get useful results that way, especially if you’re researching an uncommon name. But if you’re searching for a Smith or a Williams or a Jones you might find yourself quickly overwhelmed and frustrated.
My Search Gizmo Obit Magnet uses one piece of metadata – a death date – and creates two time-bounded queries: one for seven days after the death date, and one for fifteen days after the death date. Those queries are turned into search URLs which search Google News, Chronicling America, and other news archives. This is a useful and simple way to restrict search results when you’re searching common names and doing genealogy research.
Here’s an Obit Magnet search result for Louisa May Alcott:
As she has both a famous and an unusual name, LMA is not a difficult obituary search. But I made Obit Magnet with 7- and 15-day search spans because I wanted to see if there were ever followups to obituaries – personal reactions, responses, or even corrections.
And in this case I found a sweet story, again from the Boston Daily Transcript, about a boy whose father is about to read him a chapter from LITTLE MEN before the evening paper comes and they find out that LMA has died. And because this is still only a 15-day search, this story popped up right to the top of the search results.
Other Metadata Options
It’s not just dates that you can use to create contextual boundaries for your search, either. Backyard Scholarship uses higher education institutions in a radius of a location to find information proximate to a famous person’s birthplace, workplace, place of most productivity, etc. As you might imagine, Boston University has a lot of scholarship about Louisa May Alcott.
In this case we’re using location and higher education institutions’ specific top-level domain of .edu to, once again, carve out a section of Web space to explore that should, thanks to the use of metadata, yield richer results.
As I polish my skills and look for new ways to tunnel into Web collections, I have found that using authority has helped a lot. Not mine, but the authority of other groups and institutions who specify, gather, and publish information. In the next part we’ll talk about Marion’s Monocle, Super .Edu Search, and Congressional Social Media Explorer.