Friday, May 27, 2011

On writing good search questions

Writing a search puzzle turns out to be epistemologically interesting in a way that I'd never imagined.  

I'd always figured I could just sit down, ask a few questions, write them up and be done.  Oh no, that's not the way it works.  

Mostly, the problem boils down to What's true?  What's determinable?  and How do you know?  

In essence, the grand philosophical questions for all time.  But let's not discuss that here (at least not yet).  

These kinds of questions come up constantly when I'm doing web research.  Even a simple question such as "When and where was oil first discovered in California?" (SearchResearch 12/29/10) leads inexorably into a search for truth, or as much of it as we can find.  

In that search, you'll recall, the simple question unravels into niceties of definition--what do you mean by "first"?  What do you mean by "discovered"?  And while you can punt on many of these deeper issues if you're running a class in person, it's much harder to do it when you're trying to write a search challenge question.  They need clear answers and clearly resolvable web-findable answers. 

(I don't believe for a millisecond that all possible questions can be resolved by web-hosted content.  That's another topic for another day.)  

So, when you're writing a good search questions you need to not only pose an interesting problem, but also demonstrate that a search path clearly leads to a solution.  Let me unpack that a little... 

The best problems are interesting--I won't try to define it, but you know what I mean.  They're intrinsically interesting to a fairly broad range of people.  Abstruse and obscure issues of number theory generally don't appeal to more than a few people.  And "interesting" can be found in a number of ways.  I like to use "obscure connections" as a way to link topics together in ways you might not have thought about.  Example: In the AGAD problem of April 24th, 2011

Rembrandt painted a philosopher looking at the bust of a Greek poet. The gold medallion on the chain represents another famous Greek. Who is it?

That's a nice linkage between Rembrandt, Aristole (the philosopher) and Home (the poet).  You might have known one, but you probably didn't know the other.  That's nicely interesting, and a good example of something that's simple to do with search, but tough to know as trivia.  

Demonstrable means that there's a clear (and repeatable) sequence of search steps that gets you to the answer.  The best ones here also illustrate features of search skill that you might not know about.  We had one question (April 25, 2011) that required you to know about the Conversion feature in Google:  

If you came home from a trip with 150 South African rand, 350 Kuwaiti dinars and 200 Japanese yen, how much would you have in U.S. dollars?  

You can answer that one with the simple one-liner trick of [ 150 rand + 350 kuwaiti dinars + 200 yen in USD ] (or you could do it step-at-a-time, then add them all up with Google Calculator.  

So when we write these questions, we're looking for interest, a teachable skill and a clear way to demonstrate the skill in question.  

I've gotten just a few questions from SearchResearch readers.  Anyone else want to give it a try?  A limited-edition AGoogleADay t-shirt maybe lies in your future if your question is picked! 

Write on! 

No comments:

Post a Comment