Category Archives: research

Expertise is more than meets the eye

Last Wednesday, I shared this amazing anecdote from a research paper
on expert levels of performance (Ericsson, 1998) about the Italian
violinist, Paganini:

According to my father’s account, during one of his
concerts Paganini experienced a problem—one of the strings on his
violin broke; after a brief pause, he continued playing the music on
the remaining string. A little later another string broke but he still
resumed playing. Yet another strong broke, and Paganini finished
playing the concert on a single string while producing the most
beautiful music. Such a demonstration of an immediate unexpected
reorganization of one’s music performance is mind-boggling.

So what’s the explanation for this amazing feat of genius? Ericsson
went on to explain that in the 19th century, performers generally
composed their own music. Paganini set himself the creative challenge
of composing musical pieces that could be played on only one string,
developing new techniques along the way.

The audience didn’t know that, of course! Paganini started off by
playing the pieces on all strings. Easy enough if you’re used to
playing them on one string! Ever the showman, Paganini would sometimes
intentionally snap the other strings in the course of his performance,
finishing—to great applause—on the single string that he’d planned
all along.

So what seems like a miraculous gift is really more about lots and
lots of practice and preparation, with a little bit of trickery thrown

Now I have a sneaky suspicion that I’ve seen this trick before. Not
only that, but I completly fell for it too! I was in Tokyo watching a
shamisen performance. In the middle of a frenzied passage, the
plectrum a blur over the instrument’s three strings, snap! went one of
the strings. Not a problem! He adjusted the tuning peg and kept on
playing! I was *so* impressed. But now I’m onto you, Mr. Shamisen
Player. You probably break strings all the time while practicing. You
might have even been playing a one-string piece and had snapped your
strings intentionally to impress us. Hah!

Do we have anything like that in IT? What’s our one-string piece with
which we can astound other people?

Ericksson, K. A. (1998) “The scientific study of expert levels of
performnance: general implications for optimal learning and
creativity.” High Ability Studies, 9(1), pp75—100

Random Emacs symbol: comint-within-quotes – Function: Return t if the number of quotes between BEG and END is odd.

Follow up: JM Ibanez

and from Paul Lussier:

Practice can’t account for all of it, actual, real life experience
handling catastrophes and routing around unforseen problems on the fly
helps a lot :)

Paganini obviously knows his music and his instrument to such a level
that he’s able to pull this type of thing off. But anyone so versed
in music could do the same thing. Ever listen to extemporaneous “jam
sessions” from the likes of Eric Clapton, BB King, Jimi Hendrix, Joe
Satriani, Pat Methany? They’re all unbelievable guitarists who can
make the guitar do things you never even dreamed possible.

My manager, at 2:00am, faced with a dual drive failure in a mirrored
system running an ancient, monolithic, non-moduler Linux kernel
specially compiled with certain drivers long since lost, pulled off an
unbelievable feat of grafting a very recent, very moduler 2.6 kernel
from a Knoppix CD onto this system and was able to get it back up and
serving NFS long enough to move the entire terabyte array to a new

It’s the same thing. Years of practice and familiarity coupled with
knowing *what’s* possible, then working around catastrophe using this
knowledge and experience!

Someday you’re going to pull off the impossible too, but it won’t seem
so hard to you. It will just be “the natural thing to do” given the
circumstances at the time. To everyone else, it’ll be a miracle :)

Remember what Arthur C. Clark said:

“Any sufficiently advanced technology is indistinquishable from magic.”

Where “sufficiently advanced technology” can be defined as any area in
which you are more expert than the average person :)

Much done

Also done with practically all of the data gathering for my thesis. It
went surprisingly well! I don’t know what I was so afraid of. I have
maybe one or two more data points on Monday, but I already have enough
data to make my research supervisor comfortable.

I will celebrate with a nice long bath and a good weekend. That’s
it… Besides, I need to decompress before all of my job-related

Random Emacs symbol: ido-enter-insert-buffer – Command: Drop into insert buffer from insert file.


I showed my thesis prototype to three other researchers at the IBM
Toronto Center for Advanced Studies. They all thought it was a cool
idea! =) I’m going to do my pilot usability test tomorrow. I spent
today typing in my notes from various books… More tomorrow, too!

Random Emacs symbol: decrease-left-margin – Command: Make the left margin of the region smaller.

The tension between research and research

I’m starting to feel that my thesis is actually doable. I’ve applied
for the ethics review of my experimental protocol. The prototype needs
some tweaking based on the feedback from my IBM mentor, but it’s
coming along nicely. I’m learning to push back, too: “That’s out of
scope.” “Let’s keep it simple.” “I want to finish by June, mind you.”

I want to finish by June. And you know what, it’s starting to look
possible. I’m not as happy with my thesis as I hoped I would be,
because the practical applications for it are less obvious and less
immediate than I thought. Most people would already be happy with
using tools that exist if they only knew about those tools. I’m
sketching an idea a little further down the line, but people have to
learn how to make the most of the intermediate steps before we can get
to the point where what I’m building makes sense. That’s okay;
research is a beginning, not an end.

I just want to finish by June, leaving me some time to take care of

During our weekly update, my IBM mentor said something that intrigued
me. Stephen said, “Don’t forget, there’s a difference between doing
research and working on your thesis, and sometimes those two goals
come into conflict.”

I must have looked at him blankly. Research? Thesis? Wasn’t my thesis
supposed to be the only reason for my existence for the next few
months? Weren’t the two the same thing?

Stephen reminded me that my job in research is to wonder and be
curious—to explore. I should keep doing that. I told him about the
community work I’d been doing for LG (which is still hush hush),
which I’m really curious about.

I really appreciate having a mentor who points out all the various
forces and conflicts and will help me see them. It’s really quite
interesting how we talk about politics…

Random Emacs symbol: pgg-gpg-insert-key – Function: Insert public key at point.

Upon further reflection, I don’t hate my thesis after all

I am officially out of my I-hate-my-thesis phase. It was easy to give
in to frustration and think that hating it would make me finish it
faster. But why should I subject myself to that much unnecessary
stress? After all, if it’s going to be a big part of my life for the
next few months, I may as well focus on the positive.

And I *do* really like it. I love the feeling of working on something
new. I can’t wait to try these ideas out. I’m looking forward to
talking to IBMers and thinking of how we can make these ideas real.
And the people I’m working with are actually pretty cool, too.

Also, I’m finally able to download papers from the ACM Digital Library
again. Being able to *see* examples of good research certainly helps.
Knowing that we’re working to a higher standard than some of the
published papers also boosts my morale! ;)

I updated my schedule again, and was pleasantly surprised that despite
the Internet problems, despite my occasional grumbles and frustrated
arggghs, I’m actually ahead of schedule.

This is a nice feeling. I think I’ll keep it. Sometimes I really just
have to remind myself to focus on what’s going well…

Random Emacs symbol: gnus-auto-extend-newsgroup – Variable: *If non-nil, extend newsgroup forward and backward when requested.

Big, hairy, audacious goal

2My Big, Hairy, Audacious Goal for the next eight months is to help
huge companies imagine how they can help people connect by prototyping
my research idea, figuring out how it can be improved and made into a
product, and writing up my thesis.

A thesis is a pretty big goal. Just ask all the people who finish
everything but their thesis. Even really smart people. Even my

And I’m pushing even further than that. I want my thesis to be
practical *and* research-worthy. I want to take it into the business

I’m going to need help making this happen, and I’m going to need a lot
of personal strength, too.

Random Emacs symbol: browse-url-epiphany-sentinel – Function: Handle a change to the process communicating with Epiphany.