• Problem solving vs error correction

    Take the circumstance in a city where tourists passing from a transit stop to a popular attraction.  The most direct route passes along a normally quite industrial area, especially on evenings and weekends.  And our tourists are regularly being mugged.  What is the problem here, and how do we fix it?

    1. Do we increase police presence in the neighborhood? Maybe add an undercover task force?
    2. Add better street lighting and security cameras?
    3. Or put up directional signage to the attractions that leads the tourists around the offending area along a more heavily trafficked thoroughfare?

    Is the problem that the neighborhood is dangerous?  Or that people who don’t know the city are making themselves easy targets by passing through a deserted area?  Choices 1 and 2 strike me as typical responses, but option 3 might just be the best.  It is cheap, quick, and low maintenance.  It doesn’t really solve the problem of criminals in the neighborhood, but it will stop people from getting mugged.   Now I recognize that it could be argued that the muggers would follow the new pattern (although they’d be more likely to be seen and caught) or that tourists wouldn’t follow the signage (but most would, not being familiar with the area) — but will you at least agree that option 3 would be the easiest to try?

    This all came to mind as I finally solved a long running application problem.  In our financial system one department kept getting an error when they went to use a certain screen.  The screen was running an update procedure — the same procedure used on another screen which did not give this department any trouble.  All the other departments could use the screen with no trouble.  So at first I started digging through this one department’s budget lines and transactions looking to see what caused their error where no one else experienced one — and they have hundreds of  budget lines with thousands of transactions.  A real needle-in-a-haystack situation.  I have given up on the problem in disgust, and then returned to it several times over a period of months — no solution.  But today, in a flash of inspiration, realized that I approach the problem from a different angle.  If I move the update routine to another screen one step back in the transaction process I might just end-run this specific situation, and still keep the finances in order.  AND so far it looks like this idea works.  I didn’t really fix the error, but I did (hopefully) solve the problem.  And isn’t that what the end-user really cares about?

    Related Posts

    Tags: ,
  • Author: Randy

    In my day job I serve as Information Technology Director for the Yale School of Drama. Otherwise I garden, play guitar, build stuff out of wood, take photos, play around with technology and have been blogging since 2003.

    Share on: LinkedIn

    Stay Informed!

    Did you enjoy this post? Then subscribe to my email newsletter and have the daily posts delivered directly to your inbox. Enter your email address here:

    ADD YOUR COMMENT
    Comments are moderated.

Welcome to RodeWorks

Randall Rode's online home for thoughts, notes, and experiments with a wide range of technology topics. Visit the about page for info on my recent projects and professional background. I welcome your comments!

  • Recent Comments

  • Coming Soon

    Open Source Way in creating community

    March 17, 2010

    Drupal 7 promises big UX improvements

    March 17, 2010

    Site Topics