On my last trip through Logan airport my laptop ran afoul of a DHCP provider that handed it a duplicate address. Although I didn’t notice it right away, because I hopped onto a long flight and then spent the following day with a wired connection, the wireless capability of my Vista laptop had been turned off as a result of this incident.
When I realized what had happened, my first instinct was to start spelunking around in various administrative nooks and crannies of the system. It reminded me of another incident where Phil Windley went into overdrive to debug a display problem on his Mac. As Reed Hedges wrote in a comment:
…it’s as if you were the engineers building the MacOSX operating system and the only way to figure out problems is to rebuild it, and isolate causes.
Mindful of this, I asked myself: “What would a civilian do?” Then I took a deep breath and clicked on the Diagnose and Repair link. And lo, Vista said: “Wireless service is turned off. Would you like to turn it back on?”
Why yes, thanks, I would.
And lo, it worked.
Of course I couldn’t just carry on without knowing what had been turned off and on. For future reference, it was the WLAN AutoConfig service, which “enumerates WLAN adapters, manages WLAN connections and profiles” — something I’d forgotten about, or possibly never encountered, but that a civilian would hope never to meet.
It was a nice glimpse of how things ought to be.
That said, I’m wondering how the explanation that I found — using the ipconfig command, and the event log viewer — could be translated into something a civilian would find intelligible and useful. Recall that there was a long delay between the incident that triggered the wireless shutdown, and the appearance of the problem. I was able to figure out what had happened, but a civilian wouldn’t associate a transient popup message in an airport with a problem manifesting 18 hours later.
It’s wonderful to know that somebody can (in certain cases) wave a magic wand, say “Yes, please fix this,” and make it so. But absent a rational explanation for what went wrong, this can become a recipe for superstition. A civilian would be highly likely to attribute the whole incident to something that he or she did wrong, and then to fear doing that wrong thing again.
Of course we’d like things to always just work. But our systems are complex artifacts and things will go wrong. Intelligent diagnosis and repair is a daunting but worthwhile challenge, and it’s nice to see evidence of progress. Communicating helpfully to users about that process is an equally daunting but worthwhile challenge.