by Mark Bernstein

My day as a software engineer always starts with tech support. Morning e-mail brings news from writers in Europe who have run into problems, from West Coast students who are trying to puzzle out their reading assignments, from scholars who ran into trouble with their research notes long after midnight. Nearly all of these will be new questions about new problems. This is especially true today, because this week we're finishing a release of Tinderbox, our tool for visualizing, analyzing, and sharing notes, and the beta testers are poking at rough edges.

Today brought a bad file from Indianapolis, documentation clarifications from Portsmouth (accompanied by notes on preparing a  lobster and scallop risotto to be served with vinho verde), concept mapping queries from Oxford (with plans for a historical novel), and a proposal for a new feature that we'd never considered, that has never appeared in the research literature, and that appears to have occurred to this user on his first day. The post brought a large enigmatic poster, accompanied only by a hand-typed Rolodex card, which we presume is the first step in an elaborate author query, and a book jobber's order for a single copy of a hypertext novel.

As an industry, we have mixed feelings about tech support. We all pledge allegiance to user-centered design, to customer-driven development, to getting products out quickly so buyers can tell you how to improve them. But we also go to great lengths to shield developers and designers from tech support, lest their time be swallowed in a general clamor. Phone trees and callbacks and screeners are so effective that most of our new customers simply assume that tech support is a hollow promise.

Users are great at finding problems. Users are plentiful; they try surprising things; they sometimes have whimsical ideas about how the software works. Listening closely to tech support gives you a good sense of what needs to be polished and how your software is actually used. It also reminds you how much the software matters to people's work and how hard people's work actually is.  

Mark Bernstein is chief scientist at Eastgate Systems, where he crafts software for new ways of reading and writing.