When I was a kid working in the bread factory, my nemesis was dough. It was sticky and hard to remove and it got everywhere. I got home with specks of dough in my hair. Every shift included a couple of hours of scraping dough off of machinery. I carried dough-scrapers in my back pocket. Sometimes a huge lump of dough would go flying someplace where it shouldn't and gum up everything. I had dough nightmares.

I worked in the production side of the factory. The other side did packing and shipping. Their nemesis was crumbs. Crumbs got everywhere. The shipping crew went home with crumbs in their hair. Every shift included a couple of hours of brushing crumbs out of machinery. They carried little brushes in their back pockets. I'm sure they had crumb nightmares, too.

Pretty much any job that you can get paid for includes dealing with one gnarly problem. If you don't have dough or crumbs to deal with, maybe you work in a razor blade factory and go home with little cuts all over your fingers. Maybe you work for VMWare and have nightmares about emulating bugs in sophisticated video cards that games rely on. Maybe you work on Windows, and your nightmare is that the simplest change can cause millions of old programs and hardware devices to stop working. That's the gnarly part of your job.

One of our gnarly problems is getting FogBugz to run on our customers' own servers. Jason Fried over at

But the great thing is that as you solve each additional gnarly problem, your business and market grows substantially. Good marketing, good design, good sales, good support, and solving lots of problems for customers all amplify each other. You start out with good design, then you add some good features and triple your customer base by solving lots of problems, and then you do some marketing and triple your customer base again because now lots of people learn about your solution to their pain, and then you hire sales people and triple your customer base yet again because now the people who know about your solution are reminded to actually buy it, and then you add more features to solve more problems for even more people, and eventually you actually have a chance to reach enough people with your software to make the world a better place.

P.S. I'm not claiming here that 37signals would sell 5 times as many copies if they offered Installable Basecamp. First of all, one of the reasons we may sell so many more installable versions of FogBugz is that it appears, to some customers, to be cheaper. (It's not cheaper in the long run because you have to pay for the server and administer it yourself, but that's subjective.) Also, our support costs for the installable version are only as low as they are because 80% of our customers opt to run on Windows Server. Because Windows systems are so similar, it's much easier for us to support the lowest common denominator. The vast majority of our tech support costs are caused by the diversity in Unix platforms out there--I'd guess that the 20% of our Unix sales result in 80% of our support incidents. If an installable version of Basecamp required Unix, the support cost would be disproportionately expensive compared to a hypothetical installable Windows version. Finally, another reason our experience might not translate to 37signals is that we've been selling installable software for seven years now; the hosted version has only been out for about six months. So we have a big installed base used to running FogBugz on their own servers. If you only look at new FogBugz customers, the ratio of installable to hosted goes down to 3 to 1.

Not loving your job? Visit the

This is part three of the text of a talk delivered to the Yale Computer Science department on November 28. I despaired of finding a company to work for where programmers were treated like talent and not like typists, and decided I would have to start my own. In those days, I was seeing lots of really dumb people with really dumb business plans making internet companies, and I thought, hey, if I can be, say, 10% less dumb than them, that should be easy, maybe I can make a company too, and in my company, we’d do things right for a change. We’d treat programmers with respe