A marvelous post about Patrick's start of his "wee business" and how he got there on the time left over from his salaryman day job with a three-hour train commute to boot. Whether this overall will work for any of the rest of us, given what seems like incredible focus Patrick seems to have, is unknown.

And this is not "work four hours per week to gain huge success in life". Like everything else he writes, this is backed by expedience, external constraints, and data-backed experience.

Patrick's business is what I call the Chicago School of Startups. (I might have just made that up.) (And at the risk of tweaking Joel, as Fog Creek Software seems to be following the Chicago model.) This is, like 37Signals, a self-funded operation, and eventually self-sustaining. The SV School, or more precisely, the YC school, emphasizes the path of seeking venture capital, and has the side benefit of a seriously enviable network of other founders and former founders.

What has been most interesting is the data-driven nature of his decisions. Many longwinded discussions on HN have focused on the importance of design,. Patrick, among others, points to A/B tests (not to mention highly successful sites like Craig's List) indicating that such design obsession might be misplaced. Minor heresy: while Mac developers are very graphically intensive people who will buy software just to lick it if the UI is good enough, many Mac users are just regular people. My Mac version has a conversion rate fully twice that of the Window version, and it is not noticeably pretty.

Perhaps there is a little leeway to tease Patrick--it does not seem that this five hours per week is sustainable. It is successful enough that he is quitting his day job, and I bet he now can't help but spend more than five hours per week on it. Congratulations.