Every software developer should read this awesome takedown by Benjamin Pollack on the optimism of developers who say they can clone a successful product in a weekend.

tl;dr? Here are my favorite grafs:

That is why an open-source clone of StackOverflow will fail. Even if someone were to manage to implement most of StackOverflow “to spec,” there are some key areas that would trip them up. Badges, for example, if you’re targeting end-users, either need a GUI to configure rules, or smart developers to determine which badges are generic enough to go on all installs. What will actually happen is that the developers will bitch and moan about how you can’t implement a really comprehensive GUI for something like badges, and then bikeshed any proposals for standard badges so far into the ground that they’ll hit escape velocity coming out the other side. They’ll ultimately come up with the same solution that bug trackers like Roundup use for their workflow: the developers implement a generic mechanism by which anyone, truly anyone at all, who feels totally comfortable working with the system API in Python or PHP or whatever, can easily add their own customizations. And when PHP and Python are so easy to learn and so much more flexible than a GUI could ever be, why bother with anything else?

And the closer:

The next time you see an application you like, think very long and hard about all the user-oriented details that went into making it a pleasure to use, before decrying how you could trivially reimplement the entire damn thing in a weekend. Nine times out of ten, when you think an application was ridiculously easy to implement, you’re completely missing the user side of the story.

TRUE STORY: when I was but a wee lad working at my first startup, I had the pleasure of interview a hot shot who claimed during the interview that “what you guys have done is trivial” and that “he could write the whole thing in Perl over the weekend”. That asshat didn’t get hired.