Jeff Devine

I only tell you what to do because I love you

Prototype vs. Architecture

The CEO of Meebo, Seth Sternberg, started a TechCrunch series focusing on the decisions a young entrepreneur needs to make. His first post, ”From Nothing To Something. How To Get There” recommends you focus on building a product and forget about VC:

“At the exact moment you had your idea, ten other people had the exact same idea. There was just something in the environment that made it the right time for folks to think that one up. The race has already begun! Who’s going to execute first? Who’s going to execute best? If you want to waste nine months trying to raise VC money for that idea, great. But six months in, you’re gonna cry when you see someone else put out that same product you’re pitching me right now. Like I said, forget everything else and just get your product out the door. Now.”

This weekend I also rediscovered Martin Kleppmann’s excellent blog Yes/No/Cancel. First it was building simple REST APIs in Scala but felt strangely validated reading The Python Paradox is now the Scala Paradox. He uses an argument from Paul Graham that:

“…a company can hire smarter programmers if it chooses to write its code in a “comparatively esoteric” programming language”

While Paul Graham was arguing about Python at the time, Kleppmann argues that Scala is this year’s black and has had great success using it to build parts of Go Test It:

“…but provided the technology is suitable and won’t increase your costs disproportionately, why not do something fashionable and adventurous? In an innovation-based technology business, the quality of your developers is key. Investments into things which make your good developers happy will pay off handsomely.”

I’m with Seth that every minute I’m not working on my ideas someone else is, but I struggle with what risks I add by using tech that is new to me. Is it worth adding a few months to get a solid architecture in place as opposed to throwing together a shell that’s mostly throwaway? When you have no product you can only consider the opportunity cost.

In the end none of it matters if I don’t have a working prototype, but this does continue to occupy me… especially as I wait on GWT to compile with Snow Leopard fixes.