As t → zero

Andy Hunt asked a while ago, What happens when t approaches 0?”

One of the interesting topics that came up today at Software Trends was “what happens when the time to develop a new application approaches zero?” It will never be zero, of course, but it will, over time, asymptotically approach zero. Suppose we finally get to the stage where a single developer, sitting at a laptop, can develop an application as fast as the end user/sponsor/customer can describe what they want.

I’ve been noticing a few web projects lately that are built at light speed. There’s nothing new about this as I’m sure lots of cool programs in history were written in a day or a weekend. Still, websites are the coolest programs in history because they’re immediately accessible by the whole planet, so it’s great to read about these projects happening. Thanks to the power of social networks, blogging, etc etc, it means a new service can come on and immediately its creator gets an idea about how big it will be and where to take it.

Here are some examples:

  • Startup Weekend brought 70-odd people together to build a new startup in two days. The blog from the event and the blog of the subsequent website, VoSnap is candid and the insights and admissions fascinating.
  • TwitterVision was built in four hours.”>built in four hours. This is, at present, my canonical example of lightning-fast development and the power of the mash. Not just spiking the concept, but actually a full delivery. (Twitter itself is rumoured to have been built in a week, which seems fair as it’s more the idea that makes it what it is.)
  • TransferBigFiles.com, which claims over a million big file downloads, was built by two guys in a weekend in ’05, two normal 8am-6pm days. Here’s how
  • RailsDay challenges developers to build something real in 24 hours. (The idea is sound, though execution lacked apparently, hence no repeat this year I guess.)
  • Yahoo! HackDay encouraged fast mashups and led to the GetUsOrganized startup.
  • AnAppADay The “software jedi” wrote 30 apps in 30 days. I’ve often contemplated taking a week off and doing 7 websites in 7 days. I’ve got enough ideas and can code them fast enough to get something out with Rails, the problem is deploying in a manner that’s stable, cheap, and secure. Looking forward to the day when Rails deployment becomes as simple and cheap as PHP … or maybe I should just bake cake.

2 thoughts on As t → zero

  1. What you tell differs completely from my experience. The most problem domains I encounter are not to master in days.

    The example above are exceptions – not the standard case for an web application.

    Think about payment, user profiles, internal messaging, complex filters for huge amount of data, multi language sites, obscure javascript issues etc. etc.

  2. Pingback: List Of Tweets

Leave a Reply