Wednesday, November 14, 2007   —   Vienna 0 Comments

I'm in the process of reshaping and streamlining the backend of Wimbomedia, and wishing I had paid more mind to the cautionary voice in my head urging me to plan more and code less when I was building the site.

As the Wimbomedia grows, I've identified some potential issues with scaling the information-heavy database. It's proving a little cumbersome changing the way the system works because it's already so loaded with data.

It's amazing what you can learn between launching a site and it being pickup by a Google bot. Well, you live, you code, and you learn stuff you wish you knew before...

One lesson I'm learning pretty well is that there is NO excuse not to thoroughly figure out the framework of any application before your first line of code. Sounds pretty obvious doesn't it? Well... it's demanding a new level of discipline from me because so many of my projects are initially propelled by the urge to see an idea in work.

In many ways this is the major advantage and disadvantage of building online applications. No idea is ever born whole, so the advantage of designing software for the web allows you to quickly see the potentials of your idea "on-the-fly". But the key here is to be restrained; to work on the details early as you build fast. Wimbomedia began as a hack that seemed to work very well. And as each feature became more refined and stable, I just continued following the bread crumbs of cool possibilities...

I'm learning that there is a way to build fast and dirty without neglecting the greater framework. The simple principle is to make sure that the sum of the parts you’re hacking out should never reduce the efficiency of the whole... ever! Otherwise it will never build to scale, regardless of how great the code is, because scalability is as much about efficient organization as it is about efficient delivery.