Platforms, Top-down design and Tumblr
So, let’s start off with the cool, fun stuff: I love Tumblr. Aside from being slow to aggregate my data (their TTL might be a bit long?), it’s a neat way for me to have all of my “Web Twoooooo” activities aggregated into one place that looks neat and is easy to swallow. I love it, and am going to be giving it it’s own domain and a design at some point (until then, check out http://webba.tumblr.com/)
However, this brought up some discussion when Mike Luby was over at my place this past weekend working on some super-cool stuff with me. What happens if, say, Twitter just dies? I mean, besides how many people would get pissed off. There’d be literally dozens if not hundreds of services that would either be severely crippled (did you set CURLOPT_CONNECT_TIMEOUT right?) or just plain not work. It’s quite literally the house of cards. Which makes it important for me to get back up on my soapbox really quickly about 2 things:
1) Build your application as a platform! Don’t retrofit some crappy website 2 years later after people have been using your site as a service. Design it as a service, or use a framework that’s been built to handle being a service, from the beginning. It’s huge.
2) Top-down design. What’s that mean? It means presume and plan for the biggest, best, scaliest(?!?) outcome and work towards that end. Don’t build around pretty and slow for 2 users and wonder why when you become popular that you’re rewriting and retrofitting like crazy. Plan and design properly! It’s not hard to both make a platform usable/pretty as well as scalable/well-written.
So, cool apps are good, bad design is bad. Pretty simple, eh?
(via seeteeohh)
