I just made a new portfolio site with cactus, a python static site generator. Oh man — the joy! So if you’d like to see all the projects I’ve worked on over the past few years in one handy place, check it out (and read on below the picture for specifics and lessons learned):
Some of my projects
Until now, I’ve been using django (basically like Rails) or wordpress (e.g. this blog) for everything I build. I have all manner of shortcuts for both of them, but I it’s still slow to develop compared to other options.
Crazy part? No matter how many blog posts I read (lots), it didn’t click until I shipped something with alternate tools.
Even crazier? It didn’t have to be big. This was a one day project[1].
Now my head is spinning: what else have I been missing? How many days have I squandered trying to optimize strategy and customer development when I could have cut the project in half with better tech choices!?
This may be boring & trivial, but for the sake of anyone who is starting out, the complete list of tools used for this project was:
- Cactus (site generation & deployment to S3)
- Amazon S3 (hosting)
- Amazon Route 53 (a bit of plumbing required to use a real URL)
- Bootstrap (basic style)
- Normalize.css (more basic style)
- Sublime Text 2 (actually vim, but don’t use that) for writing the code
My next project will probably be stick with cactus while adding Parse, which I’m ashamed to say I still haven’t used.
Dex.io needs a “real” backend, so I’m happy leaning primarily on django for that. But I’m going to look elsewhere for my side projects. And I suspect that the lessons learned with make their way back into the main startup and save me a substantial number of days and weeks.
[1] 4 hours to learn the tech & build the skeleton, 4 hours to tinker with copy & layout, and 1 hour to proof & deploy. Those chunks were admittedly spread over 3 days, but you know what I mean.
Related Posts
Lean startup intro (with no buzzwords) Next Post:
Deciding which features to build for a big demo
"Wow. This book created a paradigm shift in my understanding of how to ask good questions."
I’ve been using Jekyll with some success. It’s based on Ruby but you barely notice that while using it. The nice thing about it is that GitHub supports it directly. You can host simple projects there. More dynamic stuff can be made using JavaScript. Here’s one example of this.
As you said using a static site generator isn’t enough when you need a backend. Heroku is a good alternative in that case.
The nice thing about static site generators is that you avoid using something heavy like WordPress and have the control totally on your hands (good/bad thing depending on a person). As a bonus it’s dead easy to host and it is more secure by default as there are less access points available for a possible attacker.
OMFG! There’s a new kid in prototyping town. Low-fi, hi-fi, and now function-fi. Instead of wireframes, or HTML mockups, you can actually make something functional in a day and get it in front of customers…
Since I have a decade-long PHP habit, my quest to carve out time for new startups has been heavy on cust dev too. My last attempt to bone up on modern coding was to learn node.js, but I struggled by getting snagged by a lot of undocumented conventions, and those projects went on ice.
I’m gonna take a few hours today to dust off some of those front-ends (already using Bootstrap) and plug Cactus and Parse into them instead of the node.js back end. Lets see if that opens the gates…