Launching

Yesterday, I launched a partial rewrite of a long running side project connecting people to farm shares.  Over the past few months I’d fixed a few pressing bugs and overhauled the software so that a site and data model that previously supported only cities and zips as geographic features now supported states as well.

For the past week I’d been hemming and hawing about when to release–fixing “just one more thing” or tweaking one last bit.

But yesterday I finally bit the bullet and released.  Of course, there were a couple of issues that I hadn’t addressed that only showed up in production.  But after squashing those bugs, I completed a few more tasks: moving over SEO (as best as I can) and social accounts, making sure that key features and content hadn’t been thrown overboard, and checking the logs to make sure that users are finding what they need.

Why was I so reticent to release?  It’s a big step, shipping something, but in some ways it’s the only step that matters.  I was afraid that I’d forget something, screw things up, make a mistake.  And of course, I didn’t want to take a site that helped a couple thousand people a month find local food options that work for them and ruin it.  I wasn’t sure I’d have time to support the new site.  I wanted the new site to have as close to feature parity as possible.  I worked on the old site for five years, and a lot of features crept in (as well as external dependencies, corners of the site that google and users have found, but I had forgotten).

All good reasons to think before I leapt.

But you can only plan for so much.  At some point you just have to ship.


Twitversation: how much do you converse on Twitter?

twitter photo

Photo by eldh

You know what I said a few days ago?

I’d love to have stats on this to make myself more accountable, but I wasn’t able to find an easy way to show my Twitter usage (new tweets vs replys vs retweets)–does anyone know one?

Well, I didn’t find anything and thought it’d be fun to learn some of the Twitter API, a bit of Django, Bootstrap, and how to host something on Heroku.  So, I wrote an app, Twitversation, which gives you a rough approximation of how much you converse on Twitter, as opposed to broadcasting.  You can enter your Twitter username and it presents a breakdown graph and a numeric score (I’m 60 out of 100, whereas patio11 scores 78 and Gary V scores a hefty 83.

Twitversation only pulls the last 200 tweets, so it’s not canonical, but it should be enough to give you a flavor.  Sarah Allen has a post up about her score.

What’d I learn?  Among other things:

  • Heroku is super easy to get started on. And it’s free!  Perfect for your MVP.
  • Django has an unfortunate term for the C in the MVC (they call it a view).
  • You can create a pie graph using only CSS and HTML.
  • Side projects take longer than you think.
  • Picking a side project that doesn’t require any feeding is liberating.  Twitversation will keep running without any attention on my part, as opposed to my other side project.
  • Python’s dependency management is a bear for a newbie.  I didn’t have to do much with this project, because it had its own vagrant vm, but I saw some of the complexity out of the corner of my eye.  Makes me long for the JVM and classpaths, and I never thought I’d say that.
  • Catchy names are hard to come up with.

Hope you enjoy!



© Moore Consulting, 2003-2017 +