Current status

Working on a super-secret, stealth-mode startup.

Not really. Sort of, but not really. I’ve been working on a little application/service that I needed in my household. Its target audience consists of two people: My wife and me. I shipped just ahead of a June 1 deadline, and have been using the app multiple times a day this past week.

If it shows promise, I may look into expanding that customer base–but for now, so far, it’s doing what it needs to do to start getting feedback on the general idea. A true Minimum Viable Product, in that sense.

What I’m using

• Ruby on Rails: I suspect there are better frameworks and languages for this project. I was intrigued by Meteor, in part because of its built-in support for Phonegap. (I also looked into the up-and-coming Volt framework, but ran into just enough cognitive dissonance that it was going to slow down my tight shipping schedule.)

  And I don’t need a downloadable app yet anyway; mobile web is just fine. By accepting that, and using a framework I know well, I was able to knock out the first version in a couple of hours, and make a couple of changes based on feedback within another.

• Bootstrap 3: But, but, but! Every Bootstrap application looks the same! CSS frameworks are evil!

  I don’t care. I needed a decent-looking web UI, and Bootstrap still looks better and functions than Zurb Foundation and Materialize. And I’d rather focus on the general idea versus well-structured Sass right now, so Bourbon and friends are out.

• Heroku: Free hosting, simple deployment, and simple SSL support in about 15 minutes. I’m using a herokuapp.com domain, so SSL is free, and won’t have to worry about bumping up dynos or database storage for some time based on my current customer base.

  Waiting for a web dyno to spin up can be a drag sometimes, but it’s not a deal-breaker. Much nicer than the Capistrano-induced hell I’ve been dealing with the past couple of weeks on a paying gig–and if the lag does become an issue, Heroku’s new Hobby tier should keep me well-provisioned for awhile.

• Public domain clipart: I needed an app icon. I found one in ten minutes. I like it so much that, if and when I expand it, I’ll probably keep the same icon.

Costs so far

• Time: About six hours tracked; about two hours sketching out ideas before I added it as a project in Harvest (referral link).

• Money: $0.00. Ain’t software grand?

• Test coverage: This one admittedly hurts a little. I didn’t TDD this little app, but it’s still small enough that I can know pretty quickly when I break things. I do plan to add some smoke test coverage, though, as some of the first round of functionality gets validated and I move forward on part two.

The takeaway

Sometimes I get annoyed by the Learn to Code movement. It’s not just about acquiring skills necessary to compete in an evolving job market (though that’s important). It’s also about knowing what’s possible with computers and programming, and knowing that when the offerings in the App Store aren’t what you need, you have other options–even if it means opening up an editor or IDE and building it yourself.

I also wish more professional developers and self-labeled entrepreneurs would state out loud that, sometimes, you need to make a mess before you know what you really need to build. And that’s OK. If I spent all my time scheming the perfect architecture and mocking the most delightful UX and building landing pages, I wouldn’t ship for months.

I know it’s an apples and oranges comparison, and that not every product idea can start out at a scale of two. But maybe yours can. I honestly don’t know right now if this product will ever be a public offering. The first step is to make something I know two people find useful. I’ll scale up from there, if it looks promising, and revisit my decisions as needed.