NOAH KANTROWITZ

Making Things 2013-09-03

Braincrack: (n) an idea you would rather plan about than execute.</blockquote>

First used by Ze Frank, braincrack is one of those ideas that gets lodged in your brain. That you just never quite get around to creating or executing, but always tell yourself you’ll get back to. Maybe it is your eternally unfinished novel, or maybe a funny site about cats in sombreros, or maybe it is a complex web application. In the end the problem is often the same, making things is hard. Let’s look at this process both generally and for my hypothetical cat photosharing site.

What

At heart everything starts as an idea somewhere, an idea we are committed to not allow to become the fabled braincrack. Often this is just a given, you have an idea you want to share with the world but don’t know how to proceed. For those that want to practice but don’t have a specific plan, we have special gatherings that celebrate starting an idea from scratch. Weekend hackathons have become increasingly common as a marketing vehicle, and online game jams like PyWeek and Ludum Dare encourage coming in with nothing but a passion for making things.

In the end, though, ideas are the easy part. Executing every passing concept would take many lifetimes, filtering down to just the most worthy of time and attention takes experience that paradoxically only comes from choosing poorly many times. Make wrong turns, and try to see when to put down your tools and work on something else for a while.

For my photosharing site I was fortunate to have a serendipitous realization that the world needs a dedicated place to share photos of cats in sombreros. The outline from there flows naturally. I can look at existing photosharing sites for inspiration on the interface, as well as places like PatternTap.

Who

Even the best of concepts generally needs to be grown and honed before you can jump into building it. Here is where having a social network to tap can truly shine. Posting your idea to Reddit or Hacker News can get of eyeballs on it (ex. hyperloop with over 45000 votes), however it can be both demoralizing and time consuming to tease apart the useful input from the overwhelming wall of noise.

Organizations like PyLadies have worked hard to create safe spaces to explore new ideas, though these can be few and far between. Community hackerspaces are also a good place to seek safe input. I am happy to have seen rapid growth and proliferation in these kinds of communities, and look forward to them being more available to everyone. To slightly misquote John Green, we need to keep “making places where people can fall and get caught”.

Generally when talking about functions within a team (in the context of something this early in its life cycle) you would hear about three main roles: designer, developer, and sysadmin. These three silos have far more in common than they differ, but they do often remain unfortunately far apart. Make sure to balance your own skills with those of others, and cultivate relationships with as many other domains and domain experts as you can, your second drafts will thank you.

Back to my brilliant idea. A quick IM with some friends revealed that catsinsombreros.com was available and is a natural fit. We quickly brainstormed some ideas about how to run the service and what kind of community it would be. We quickly iterated on a logo to share in the hopes of rallying others to our cause. All of these things rely on having a community to work with, that you trust to provide good feedback and be honest with you.

Where

If your idea is entirely or largely informational, something you just want to share with the world, there is now some good news. Between WordPress and Tumblr, it is easier than ever to launch a site with some quick content. If these don’t address your needs, unfortunately the path ahead drops out from under you very quickly. Many Platform-as-a-Service (PaaS, ex. Heroku, Google App Engine) services offer robust free tiers which can be an enormous boon to a nascent project. As before though, these services have a very proscribed way of doing things and deviation will lead to an even sharper drop than last time. After that the next place to look is generally a public cloud vendor, some of which offer limited free resources but in general you will have to be willing to put up some money in one way or another.

Beyond just hosting options there are a lot more moving pieces in just getting something up where people can see it. Domain registration, DNS configuration, mail server records, all of these can present serious roadblocks as you need to integrate many disparate moving pieces.

The emergence of well-supported and clearly defined free tiers has been a colossal leap forward in allowing people to play with ideas. If you are a service provider of any kind, I would strongly encourage you to look in to offering one.

For catsinsombreros.com my first stop is my domain registrar, Hover. My registrar offers free DNS hosting, as do many others, so I don’t have to find a separate host for that though I do have to configure it myself. For an application this simple, and given that I expect my launch to be small with the current surplus of photosharing sites I can easily start with a free tier application on Heroku and either scale up there when my service takes off, or migrate to my own servers.

How

HTML, CSS, Javascript, Objective-C, Java, Django, Rails, Express, Chef, Puppet, Ansible, Fabric, Nagios, Graphite, PagerDuty, Git, Logstash, NPM, Rubygems, Bunder, Pip, Setuptools, PyPy, Nginx, HAProxy, Apache, Linux, Docker, Illumos, FPM, Omnibus, ZooKeeper, Subversion, Mercurial, Perl, C, Python, Ruby, PHP, Go, Scala, Clojure, Postgres, MySQL, Redis, Riak, MongoDB, EC2, Rackspace, OpenStack, HTTP, TLS, bcrypt, SHA2, HMAC, OAuth, REST, Jenkins, Travis, Sublime Text, Vim, Emacs, TextMate, Twitter, Facebook, Google+, Reddit, Hacker News, Marketing, Sales, Support.

We have built an astonishing number of tools, frameworks, and communities around making things with computers. This is not an indictment of any of them, but the fact remains that mastering even a small subset can take decades, by which point your skills will surely be outdated. Most people, myself included, advocate one particular stack of tools or another, sometimes to the point of dogma and fanaticism. Ignore us, all of us. Do it wrong, ruin your first deploy because you forgot to rebuild your CSS, get woken up at 3AM because you didn’t setup automatic fail-over on your database, but get something up there.

Will you regret not doing it the “right way” during that 3AM outage when your customers or friends are annoyed at you? Absolutely, but you will regret doing nothing far more, and the world will be lesser for it. Don’t forget all those pain points either, share them, make sure we all keep talking about them so when it is the next maker’s turn, they have an easier time of it.

For my site I can pick some simple components to build up from. Heroku provides a hosted Postgres database, so that is a natural fit for data storage. For storing the actual pictures I can use Amazon S3, and for the web application I can use Django. There is a lot not covered in this, things like SSL or monitoring or off-site backups can realistically wait until after the main site is up and running, but be sure you don’t leave it for too long or you are in for one of those nasty 3AM phone calls.

The Exception

In a complete 180 from the above, there is one place where “don’t worry, just fail” can’t usually apply: security. While you don’t have to immediately deploy TLS Perfect Forward Secrecy or DNSSEC, you do need a certain baseline to ensure you don’t put data from users at risk. One option is to just not have any data from users, thus sidestepping any hazards. Another is to make use of the password and user data storage options in a popular framework, which often gives you the current best practices for things like key derivation functions. Unfortunately getting experienced help with internet security concerns can be very difficult, even I only have a few resources so my only suggestion is that if you are at all unsure if you are about to do something unsafe please ask me. I don’t claim to be an expert, but hopefully I can point you at one if needed.

When

Now. Let go of the braincrack. It won’t be perfect the first time around, or probably even all that good, but it will be real. Only there can you start the journey towards awesomeness.

Why

I am a maker, and I think you should be too. I can’t help you write that novel or finish that painting, but I can help you put all these and more on the internet. As time goes on, we are moving closer and closer to ideas being the true currency of the world, but an idea must be made manifest before you can show it to others. One of the original goals of the OLPC was to create a whole generation of makers, to unleash upon the world children that had no preconceptions about it being difficult to make and share things across the room or around the world. Much progress has been made, and we build on the shoulders of giants.


As I lead off with, making things is hard. Specifically of interest to me is that launching a website is at least an order of magnitude harder than it needs to be. I am going to fix this. If you want to fix this too, maybe you should hire me.


Back to articles