Teach people to magnify their mind, not write code

Coding is not the new literacy:

When we say that coding is the new literacy, we’re arguing that wielding a pencil and paper is the old one. Coding, like writing, is a mechanical act. All we’ve done is upgrade the storage medium. Writing if statements and for loops is straightforward to teach people, but it doesn’t make them any more capable. Just like writing, we have to know how to solidify our thoughts and get them out of our head. In the case of programming though, if we manage to do that in a certain way, a computer can do more than just store them. It can compute with them.

That is, it’s not enough to write a loop in Ruby, a class in Java, or use a channel in Go. You’ve got to learn way more “material” than that: how to run your code in an application server, how to store rows in a database, how to deploy all your code to another machine. And then: how to have good taste, how to correct oversight, how to avoid bugs! And then, worst of all: knowing all the little miniutae like platform bugs, slow code paths, unstable code, dependency hell.

We shouldn’t put that upon people just because that’s how most programmers interact with computers. We should keep looking to help folks leverage systems as part of their work, not learn how to build systems to leverage systems to do their work.

Hence, the ending:

Alan Kay did a talk at OOPSLA in 1997 titled “The computer revolution hasn’t happened yet,” in which he argued that we haven’t realized the potential that computers can provide for us. Eighteen years later, I still agree with him – it hasn’t happened yet. And teaching people how to loop over a list won’t make it happen either. To realize the potential of computers, we have to focus on the fundamental skills that allow us to harness external computation. We have to create a new generation of tools that allow us to express our models without switching professions and a new generation of modelers who wield them.

We’ve succeeded in magnifying our voices with computers. I like forward to standing back and looking with wonder at how much we’ve magnified our minds.

When software loses its hair

Software’s Receding Hairline:

This is interesting, because the mechanism of growing a comb-over applies to software development. A comb-over is the accumulation of years of deciding that today is not the day to change things. A comb-over is the result of years of procrastination, years of decisions that seem right when you’re in a hurry to get ready for work but in retrospect one of those days should have included a trip to the barber and a bold decision to accept your baldness or take some other action as you saw fit.

A comb-over is a local maximum for improving baldness. You can’t really escape baldness of your head. You can escape a local maximum in your software, if you’re thoughtful about managing tradeoffs between product progress and technical progress.

It would have been better if you hadn’t gone bald in the first place though. Same with software!

(Analogies are just great. Thanks for this one, Reginald.)

Dining at the source code buffet

Let me start with a quote from wonderful person James Edward Gray II:

One of my favorite techniques for really learning a new language is to read the core API like a good novel. I’m a hit at parties!

I’ve had great success with this approach as well. I probably read the majority of Rails, the source of every RubyGem I used, and chunks of Ruby’s standard library in my first few years of working with Ruby. I picked up new tricks, figured out how things worked, and got myself out of a lot of tricky corners by reading code.

That we can do this is, to me, the real wonder of working with an open source stack. If I’m curious, I can dig into the framework, language, database, compiler, and operating system I’m using. If something goes weird, I can dig into it. I probably won’t end up changing or fixing anything below my app in the stack, but the ability to peel back the layers is a huge deal.

Given the choice of digging into why software sometimes goes weird and complaining or giving up, always chose digging into the source to see what’s going on.

When you work with open source software, you can always chose to figure out what’s going on around your app. Eating at the source code buffet is awesome!

Sam Stephenson, understated and excellent

I’ve enjoyed Sam Stephenson’s work for a long time. Even before “sheesh”, the most polite dismantling of an over-privileged open source user, Sam’s work has been top notch. Prototype is the library that made JavaScript palatable and learnable for me. pow and rbenv, in concert with ruby-build, are a lovely simplification of the weird problem of maintaining Ruby development environments.

The thing that pulls it all together, I think, is how well suited his solutions to diverse problems are. There aren’t a bunch of moving parts. Prototype was very much a library, and not a framework. His code is very much of the tool, playing well in the environment, be it Ruby, CoffeeScript, or even shell scripts. rbenv, ruby-build, and pow all play to the strength of bash and Node rather than trying to extend them to become something they’re not.

I was tempted to say his work is minimalistic. On second thought, I think it’s understated. Look at his website or photostream. The quality of just enough, but not too much, isn’t luck. It’s Sam Stephenson’s calling card. I love it.

How to succeed at Rails by trying

I think most teams, probably 90% of them, should start and stick with Rails conventions. Intelligently apply design principles, watch out for coupling that’s not worthwhile, carefully add dependencies when you must, sure. But don’t worry too much about erecting a wall between your app and Rails, building microservices, or whatever fashion dictates when you run rails new.

That said, I don’t think strict adherence to Rails’ opinions is the only way to succeed when using Ruby to build for the web. You can adopt the principles of Rails’ opinions, e.g. use code over configuration to fight boilerplate or reduce the number of choices developers need to make by curating some libraries. You could document those principles and invest in new teammates by mentoring them up on your framework and tools.

Actually, you should do that anyway! But there are reasons you may not be able to do that: the team is too junior, time is tight, you need to explore new technical ground in other areas of the project. If that sounds like your team, you will benefit a lot from letting Rails do much of the tool-building, principle-seeking, and training for you.

The wolf moves fast…

The Wolf:

The Wolf moves fast because he or she is able to avoid the encumbering necessities of a group of people building at scale. This avoidance of most things process related combined with exceptional engineering ability allows them to move at speed which makes them unusually productive. It’s this productivity that the rest of the team can… smell. It’s this scent of pure productivity that allows them to further skirt documentation, meetings, and annual reviews.

See also, The Grinder.

Well-tuned judgement

Lessons From A Lifetime Of Being A Programmer:

Never stop learning, the technology steamroller is right behind you waiting for you to stop.

I’ve taken this one seriously in the past, almost aways tinkering with languages, databases, frameworks, etc. I think it’s served me up to a point, expanding my mind and learning different ways to do to things.

The problem is I’ve reached the point of diminishing returns. I could go learn a stack-based language like Factor, or bend my brain around a oddly shaped database like Datomic. I’m not sure it would make me much better as a developer and leader of software teams.

Instead, the steamroller I think I need to keep ahead of is practice. Given a problem, what are three different solutions? What are their tradeoffs? Which approaches seem nice on paper, or in a blog post, but don’t work out a few hours down the road?

To wit:

This isn’t obvious to everyone, but the ability to see something new, or see what others are doing, or to compare multiple ways of doing something and then pick the best option for you, your team, your project or even your company is incredibly valuable. Most people I’ve seen are not very good at this. Most leaders are really terrible at this. It’s easy to just do what someone tells you you should do or something you read in a blog or just do what everyone else is doing. It’s much more difficult to look at things from all sides and your needs and pick something that seems to be best at that point. Of course you have to make some decision, people are often paralyzed by having to evaluate which often leads to picking something random or following the herd.

Well-tuned judgement is where I’m hoping to go next. Part of that is experience, knowing the forces and tradeoffs that apply to the possible solutions. Part of that is the ability to communicate it with teammates, sometimes face-to-face and sometimes asychronously. The really challenging part is letting your teammates run with the result of that judgement and collaboration.

A good developer makes good decisions for their own implementation; a great developer helps the whole team implement good decisions.

Conservation of complexity

You can’t fight the Law of conservation of complexity:

The law of conservation of complexity in human–computer interaction states that every application has an inherent amount of complexity that cannot be removed or hidden. Instead, it must be dealt with, either in product development or in user interaction.

Turns out one of my criticisms of microservices and microlibraries is a law. A LAW PEOPLE, YOUR ARGUMENT IS INVALID. Hilarious narcissism aside, keep an eye out for practices whose tradeoffs don't fit inside the depth of reasoning a blog post (like this one!) afford. Turning monoliths into services begets operational challenges. Microlibraries beget choices and wiring things up. Maybe the former is your thing, maybe it's the latter. Tradeoffs happen!

Executables deciphered

What's inside a compiled Hello, World program? Julia Evans is on that. How to read an executable:

Executable file formats are regular file formats that you can understand. I’ll explain some simple tools to start! We’ll working on Linux, with ELF binaries. (binaries are kind of the definition of platform-specific, so this is all platform-specific.)

I thought I had a rough grasp of how executables worked, and I still learned things. I love this format too. Julia Evans writes these fearless, curious posts about the deeply mysterious underpinnings of our computers and I learn a lot every time. More like this, please!

Microservices for grumpy old men and women

Microservices? I’m not entirely sure what they are. The term seems to exist on all parts of the hype cycle simultaneously. It’s on the ascent of excitement, the descent of disillusionment, and the plateau of productivity for different people, simultaneously. Some folks know exactly what it means, others know entirely nothing about what it means. Weirdness ensues. People talk past each other.

It’s a mess. Grumpy old man mode is in full force. (FWIW, I am the grumpy old man in this metaphor. I can not speak to the stature of Fowler or Feathers at this time.)

If I’m pessimistic, I nod along with Michael Feathers. He’s on to something when he observes the use of microservices as a blunt weapon against failures of encapsulation. Microservices become SOLID principles using units of deployment and even teams as a barrier between concerns. I feel that’s rather draconian.

If I’m optimistic, I cite Martin Fowler. To his credit, Fowler is doing the best work sifting through the noise and sensibly organizing what microservices might, in fact, be about. They’re probably not distributed objects, if you do it right. But you should understand the details, lest you get swallowed by the novelty and forget to realize the benefits. Make sure you’re tall enough for the ride your operations team is about to endure.

I get the feeling that radical approaches to working with Rails and microservices are tilting at the same windmill. You won’t fix the human tendency to build complicated structures with more software. Employing more buckets, or smaller buckets, in which to put your software doesn’t solve it either. You need a human factor to jump in and say “Hey, this is complicated! It might even be essentially complex. Let’s manage that complexity.” If you’re actively managing complexity, microservices or monoliths, Rails way or your own way, microservices and decoupling are a lot more of a detail than a foundational principle.