The perils of programmer education in the bazaar

Credentials vs portfolios isn't an either or argument. And "worse" is a loaded term.

If you have the credential of an undergrad degree at MIT with a good GPA, combined with a credential of working for Google for 5 years, there is a very high likelihood you are a good engineer.

On the other hand if you have made serious contributions to the Linux kernel, there is a very high likelihood you are a good engineer.

But most developers don't go to MIT and work at Google. And they also don't make serious contributions to large open source projects. For most developers credentials and portfolios are not effective methods of hiring.

Maybe! Although (for better or for worse), it's a word you rarely hear in the software industry.

I actually may plan to do some feedback with cancellations in the future, but "you not publishing for half a year and giving everything away for free" seems like a pretty compelling reason to quit a service.

However, the main feedback I got from people (and it happened often enough where I'm convinced it was a common line of thinking) was that folks just weren't involved in Ruby anymore, or weren't doing a ton of online reading because got busy with other things. When the average age of our remaining supporter accounts is something like 3-5 years), that's understandable.

I've had two children since Practicing Ruby was started. Things were hard for a while. I wanted to quit the business many times, but left it in a zombie state in the hopes to do right by my subscribers sooner or later.

Now things are getting a little better for me, so that's what I'm going to try and do.

As for eBooks, I've thought about that 100 times. I would love to get some of the existing content into a nicely tied together collection and release it as an eBook. But it's a big enough undertaking where I feel it'd take me away from other more valuable work I could do.

If we get our cashflow situation to be even a little better where I could fund a couple days a month to work on that, I'll go ahead and do it. But for now, I only really can afford to fund about 1.5 days a week of my own work, so that isn't much to spread around.

Best option is actually to get a job where you'll be working with other senior programmers and then understand your company's source code. That's how I did it - my first job had Ken Arnold as the CTO, then when I got to Google I'd lurk on Jeff Dean or Rob Pike's code reviews (everything is public internally at Google, you can see what everyone's working on). The big advantage of this is that you understand the business context behind the code; it's really hard to understand code in isolation without understanding the purpose behind the code and the goals for which it was written. A secondary advantage is that you've got ready-made mentors.

If you can't get a job and need to go the open-source route, then Rust, Chromium, V8, Django, and the Linux Kernel are all good options. Again, the guiding principle is to pick software where you're familiar with the outside, so you have context when you go look at the inside. Reading software in isolation is hard, so you're often best off volunteering to fix a couple low-priority bugs on the bug queue, and then diving in to start with them. Existing volunteers will often point you to the right spot in the code to get started if you're doing productive work for them.

I feel like I addressed this very directly in the article, but I'll repost the main points of "what's wrong" here for further discussion...

This economic model is so ingrained in our DNA, that we hardly ever question it, and many would go as far as to hold it up as a triumph. But the truth is, even if this system works at the grand scale, it’s hurting both learners and teachers in countless subtle and not-so-subtle ways. Here are just a few:

1) The system we’ve created is assumed to be a meritocracy, but is actually driven far more by popularity and market trends than it is by the quality, relevance, depth, and effectiveness of our learning materials. Although it is far from a zero-sum game, the current model is much more competitive than it is cooperative, and so suffers the consequences that come along with any competitive environment.

2) Because being highly skilled at both writing and software development is rare, those who can do it well are often get the most attention and influence from the software world, turning these folks into “thought leaders” that drive the overall direction of the community.

3) Producing high-quality educational resources is obscenely, ridiculously difficult and time consuming. So the people who can invest the effort are typically either from a position of economic advantage, or are backed by monied interests.

4) As a result, the materials that get produced, and the topics that get covered widely are based on where the money and social buzz is at right now. Those who are doing original research, particularly things that are experimental or exploratory in nature, are not well supported at all.

5) Because we’ve tied blogging and teaching and tweeting to hireability, plenty of people enter into the bazaar not because they particularly want to, but because they must do so in order to “prove their worth” to potential employers or clients. Again, this clearly benefits those who have lots of time and resources to burn, and burdens everyone else with massive amounts of shadow work. We cannot be surprised at the lack of socioeconomic and cultural diversity in light of this point.

6) The success of the few enables the selling of dreams to the many, creating a cottage industry of infomarketers. In effect, these are the folks who sell pickaxes and maps to gold prospectors, and it’s no surprise that their business is booming. Some of these folks are brilliant and insightful, but many are hucksters.

In light of all this, we cannot say “this is just how the internet works”, as true as that statement is now. We have to ask, how do we make things better for our future selves and those who will come after us down this road?