Of hacker schools and degrees

May 23, 2014

tl;dr: Hacker schools focus on practice. Degrees focus on theory. I don't care which you prefer, as long as you understand they're complementary. Those from one side of the fence can well learn from the other.

Recently I saw a tweet from Mike Bayer, the author of SQLAlchemy. He brought up an interesting thought comparing hacker schools, degrees, and woodwork. I respect him and his work a great deal but do feel the analogy doesn't ring true.

Whilst I'm not well enough versed in woodworking to make a direct comparison, I know a deal about the field of computing. The field of computing is an intricate and complex web. To gain a strong understanding of the many facets of computing is a difficult endeavour. A good degree program aims to introduce you to these many facets. It's still the way I'd suggest to go, even if it could use work.

The advice I have for those asking, and advice I have given to friends and students, is to get a degree. During the degree you should do internships at places that will extend you and give you the practical knowledge you need. You should also be doing your own projects, though hopefully you're the kind of person to whom that comes naturally.

Depending on where you are in the world, a degree can be expensive. A degree is a necessary tick in many boxes: employment, immigration, and so on. That's not the point I'm arguing though. If the expense is bearable, and the degree program's quality is good, it's worth it.

Given that, many of the coders I trust are self-taught, either since early days or due to moving over late from an unrelated field. It's not impossible to get a well rounded education through self-education, just more difficult.

Hacker schools are aiming to provide a shorter path to becoming a developer who is productive enough that they can be employed.

Dot point comparison for thinking

Issues with hacker schools

• Aim to find the practical and minimal solution to a problem
• The more questionable ones focus on "becoming a coder" in a matter of weeks not months or years
• Feeder schools directly into industry, commonly with the companies themselves sponsoring the places (which can be a positive or negative depending on what the students learn at the companies that employ them)
• Lack of emphasis on a computer science or a broader scientific background (no easy way to mix computing with other fields such as bioinformatics and so on)
• Lack expansive background to tackle certain problems the correct way (harder to avoid re-inventing the wheel poorly as you might not know that this has already been solved and been solved well)

Issues with degrees

• Expands on forefront of knowledge and technology not yet ready for industry
• Potential to mix an interest in technology with other fields (physics, biology, economics, genetics, ...)
• Focus on theory at the lack of practical concerns (see: proper version control, release practices, testing, code review, code style, ...) which is genuinely an issue and something universities need to fix
• The act of publishing papers is rarely focused on real world systems as what is useful is what is publishable
• The lack of a practical background to understand the gap between theory and practice leads to many poor practices

Augmenting hacker schools

• Hacker schools give you a strong initial foundation but must be rounded out with further education, either by yourself, your future employer, or something akin to a degree program
• Hacker schools prepare you to enter a workplace where you can find educational opportunities: if the workplace doesn't treat you well, you may find yourself stuck
• The education at hacker schools is commonly free or low cost as it's directly sponsored by industry

Augmenting degrees

• Round out your practical skills via practical internships with companies (practical as there are many internships which won't give you any sane or useful skills and/or the company avoids best practices like version control, testing, and so on)
• Hacker schools or one of the higher academic level variants such as data science / data engineering schools

Off-topic: my pet hatred of degree programs is the lack of practical skills. No program should skip teaching version control, testing, or code review practices. These skills are just as vital to have reproducible research as they are vital to have sound code in industry. That's a separate discussion to have however.