Untitled topic

For nearly all of the software I rely on, I haven't examined the code. I'm not in a position to judge it.

For most of that software, of course, I'm not in a position to judge the developer either; but if all I know is that the developer isn't familiar with binary and hex, then I wouldn't expect her to be competent to write, for example, a brochure website, let alone a webserver. URL encoding depends on hex. Debugging often depends on hex. Arithmetic overflow and carry are binary. Twos-complement notation for signed integers is a binary convention.

I wouldn't hire a developer who couldn't explain binary notation. In fact I don't think I've ever met one like that.

As I suggested, perhaps I'm old-fashioned.

> But almost undoubtedly you'll be a better programmer if you do know CS (I'm using binary here as a proxy for CS since I can't imagine having a deep knowledge of CS without knowing something fundamental like binary). How many times over those 25 years could a problem have been more efficiently solved (in programmer time or CPU time) if you had known about problem solving techniques (perhaps those related to binary knowledge for example) that you don't know from the world of CS? You'll never know...

You're both right and wrong.

Principles are a lot of cognitive debt to take on, and not strictly necessary to be functional. Literally anybody can sit down with an IDE and write code to merge spreadsheets or whatever the actual business need is. We teach this stuff to anybody with any interest. Kids, even!

If someone thinks they want to be a plumber, they shouldn't start with a 4-year commitment to learning principles of hydraulics and mechanical engineering-- makes much more sense to apprentice, lay some pipe, and if it's the job for you, then go back and learn it at a deeper level. Doing it the other way is why college is so ridiculously expensive and most spend 5 to 6 figures on a major and then go manage at Target.

After failing discrete math three times and giving up, I managed to make it almost 20 years in this industry before needing to learn anything about binary math-- and even then, it was only so I could understand how flags in old MUD code worked, for fun. Truth tables are important, but there has never come a logic problem I couldn't solve by taking a few extra steps to be explicit where you might reduce the logic to a compact blob of symbols. I'll never optimize code as well as someone classically-taught. I don't know what Big-O is-- and outside of a botched Google interview absolutely not a single employer or client has ever given a shit. Nobody has ever been victimized by my code. The "CS" way implies the academic approach is the only way to do things. It's the textbook definition of gatekeeping. All academia did was appropriate and institutionalize experiences yahoos like myself have always learned through ingenuity and trial-and-error.

You've learned more than I do, so you have more tools in your bag. The only thing that sets us apart is that I won't be advancing the industry or publishing anything on arxiv anytime soon. Outside of that, you're going to struggle with quantifying how you're better than the self-taught without resorting to speculation or guild/union mentality (gatekeeping).

The average web CRUD developer never needs to touch any of this stuff.

- Numeric types? Who cares? I know min, I know max. I take number from user and insert it in database. For calculations with money, I use integer cents.

- Bit-fields? I work in Java, what are bitfields?

- IP addresses? I am web dev loper, not network engineer. I don't need to deal with netmasks.

- Cryptography? Me no understand. Me use Let's Encrypt. Is secure, no?

- Compression? Browser do gzip for me. Me no care.

- Colors? I pick the nice color from the color wheel.

- Binary? What is binary? I only use binary when I want users to upload a file. Then I put the files on S3.

Well I do embedded dev as a hobby now so I know this stuff. But for a long time I didn't know how many bits were in a byte simply because I never really needed that knowledge.

I think the generalization skill that allows to go from one to another is essential. If someone has only some basic understanding of boolean expressions, but cannot apply this knowledge to binary numbers, this is a good characterization of their programming skills in general.

> Nobody has ever been victimized by my code.

You don't know that. Actually all code that is inefficient is victimizing both the user (via performance) and the environment (unnecessary energy usage). I'm not saying you've done anything wrong, I'm just saying we all don't know what we don't know. I'm sure my inefficient code has had many users and electric bills as victims. I released software before that subsequent versions where I had better CS techniques improved 100 fold in performance. I wasted my users' time before I learned how to do it better.

> You've learned more than I do, so you have more tools in your bag. The only thing that sets us apart is that I won't be advancing the industry or publishing anything on arxiv anytime soon. Outside of that, you're going to struggle with quantifying how you're better than the self-taught without resorting to speculation or guild/union mentality (gatekeeping).

You made a lot of assumptions about me without knowing my background. I was a self-taught programmer as a child/teenager and my undergrad degree was in economics, not CS. I went back to school to get a masters in CS which was difficult coming from a self-taught background. I did programming both paid and hobbyist for over a decade before that more formal education. And I write books read largely by self-taught programmers.

Saying a full software development education includes binary and the fundamentals of CS is not gatekeeping, it's a low bar if we don't want inefficient software wasting users time and sucking energy. I'm not saying you have to start there, I'm saying it should be part of your education as a programmer whether self-taught or formally taught.