Hacker News Reader

I notice that compilers are often singled out (unfairly imo) in these discussions. I found the language theory + compilers sequence to be one of the most practically useful parts of the entire major. in two semesters, I went from helpless googling to actually understanding compiler errors. I'll concede that the typical "toy compiler" projects are maybe not the best use of time for someone who wants to be a software engineer, but understanding PDAs and context-free languages is super helpful when you don't have the luxury of using a friendly compiler. most of the value is front-loaded in language theory, but a compilers course really drives the concepts home.

"Look man I just used that regex to parse the HTML response. Easy"

"You can't parse HTML like that, it's not regular"

"What do you mean regular? Look the regex works no need to complicate things we're trying to resolve this JIRA ticket"

See, no need for these fancy compilers!

And yet, these things are easily comprehensible facts without you needing to understand how to implement an LL1 parser.

More generally, I can always come up with something you can't understand without some specialised knowledge. I could argue that you couldn't possibly be a computer scientist / software engineer unless you understand the foundations of digital logic -- how else to understand safe implementation of crypto, caching and VM? Who cares about big O when people write pointer hungry code that faults to main memory every cycle?

I guess we might both be arguing for 'knowing what we don't know' and where to start looking.

You can parse HTML using regex. Not all HTML, but specific HTML pages are almost always regular as they typically have well formed nesting. It isn't a good idea but language theory doesn't say it is impossible.

I don't think I've ever heard anyone claim otherwise. Someone with a bachelor's and no experience is only qualified for entry-level jobs, where they will generally be working under close supervision.

The sentiment I get from some people is that traditional, pure CS educations are in some way uniquely qualifing, and software eng subjects (config mgmt, project mgmt, requirements/design/V&V) are trade skills and paperwork that don't have rigour, whereas in reality they are foundational skills for big/complex software projects. It is as reductionist as saying CS is just programming.

> that leaves you fundamentally unqualified for many tasks, and you'll have learn (via painful experience, apprenticeship [if that is emotionally possible for Gen Y/Z], or self study) many things

This is not a generational thing. I know at least as many entitled older folks as I do detached younger folks.

Moreover, part of the drive for saving dying industries is because the folks in those industries don't want to pivot because that's hard.

Some people just want life to be easier and give in to that. It is easy to sympathize with that because life is hard and it would be nice.

Someone that can take something confusing they've written and then refactor it into something easy to read and modify is better than most compsci skills. Because that person can usually suffer through hours of googling to find a math solution to a problem.

ok boomer

>painful experience, apprenticeship [if that is emotionally possible for Gen Y/Z]

Nice flamebait.

Sure. But I've met too many young programmers who absolutely reject any attempt at guidance, build towers of fragile code, and just seem unable to ask for advice. They have a strong tendency to want to be in directly in charge of something, regardless of experience or aptitude for that. It has been very noticeable the last 10 years.

> apprenticeship [if that is emotionally possible for Gen Y/Z]

ok boomer