Hacker News Reader

Comment by DarkWiiPlayer | Hacker News Reader

I completely agree with both points.

1) No, universities shouldn't waste too much time preparing people for the job; they should prepare give people the more scientific types of skills that one doesn't really pick up on the job: Data structures, algorithms, paradigms like OOP and FP, etc.

2) Unleashing a horde of yet incompetent newbies on the open source world with a mission to contribute won't be helping FOSS; what's more, if this was the norm, it would cause much more harm than good by spamming projects with low-quality contributions. See Goodharts law.

View on HN

but there's already significant amount of ds, algos, OOP in academia, also I don't see how those are "scientific types of skills"

I do also wonder what people *precisely* mean when they talk about "data structures and algorithms" in context of academia

You talk from perspective of competitive programming?

Yes, there is already a significant amount of it; and that's good. My opinion is that academia should not waste too much time preparing people for a job, but give them as much as possible of precisely the kind of problem solving skills that they won't be picking up on the job.

Playing devil's side:

It's dependent on the job, but anyway they can learn this stuff not only during work hours (learning hours or whatever it is called), but also in their free time

It's not like only Academia is authorized to teach DS&A, FP

I mean, yes, you can learn these things on your own, but that's by no means an excuse that higher education should focus less on these aspects.

It shouldn't be the default that university teaches you job skills and leaves you to learn the theory on your own.

Honestly, I really don't see a solid case there.

How can you learn to solve engineering problems without doing engineering?

University teaches you about theoretical computer science, not problem solving skills for software engineering.

No university teaches you about how to assess an engineering problem, how to break it down, how to create at timeline/estimate, etc.

> No university teaches you about how to assess an engineering problem, how to break it down, how to create at timeline/estimate, etc.

Many of them do actually. Obviously you will not be very good at it before having actual work experience, but it is useful to be open to thinking about it.

> How can you learn to solve engineering problems without doing engineering?

Yet that's the opposite of what the article focuses on: mostly soft-skills and all the fluff that surrounds the actual engineering work, like dealing with project maintainers, figuring out github, etc.

This approach, when taken to the extreme*, would lead people to be very good at explaining why they opened a certain pull request, but couldn't even merge it on the CLI because that's not something you need for a FOSS contribution.

* note this part

For DS and algos, formally proving properties of things would definitely be considerd an academic activity.

> mean when they talk about "data structures and algorithms" in context of academia

Have you taken a senior/graduate level DS&A class before? Part of it is about proving the correctness of the algorithms and making guarantees about the DSs. It's not about competitive programing or acing leetcode interviews at all.

In the class I took we didn't write hardly any actual code.

>Have you taken a senior/graduate level DS&A class before?

I've taken DS&A class during CS Eng. degree, it was bullshit about sorting algorithms while using 1-indexed arrays in languages that used 0-indexed approach, nothing fancy.

I wish it were about TLA+ and formal proving

Reply on news.ycombinator.com