Hacker News Reader

Comment by qntty | Hacker News Reader

Surprised not to read anything about offering classes to people who choose to live in these remote-work communities. That's the only reason it would appeal to me personally. I think a lot of people would like to take some graduate classes while working, maybe even get a degree out of it.

Our education system for things like computer science seem to be out of sync with the way that people actually learn. Not because they are too theoretical, but because there are some things, like computer science, where it's really impossible to internalize the practical lessons of learning about different kinds of software systems if you don't have some experience building software systems yourself. And this kind of experience takes years of working to develop.

View on HN

Computer Science is not Software Engineering. That’s my conclusion after the degree and two decades of professional work.

absolutely, my alma mater even has separate tracks for them (also for at least two decades, even if I only graduated a a bit over a decade ago)

Famously, computer science isn't a science, nor is it about computers.

Is math a science? My CS education was largely math classes. Algorithms are definitely math.

I would say no. STEM is science, technology, engineering, math. compsci is a math degree.

Maybe math is a tool/framework used by scientists? Physics is another example that is largely math.

Math is both a philosophical enquiry (pure math) and a tool (applied math & statistics).

Physics as a science uses math as a modelling and analysis tool.

I've always parsed mathematics more as a language than a science, because it's not necessarily empirical.

Perhaps as a language combined with rules to operate upon that language, transforming true statements into other true statements.

Physics : Engineering :: Computer Science : Software Engineering

Pursuing this question is quite the rabbit hole, but doing research in mathematics makes me believe that math is a science: you have questions, you collect evidence by exploring examples (i.e., conducting experiments), and you formulate hypotheses as conjectures. If you're lucky, your conjectures are correct and you prove them! Proofs are even better than the accumulated evidence on which theories in other sciences depend, in that proven theorems are immune to falsifiability within the axiomatic system in use.

I guess not a physical science (although there was some crossover last year between some results in quantum computing and physics, there's probably lots other examples that I'm not familiar with too) [1] but it fits squarely in the formal sciences (which includes math, logic, etc). And it doesn't concern any particular physical computer, but it does have topics on what an idealized/ theoretical computer could do haha

[1] https://www.quantamagazine.org/landmark-computer-science-pro...

computer science is applied math, and software engineering is applied philosophy

Following that reasoning, wouldn't any type of engineering be applied philosophy?

Hence doctors of philosophy.

If what you’re doing isn’t research it isn’t technically philosophy. The more applied a field is, the less philosophical it is, because a general “way” of doing things has been established. That’s why you can have an entire undergraduate about “solved problems”, and only after that do you start advancing the field by tackling unsolved problems.

What would be a better term for it? Maybe "the mathematics of computation"?

Some UK institutions (including some pretty respected ones) adopted the term "School of Informatics" and confer Informatics degrees instead of using the term Computer Science.

I suspect the university marketing teams will eventually change this to Computer Science much like "Natural Philosophy" evolved into "Physics" at many places though...

https://en.wikipedia.org/wiki/Informatics

https://www.ed.ac.uk/informatics

KCL was Informatics and it was broadly classed w/ maths/'natural sciences' when I was there 10 years ago - according to a younger colleague they have since rebranded to CS.

In the UK, the A level I took was called "Computing". My university called it that as well and treated it as an MEng rather than an MSc, but most UK universities still call it Computer Science I think.

Algorithmic Engineers? CS is essentially the study of automatic algorithmic processes that scale, so something along those lines.

Philosophy of Computer Science [pdf] ( https://cse.buffalo.edu/~rapaport/Papers/phics.pdf )

And the HN post about it - https://news.ycombinator.com/item?id=20912718

There more than a few thoughts in there about what computer science is and isn't.

Weightlifting isn't football. But if you get huge enough, your size from weightlifting predicts success in football.

Even if CS doesn't directly translate, if you can jump through the mental hoops, it likely predicts success in software engineering. You will likely do well in S/W Engineering if you have a high IQ and you have attention to detail. Since it is illegal to filter people by IQ in the USA, you have to jump through an expensive hoop to prove you have a high IQ, which also happens to vet for attention to detail and ability to focus on tasks. There is no reason that s/w engineering couldn't be taught using an apprenticeship model after doing an IQ test, but the later is illegal.

A lot of software writing jobs don’t require a degree although in the the usa.

Those tend to be primarily web development roles (i.e HTML/CSS) or doing some basic scripting not for "full stack" devs creating complex web applications.

I do agree though that somebody wanting to enter a software development role might as well skip the CS degree and learn the basics to enter these 'entry level' dev jobs while learning the more complex stuff in parallel to level up

I'm not aware of it being illegal to use IQ as criteria in hiring in the USA, do you have a source for that?

I'm sure the reason it's not done is because it would be an extremely bad predictor of success in a job, not because it isn't legal.

https://richardhanania.substack.com/p/woke-institutions-is-j... :

> … in Griggs v. Duke Power Co. (1971) [1], … the Supreme Court ruled that intelligence tests, because they were not shown to be directly related to job performance, could not be used in hiring since blacks scored lower on them, and it did not matter whether there was any intent to discriminate.

> [1] https://en.wikipedia.org/wiki/Griggs_v._Duke_Power_Co.

For some reason, there are a few parallel IQ tests that are allowed to continue, despite the Supreme Court ruling against IQ tests. The NFL still uses the Wonderlic test and the US military has IQ tests. If the Wonderlic test did nothing, might not the NFL drop it? As a single number, the IQ test has indeed been found to be a fantastic predictor of the success of a person, despite objections. People hate on it all day long, and perhaps their reasons to hate it are correct, but it remains an extremely useful as a _heuristic_, if you're allowed to measure it. The alternative seems to be 4 years of schooling that saddles the student with crushing and increasingly-crushing amounts of debt.

* https://en.wikipedia.org/wiki/Wonderlic_test * https://en.wikipedia.org/wiki/Army_General_Classification_Te... * https://en.wikipedia.org/wiki/Intelligence_quotient

Off-topic, but I've known a few people that went pro, and in all cases, the causality definitely went in the opposite direction. They all hardly ever touched weights until they were told they'd have to for the combine. And then they got very good at it, but I'm not sure it made them any better. We also had a guy that set a youth world record in the deadlift. He was pretty good, but no pro. The agility tests are supposed to be better, but I've never seen an instance of anybody getting a whole lot better at them through training. The basketball combine is even harder to analyze. One thing that stood out as predictive was the slope of the unweighting curve prior to motion. Good luck training that. They lift weights mostly for injury prevention.

Not saying there's any relevant analogy to CS.

> the slope of the unweighting curve prior to motion

What in the dickens is that? A graph of instantaneous force production?

At rest the player exerts force (weight) on the ground. When the player begins to jump or pivot, they have to pull their feet upward to either move them, or lower their center of gravity. This can be measured as a decrease in weight (force) exerted on the ground.

This is almost universally true in hockey. Promising prospect makes the NHL, disappoints in his first season, hits the gym and packs on 20lbs, comes back and lights it up.

I know very little about what actually happens post-draft, but I’ll admit that many superstar players seem to add supraphysiological muscle mass over their careers.

> Weightlifting isn't football. But if you get huge enough, your size from weightlifting predicts success in football.

Yes, but I imagine most people that got (some theoretical) degree in weight lifting or physical science wouldn't lament that it didn't teach them all the things they expected it to about being a good football player, but that's exactly what we get with Computer Science and Software Engineering.

A good foundation helps with some aspects of good software engineering, or helps steer some decisions, but by no means does it necessarily teach good software engineering (some programs might have some classes). I doubt that keeps the students from complaining about how they aren't actually being taught specifically to program, just like they did when I was in a CS program a couple decades ago.

I wonder if the training colleges that teach programming actually focus more on good software engineering practices? If so, they might actually come out of those institutions better prepared for the low end of the market than CS grads. For the middle of the market, a CS degree might give a leg up. For the high end, I imagine personal drive to learn and aptitude has long surpassed whatever advantage a CS degree gave. From what I've seen, the most knowledgeable and respected and accomplished people don't seem all that heavily weighted towards having a CS degree over another one.

My first undergrad degree was in another engineering major that wasn't CS -- you can guess which from my name. Wall Street was heavily recruiting from my major, despite us being taught absolutely nothing about finance or even business and only having the basics of CS teaching. The reason they were doing this is because we were smart and knew how to decompose really hard problems. We basically knew nothing about financial tech, but they wanted us because if you can do X, then you can probably do Y.

Even in my other engineering major, the complaints were the same. Everybody complained that You Aint Gonna Need It (YAGNI) for many classes and that _real_ industry wouldn't use it. However, if you can learn and use these really hard concepts, you can probably learn and use these _other_ really hard concepts in industry.

> There is no reason that s/w engineering couldn't be taught using an apprenticeship model after doing an IQ test, but the later is illegal.

This is literally how Software Engineering is taught. The IQ test is the brain teaser questions that you get from the interviewers, and the apprenticeship is the 2-3 years of work that you do for the company before you bounce to the next place for the pay raise.

My point is that the current system isn't even that great at teaching academic CS. Even learning a version of CS which doesn't pretend to be SE well requires developing practical skills which many students don't have the opportunity to acquire until after they graduate. This isn't a flaw that's specific to CS, it's true of many disciplines. In fact, in some ways CS is less guilty of it than others.

Does it matter? If remote working is completely viable, I don't see why remote part time formal education wouldn't be. I'm not sure why someone who preferred WFH full time would otherwise decide that in person classes were important.

Very good point. This is similar to stiff you do in business school, where you sure can talk about why BetaMax, but you can't experience it. Similarly you can be told that it's a good idea to think about your abstractions when coding, but if you don't get enough time to build a ball of spaghetti, you might not get it straight away.

I think a lot of uni stuff is like this actually.

Some of the online CS master's programs allow you to take in person classes if you want, so when work is busy you can make progress online and you can pick the in person classes that really interest you. The program you're talking about exists, though it may not be marketed that way.

I did this a few years ago and highly recommend it.

Reply on news.ycombinator.com