Hacker News Reader

Comment by theonething | Hacker News Reader

Lots of laudatory comments here about this being essential but missing teaching but I have a different take. The content looks good, nothing wrong with teaching these things. But these things can be and are learned on the job fairly quickly for anyone interested enough in the field and with enough aptitude. In fact, I would say these things can be learned on your own time as a side effect of being interested in computers.

So I would say it's good content, but not essential for a CS program.

View on HN

I think it can be useful to distinguish between the "known and unknown unknowns" here. For example, everyone will quickly realise that that they need to know git, and they will learn what they need to know (a known unknown). A university course would maybe save them time, but it would not really change what you know after 2 years in industry. Compared to e.g awk or shell scripting which can be incredible usefull, but maybe not something people realise by themselves that they need (a unknown unknown). A university should make people at least aware of these latter tools.

Totally agree. Further, these skills actively support rapid iteration on learning other core university outcomes. In CS it’s very easy to waste lots of time doing things the wrong way that could otherwise be spent doing something useful.

Good point. Some things you will learn by necessity, but if you dont know about say regexes you probably wont think of searching specifically for such a tool.

Whether you think that:

a) Universities are places of higher learning that do not need to cater to industry

b) There is an implicit social contract whereby universities should produce industry-ready graduates

c) Something in-between

… the skills taught in this course are as useful for the PhD candidate as for the junior software engineer.

Why leave these out or up to the student, then?

Furthermore, in many other academic disciplines, it’s very common to teach applied technique (eg, on writing essays or structuring research), is this any different?

I don't think you're wrong that you can learn this stuff on the job, but a "primer" kind of class like this which surveys several useful tools helps new engineers develop pattern matching skills around how to continue self-teaching these kinds of things. Shell/scripty/vim-ey,linux-ey stuff can be really challenging to learn how to learn for some people.

> The class is being run during MIT’s “Independent Activities Period” in January 2020 — a one-month semester that features shorter student-run classes. While the lectures themselves are only available to MIT students, we will provide all lecture materials along with video recordings of lectures to the public.

It seems that the course is only 1 month and runs alongside some student-run classes. So this is from the get go not an essential course of a CS program.

There's quite a broad mix of offerings during IAP--a lot of which isn't part of a regular academic program and a lot of which is just for fun/intellectual interest. It's mostly all non-credit. And there are a fair number of sessions dealing with practical details that may not be covered in regular courses.

I think it’s dependent on the person learning the material. My college had a similar course and it was also highly regarded as useful by students. However, while I found the course fun I already knew most of the basics from my research job, and the advanced stuff hasn’t really come up again even in work (e.g. fancy git or gdb commands), so I’ve entirely forgotten it; my biggest takeaway is probably ctrl r for searching in shell. But I can see why a guided intro would be really helpful to someone who had no experience or some trouble getting started learning this kind of material (which is very different from programming or computer science).

There’s probably an aptitude threshold past which the course’s value diminishes - don’t mean any disrespect to anyone, just trying to expand on your point. The top students either have already figured it out or will do so easily, so they might be better off doing something else with the time they would’ve invested in this. But for a lot of students learning about these tools and concepts can be a real force multiplier, more so than a random upper level course.

This is also what i thought too. The content is effectively a roadmap on how one could use the computer well to perform tasks (without the algorithmic/programming portion).

If your main usage environment is windows, none of these are that helpful, but the ideas could be translated mostly (windows shell is similar enough, that you can just as easily script them with batch files).

Yeah I pretty much agree with this comment. But I think my take on whether universities should be training for academia or for industry is "why not both?". I think these are just different valid tracks. There is a lot of education that overlaps between "computer science researcher academic" and "software engineering professional with strong fundamentals". So that should be the required set of credits. But then there are different tracks toward where you want to take your degree. Those sound like electives. Academics should learn more about research techniques and publishing and presenting at academic conferences, etc. And those on the professional track should learn more about tools and techniques used in industry.

There's no conflict or contradiction here, just different strokes for different folks.

Reply on news.ycombinator.com