Item — HN Reader

I have History and CS degrees. The CS experience (at a top engineering school) felt fairly vocational. But being rigorously taught how to research, construct arguments, and communicate ideas in my Liberal Arts degree has been what made me successful past my first hire. I definitely recommend double majoring in a Liberal Arts degree.

Also perhaps there’s a lesson here about the CS curriculum being too vocational at most colleges? Or not needing a 4 year college if it aims to be vocational?

4 comments

The problem with colleges teaching vocations, such as CS or engineering, is that they are too often remote from the realities of what is actually needed by professionals working in the field. To be honest, most vocations are better taught by apprenticeship, but we turned our back on that path a long time ago. I know a professional journalist who believes the same of his profession. College is not an especially good way of training someone for a profession, it's just that we have nearly eliminated the other ways of doing so, and thus by default it is pressed into service for a task it is not especially suited for.

> Also perhaps there’s a lesson here about the CS curriculum being too vocational at most colleges? Or not needing a 4 year college if it aims to be vocational?

I heard a claim that sounded plausible to me: CS in the US originated from electrical engineering, and so is based on a degree that is based on the ability to directly influence things and make stuff happen - pretty vocational. In Germany CS is derived from maths and places a greater emphasis on the theoretical underpinnings, and yes, the ability to do scientific work and research.

I guess the optimum lies somewhere in the middle: It's mildly shocking to me to hear of US CS bachelors who never heard of complexity (Big-O and all that), while it's also a bit mad that it's possible to obtain a CS bachelor degree in Germany with the only experience in making computers do stuff being 100 lines of Pascal (although that would require a rather arcane set of elective courses, so it probably won't happen all too often)

This actually depends on the university, with CS in some colleges in the US splitting off from the Math department and others from EE or CE.

In Italy there are both disciplines: informatica, that is derived from maths and ingegneria informatica that is derived from engineering. I thought that everywhere in the world there was the same distinction, I had no idea that entire countries had only one or the other but not both choices together.

I also studied both, but only graduated with a History degree. I find some of the things History has taught me that CS and engineering lacked were: critical thinking, understanding power in human relationships, writing to think deeply. But the main topic that it has drilled into me is to consider Second Order Effects.

> Or not needing a 4 year college if it aims to be vocational?

US colleges could absolutely drop the fourth year tomorrow if they wanted to. Many excellent European universities get it done in three and get results just as good.

The real waste of time is the US PhD - doing in seven years what could be done in two or three.

The reason the US PhD takes five to seven years is because most students enter it without a Master's degree. In the UK it only takes two or three years because they generally need to complete a relevant Master's degree first.

> they generally need to complete a relevant Master's degree first

No that's not the case at all - most people do a three-years bachelors and then a three-year PhD. That's the expected route for a PhD in the UK. Doing a masters in between would look like you were starting to drag it out a bit. And the first year of your PhD is worth a masters if you chose to drop out at that point.

For mathematics, all the top-tier UK universities expect you to do a one-year taught masters (i.e. not a research masters) before starting a PhD. This certainly includes Cambridge, Oxford, Warwick, Imperial, UCL and KCL. Cambridge calls its masters a "Part III" or "Certificate of Advance Studies" but it's definitely just a regular masters, albeit a really hard one. Mathematics PhDs in most of these institutions have a sort of informal target length of three years but usually take four years in practice. Funding is usually three years but often four, and the fees are often a tiny fraction of the usual amount in the final year to reflect that there is often a difference.

I have much less knowledge of other subjects, but from the little I've heard the same seems to be true of chemistry and computer science.

In short: I agree with the parent comment, at least for mathematics (and probably for all STEM subjects).

Cambridge for example does not require a masters to begin a CS PhD (it doesn't even require your undergraduate degree to be in CS.)

I don't know about other fields, but applications to do a CS PhD in the UK are usually pretty informal. It's a case of finding an advisor and if they want to take you on that's usually basically the only requirement. How they judge that is up to them.

Not many candidates finish a CS PhD in three years. Four is probably the mean, with three and five being the typical upper and lower ranges.

Often the funding dries up after three years, but the student soldiers on anyway because they're well past the point of no return, or because their advisor can employ them in some other capacity like a teaching or research assistant.

Really? Is this a relatively recent change? I have a degree from the UK and what you're saying runs counter to my experience...

Does this depend on the field? In my experience most students do a three year bachelors, a one or two year Master's, and then a three year PhD.

>The reason the US PhD takes five to seven years is because most students enter it without a Master's degree. In the UK it only takes two or three years because they generally need to complete a relevant Master's degree first.

In a number of countries, the MS is 2 years, and the PhD is capped at 3 years. That's still a total of 5 years. If you don't get the PhD done in 3 years, that's fine, but the funding is capped at 3 years, so pretty much everyone gets it done in that time frame.

The reason it takes 5-7 in the US is because the advisors want it to be that way, and because of the glut in funding in some programs.