Hacker News Reader

Comment by netsharc | Hacker News Reader

A CS prof of mine said studying CS to later do programming is like studying architecture to later be a house builder. I find it an apt comparison.

View on HN

Many (FAANG and other) companies interview with algorithm puzzles that are fairly well aligned with undergraduate algorithms courses. Perhaps one purpose of such interviews is to select people based on how recently they took an undergraduate algorithms course or equivalent.

This is the only reason I can think of for still using this interview technique.

They use it since it correlates better to later performance review scores than anything else they have tried.

The test they use to select employees correlates with their rating of employees is pretty circular. Any arbitrary test will do that.

The only way they could know that is if they occasionally hired someone who failed the interview to see if they also fail at the job.

It's brilliant really. How else would you select for people straight out of school (age), with the time (few commitments and probably no dependents) and willingness (willing to jump through arbitrary hoops before you're even their boss) to grind all variations of the interview problems, all while maintaining plausible deniability to discriminating in these ways?

As someone who graduated recently and is going through this process right now - I always thought I would be set because I'm a decent problem solver, aced all my algorithms and data structures classes and generally could solve most interview style problems I came across. I've learned that this is not enough. An organic problem solving process might involve trying several promising approaches, or starting with a suboptimal algorithm and realising improvements to it, and then you might figure out an optimal solution. In many of these interviews the time constraints can be absurd, you basically have to have done the questions or variants of them recently to be able to write down the optimal solution in your first iteration.

Today my friend had a first round online screening from Atlassian - 5 questions in 90 minutes, and none of them were trivial warm up level questions. Compound this with the fact that it's often harder to solve problems and think creatively when you're under time pressure in an interview, and you realise that your only option is to do 200+ Leetcode problems and just hope your interview overlaps with those problems.

Maybe so, but I'd still rather have my house built by someone who studied architecture, because they will have unique insights in the moment that others would miss.

Also you get a better house if the architect who designed it also builds it.

And eventually the house builder gets enough experience to become a much more useful architect because they know what's actually possible and what is purely academic.

I think the point the original commenter is missing, is that in tech the architect and the builder is usually (almost always) the same person. We don't have the same separation between architects and carpenters that the construction industry does.

This comparison is absolutely correct. I've been in the computer field for over 20 years with a CS degree, and never really needed it.

Most businesses only care about making money. Unless your unique algorithmn costs less to implement and can make money right away, you are out of luck.

Hell, all the stuff you learn at school about algorithms and faster ways of doing work, has been negated by hardware advances that mask poor coding practices.

Go get an information systems degree, learn to code, and make money. The world is going to need coders and IT people for a very long time.

All business only care about making money.

Well, that or getting acquired....

That’s just a roundabout way of making money.

> all the stuff you learn at school about algorithms and faster ways of doing work, has been negated by hardware advances that mask poor coding practices

Sometimes it isn’t, and we end up with Electron on the desktop :(

This is extremely dismissive. You don't want to live in a world where none of the people building things have CS degrees.

Side story: this is even true for pentesting (disclaimer: not a pentester, I was helping a friend to break into the career). The most recognized certification (OSCP) is about killing CVE's and dodging rabbit holes, that's it.

What I did on hackthebox.eu recently was:

reverse engineering x64 assembly into C (silly, I should've used Ghidra instead). Then I started exploiting the heap by poisoning the t-cache, with that I could overwrite the GOT (getting an arbitrary read/write). Leak libc's base address bypassing ASLR, then using return to libc to exploit the thing (a ROP technique).

</technobabble>

Most of these things I learned during my CS master and it's apparently overkill for something like OSCP. More importantly, not all skills overlap. Dodging rabbit holes is not something you learn well. Understanding how to reverse engineer a binary has not much merit when it comes to finding a CVE and using metasploit, or some script on Github. Dodging rabbit holes and being fast at finding the right CVE's and using them properly to exploit a target are skills in their own right.

In part I agree, and of course I am biased since I studied CS but my goal is to be a house builder for a while and then have enough experience to be a proper architect. That's when the CS background should help again. So, study to be an architect, be a house builder for a couple of years and then be a properly educated and experienced architect from there on out?

I wholly agree with this. I loved the mind bending nature of complex CS problems, including Algos & Automata. Yet coding is a snore fest for more than a few days.

I felt like I went from a forward looking, enlightening major - to become where you end up in the details worrying about semantics. CS is mind expanding; coding is specialist mind numbing boredom.

I'm not sure it's that apt... maybe it is, so long as you grant that architrctural study can help you design and build better houses.

Sure, you can get the knowledge in other ways, and sure, some of the houses built by people who haven't studied architecture are great too.

Some aren't though, and their builders could have benefitted from a bit of theory ahead of time.

It's almost more like the relationship between materials science and hoise building - it can enable you to do a better job.

For those saying you've never been helped by your CS degree, I think you either took a bad degree or you'be been stuck in un-taxing work. You've seriously never bothered thinking about, say, algorithmic complexity? Or other topics that should have been covered?

>For those saying you've never been helped by your CS degree, [...] You've seriously never bothered thinking about, say, algorithmic complexity?

There are lots of successful working programmers that will never need to deal with "algorithmic complexity" because "programming" encompasses a very wide spectrum of skills. I attempted to explain a rough categorization of 2 groups:

https://news.ycombinator.com/item?id=12079697

So yes, a programmer in group 1 -- let's say a 30-year veteran of COBOL -- can conceivably never have to make a decision about designing an algorithm & data structure to be O(n) vs O(log n) because he works at a higher abstraction level using COBOL. His domain of work doesn't need to deal with low-level things like raw pointers to linked-lists in C/C++ where Big-O complexity is considered when engineering a custom b-tree from scratch.

The programmers in group 1 are not any "worse" than group 2. They just do a different type of programming.

And, on the flip side, you don't need a goddamn degree to think about algorithmic complexity. It's pretty basic stuff at the level most developers ever get to care about it.

You don't need a "goddamn" degree to learn any single thing you can learn on a degree, and this applies to pretty much any degree you can think of, the books are available, as is online research.

The point is during a degree you are lead to some topics and guided through them.

> "...let's say a 30-year veteran of COBOL..."

That would be around the era of the Intel 80486 processor. All of us cared a hell of a lot about both algorithmic and data structure efficiency back then since computers were so feeble compared to today's systems; people who didn't couldn't produce programs that functioned properly.

>the era of the Intel 80486 processor. All of us cared a hell of a lot about both algorithmic and data structure efficiency back then since computers were so feeble compared to today's systems;

In the ancient 1990s, A COBOL programmer writing "green screen" terminal apps to hit a database on an IBM 360 mainframe doesn't deploy to 80486 desktop pcs.

Even today, if you type "COBOL programmer" into monster.com to search for jobs, most of the results will be programming work for mainframe environments instead of targeting consumer computers like iOS/Android/Windows/macOS/etc.

In the 90s, I wrote C and Fortran for DEC VAX and Stratus VOS mainframes. Even then our major concerns weren’t “algorithmic complexity” and being able to reverse a binary tree on the whiteboard.

You don't need to reverse a binary tree on a whiteboard, but you should have some idea of the complexity of what you're writing, or you'll write crap. Even in a CRUD context.

A 486 was more than powerful enough to run most workloads at the time even in high level languages without too much optimization. Anyone who thinks a 486 was “feeble” didn’t spend time optimizing assembly on a 1Mhz 65C02 by doing as much zero page memory access as possible to save a clock cycle.

If your COBOL veteran hasn't got at least some grasp of complexity, he's a pretty poor programmer. In fact that applies regardless of the language.

I worked in a building designed by one of the great architects. The roof leaked.

I ran into a guy responsible for roof repair for company buildings on a flight. He said, "the greater the architect; the worse the roof leaks."

And that’s the issue. For most people college isn’t about “being a better citizen of the world”. It’s about getting a job. There are relatively few positions for pure “computer scientists”. There is also a reason (as stated in the article) that many companies outside of Silicon Valley won’t hire new college graduates. Many of them are basically useless as developers.

except house builders aren't required to licensed architects to get a job as a house builder.

If you only need to glue code (house building) - no degree (or same self taught CS knowledge) is okay. However, most business problems will require some design choices beyond 3rd party vendor offers. Be it a library API design, software system architecture or something similar

Undergraduate CS courses do not teach library API design or software system architecture.

they do that by teaching fundamentals - problem solutions(thinking) with algos and their use cases, math, how to make/read a engineering chart, what is involved in IT systems and etc. You cannot skip those things. They need to be learned to design good APIs and software systems.

Reply on news.ycombinator.com