Hacker News Reader

Comment by leesec | Hacker News Reader

>I’d love to understand this in detail.

Whatcha want to know?

>When I was at Amazon, we didn’t take boot camp grads very seriously, and in recent years the company only looked into this to increase diversity hiring, specifically boot camps oriented towards either women or the LGBT community.

I think because average candidate quality of bootcamp grad probably pretty low, due to shortness of the program and people thinking all they need to do is complete the bootcamp, when really you need to stay learning for years.

Also Amazon is top tier, wouldn't expect someone with 3 months experience being able to keep up unless they were really top notch.

View on HN

Well, the thing is, if you ever do student interviews you'll realize many people in CS programs appear not to be getting that much from the classes either.

That's a problem with the students, not the CS program. "You can lead a horse to water but you can't make it drink."

Maybe so, but if the argument against entertaining hiring boot camp grads is "many of them aren't good," then one must consider that the argument applies equally well to the other place most people try to source junior devs.

I think the implied problem is that the CS programs are passing these people.

Someone signing up for a class and not learning anything is a problem. Someone passing a class and not learning anything is a big problem for the signaling power of that degree.

My CS minor program didn't cover GIT.

A CS course might cover the architecture of git, or the technical underpinnings but probably shouldn't teach you how to do the basics. In my program this sort of skill was expected to be learned on your own, or maybe a single optional tutorial session with the TA

My uni has an on-prem installation of Github and this is where you upload and collaborate on your solutions. This isn't course specific, this is just what you do.

It’s odd to me that you’d expect them to. The CS program I was in used github for collaboration on a couple projects, but nothing really covering git tooling. All my classmates were either already familiar with it, or more than capable of figuring it out with a search or two. I’m guessing your program covered material that required a little more effort than a google search to build basic competency in?

How do you expect people to collaborate if they can't git? Git is not that easy to grok, not even for Ph.d. students who successfuly completed some very complex scientific research (I am trying to teach Git a few of them so it's very visible to me).

Making it the only way to send in homework motivates people to learn the absolute basics (I haven't seen anyone learn branching like this), but also pisses them off and a lot of times they don't want to use Git at all after that experience.

> All my classmates were either already familiar with it, or more than capable of figuring it out with a search or two. I’m guessing your program covered material that required a little more effort than a google search to build basic competency in?

I did come across this charitable expectation in my CS coursework where you were given tools and expected already to know or to learn on your own. Which is fine if you want to weed people out. I can understand where someone might be frustrated if you hand them a tool and expect them to know or figure it out while they're learning new material.

Problem is, schools expect students to do this with like 100 tools and give no guidance on what to do, and why. It's absolutely overwhelming.

Even as a professional SW engineer for 10+ years, when I join a new job, the amount of new tools and workflows is overwhelming - and that's when I at least understand the descriptions. Dumping all that on a student is just idiotic, not good teaching.

I get what you’re saying. I didn’t feel overwhelmed by git, but I’d been working with it for a couple of years, so there’s a clear bias. I did run into frustration with other tools, though.

Maybe git, in particular, does deserve some specific focus given its unique status and ubiquity in the industry.

Thank you for the acknowledgement. I appreciate your consideration of an alternative view. Not everyone gits it at first and it takes some effort to learn.

If a student is technically inclined, presumably anyone in a CS program, then the bare minimum of git should be accessible. At least what is necessary for basic collaboration. Re-basing & cherry-picking, bisecting, etc. and the implementation details of git along with other more advanced bits aren’t remotely necessary for this kind of collab.

A decent grasp of cloning, committing, and pushing seemed to be enough. Conflicts are inevitable, and will present a challenge to someone new to the system, but they’re not insurmountable. Good IDE integrated tooling probably goes a long way on this point.

Git will be a speed bump for a student that hasn’t already encountered an scm before. I didn’t take a poll, but I don’t recall any of my classmates encountering issues that they couldn’t work out.

CS programs that don't cover version control (as a concept) are a real head-scratcher to me.

Git didn't exist yet when I was in college, but day 1 was still going over version control as a concept and it was made clear you'd get a zero on any assignments if you didn't use it.

> Also Amazon is top tier, wouldn't expect someone with 3 months experience being able to keep up unless they were really top notch.

I attended a bootcamp in 2015 and know two people, one in my cohort, and one in the next, who both got jobs at Amazon right out of the bootcamp. At the time, cohorts were every 6 weeks with 80 people/cohort, so that's ~1 out of 80 bootcamp grads who made it into Amazon (granted, with a small sample size), with more in FAANG generally (Google was more open to bootcamp grads than the others).

Reply on news.ycombinator.com