zlacker

[parent] [thread] 2 comments
1. ordina+(OP)[view] [source] 2018-05-01 19:25:57
Your "shibboleths" come off as somewhat smug and self-congratulatory, IMHO.

E.g.: > You think comments are a last resort for documenting an algorithm.

Right, would like to see you implement tf-idf weights in a vector-space model without any comments. Super helpful for devs who come in after you.

> You understand the illusion of complexity and why simplicity is always better.

Who's pro-complexity? What dev sits around and says, "Wow, this is too easy to understand, I need to make it more confusing"? Complexity is a result of many factors but intent is typically not one of them.

> You can and have extracted declarative implementations out of imperative code. You know when it’s worth it and when it’s not.

Speaking as a hiring manager I'm not sure I can think of a less applicable signal as to whether or not a candidate would be a good hire than something this specific.

What this list really should be titled is, "A List of Things I Know and Am Proud of" not potential signals for good hires.

Not saying this to cut you down, just to let you know it's written in a tone that's over-specific and (to the average reader) arrogant. If I may I suggest rewriting it sound more inclusive, exciting and positive instead of "I'm so great because I know these things and you should too."

Just my two cents.

replies(2): >>aethan+Q8 >>silvio+Mn
2. aethan+Q8[view] [source] 2018-05-01 20:17:10
>>ordina+(OP)
I agree. I read through this and, despite being able to speak to most of the points, my first thought was "this is not a group of people I want to work with."
3. silvio+Mn[view] [source] 2018-05-01 21:44:40
>>ordina+(OP)
Thanks for the feedback! I'm surprised, for previous rounds, the shibboleths were cited in our submissions as one of their favorite parts. I guess tastes vary. We had two amazing hires off these posts. But that's anecdata.

Point by point:

Comments: sure, that's an excellent case for a comment. I meant more for a simple loop over items, etc. In those cases, if you use clear variable names and idiomatic syntax, it'll work. But for vector-space models? Go all out...

Pro-complexity: this is "pro-complexity" in my opinion: https://engineering.hellofresh.com/front-end-microservices-a... . HN seemed to agree: https://news.ycombinator.com/item?id=15566339

Declarative: fair enough. It's been instrumental in our current project, so it's just something we look for. Perhaps too heavily, based on this feedback.

Overall feedback: that's great to know, I can look at making them more inclusive. Many of these are not at all my accomplishments but things we've collectively picked up from HN in general. Note the self-deprecating humor.

Again, thanks!

[go to top]