I thought the conversation about testing was fairly interesting, even if it did feel like ships passing in the night at times. There were several points I thought were well stated. Off the top of my head those were

1) testing to avoid future mistakes, not past ones
2) TDD as a way of forcing some better design decisions (i.e. good for testability is usually also just good)
3) Coverage being a meh metric

I don't recall if it was mentioned but I also like tests as a way to understand what someone thought this code was supposed to do. I think good tests are useful as a form of documentation, which if maintained properly doesn't get out of date as fast as other forms? That might be optimistic in some cases :P

Took a call today for a role where they want a "react/redux/kubernetes/hadoop" expert. Good luck with that.

If you're trolling, it's elite level.

This industry is weird. The industry is really broad so there is tons of room for folks to have different viewpoints, strengths, and expertise even within fairly narrow niches. It's very easy to find folks that are good at things that you don't value highly and vice versa. Because of this I'm hesitant to dismiss anyone as a troll, idiot, etc and more inclined to try to see how someone may be right than to dismiss their comments. Not to bash davet but if you looked only at his posts on interviewing and job hunting it would be very easy to be dismissive of him, but when it comes to specific technical discussion he's a great contributor.

In any substantive discussion or code sample there is something to be learned (usually) though it may take more or less effort to uncover.

Hah, do they actually "require" all of those things?

We have a role with a very similar list of technologies, but I don't think anybody is expecting to hire someone that perfectly fits those requirements.

I'm totally with you - I just meant in a provocative way, maybe worse is better and doing the right thing and learning to become better and all that are just not worth the trouble. So as much as I like to push people towards learning the right way and doing the right thing and understanding how things work from different angles, I think it's intellectually honest to also admit that, maybe it doesn't matter as much and it's sometimes better to continue to do the wrong thing, live in ignorance and just attack and ignore these preachy busybodies who dare suggest that you don't already know everything.

cb,

I def think you're a troll, but that doesn't mean your posts don't have value. Many high end trolls simply belabor esoteric details which is generally every single argument wrt programming.

wrt ds and ml: k. Also the "deep learning and GANs and other cutting edge stuff" is just not useful for 99.999% of stuff. I guess if you wanna become a 1%er data scientist and work on problems that those are appropriate tools for, great! But the vast bulk of data science is basic scraping/munging/Stats 101.

I've expressed harsh words about doing FP in non-FP languages many times in this forum, for many reasons.

I only react to the situations I run into the best to my ability. Coming from a different baseline of life experience and education doesn't mean that I am axiomatically contrarian; it means that I don't have the frameworks to understand the situations as you do.

My self-learning is very much as close to standard as I can muster, and it seems consistent to me that I'd suggest everyone be on the same page when programming, for instance. I think that, as a strict autodidact, I value the desire to learn the rules before breaking them much more than others.

Well it was a (senior, 1st party) recruiter so I told him he'd have a better chance of finding a leprechaun riding a unicorn. More or less . I'm pretty sure they want a react guy who knows some nosql so yeah.

Speaking of belaboring minutiae, I'd argue this is only true because the vast majority of what people call data science isn't data science.

At a certain point I guess the common usage defines the term rather than the original definition, so... I guess your point stands.

That's fine because I don't have much interest in the practice of data science and have no interest in becoming a data scientist - I mentioned the phrase because there's an overlap between data science and more serious AI research, both in terms of tooling and theory.

https://en.wikipedia.org/wiki/Data_science

Seems like an apropos time to educate yourself.

lol Thremp.

Edit: By the way, I know what you do and I know what I do. I'm pretty comfortable in my education level here.

having had many discussions with cb, on points of both agreement and disagreement, i can attest he is not a troll by any stretch.

jj,

"lol Thremp" is your best rebuttal followed by an ETA with an oblique "I know things about you, but you don't know them about me"? Grow the **** up.

gm,

That was probably unfair on my part due to the negative connotations. You can sub out "intentionally provocative and pedantic".

What did you want? You posted a random quote from Wikipedia that didn't even prove your point. Then you ended with your standard tactic of implying everyone is more ignorant than you. Meh.

Are you really arguing that data science = computer science, and you're rustled because that archaic meaning that is no longer in use has been supplanted by another meaning?

My apologies for assuming you had moderate levels of reading comprehension btw. I'll bold the parts that I think merit extra attention next time we do this.

It was mentioned by me when I said getting an A+ in failure is still failure. I think testing is better with permission to rewrite code, as there is no real good reason to test against bad assumptions (or is there?).

If data science is mostly scraping/munging than I get to call myself a data scientist, and that sounds kinda cool, so I'm on board.

Put both cb and thremp on ignore plz. The former while probably a smart and capable programmer is incredibly irritating to interact with, the latter is a textbook troll. Why waste your time?

I'm not really sure what you mean by "A+ in failure" but I'm thinking of writing tests as something you do when you write the code, i.e. TDD, not something you do afterwards. I agree that rewriting bad code to be better (with new tests!) is better than writing tests to make sure that bad code works as badly designed.

Thremp, the extra funny part is that I wasn't even disagreeing with you!

(And no, I don't mean "data science = computer science", I have no idea where you got that).

If you really want to get ahead of the hotness, call yourself a ML Engineer.

wn,

Just get hired by a guy who doesn't actually do anything but has this title: https://careers.microsoft.com/jobdet...ed&show_desc=0

You made it!

Reverting to basic, useless drivel in the programming forum. Not a good look. Your previous nonsense about "code quality" was at least amusing.

hahahahaha

I think you meant == or maybe === if you are one of those people who think type checking actually catches bugs (lol).