Hacker News Reader

Comment by aunty_helen | Hacker News Reader

I've done 433 google searches in the past week with about 70-80% of them being coding related.

One of the things that suprised my bootcamp students when I was doing a live demo was how messy actual development looked. If I've done something before, I'll jump to that bit of code. Copy paste if it's not worth doing a refactoring / it's in a different system.

Straight to google if it's something I don't know.

If I want to check params etc I'll jump into the source with cmd b in pycharm.

If I'm doing too many things that I already know how to do I would be wondering if I'm using my time efficiently.

View on HN

That’s awesome you show students that, it seems like a lot of people coming into this profession are really intimidated by developing.

Showing them that it’s a messy process with few perfect answers is probably one of the best things you could do for them.

I really enjoy my teaching of and in R, where it is preferable to develop most commands yourself. Preferable because it is just hard to google even fundamental tasks because it is hard to formulate your problem. You must use Google only strategically.

Then I taught a class on applied software development in Python and I had to learn that convention cannot be developed by just yourself, so most of the times it is indeed "Straight to google if it's something I don't know."

I think for the latter it is the right way to teach the process as messy as it is, but what does that say about it?

And not to be misunderstood: Most of the problems we face in R are harder to google because the user base for what we do is so much smaller.

I agree.

I'm writing a lot of CI/CD for R and R Shiny apps with various SQL backends for my employer and the broader data science team.

There's so little good documentation out there for R, I've spent weeks in trial and error trying to understand why ODBC connection strings fail and R SQL code is failing. And even more time deciphering why certain R packages in Debian don't work at all.

I've taken the effort to compile a lot of them myself from source (GitHub) and writing the appropriate love letters (aka documentation) for my future self.

Working with R for the first time has been an insanely difficult learning curve for an R language novice. Especially when compared to writing Python where most questions I can't answer immediately are only a search engine away.

R has been difficult, but definitely rewarding.

Reply on news.ycombinator.com