Hackernews Julia criticism bingo

@Harmen Stoppels bought this up on slack a while ago

Harmen Stoppels

do we have a hackernews Julia criticism bingo card?

Viral Shah

Seems like we should put it up on julialang.org.

Chris Rackauckas

Yeah, it’s best to just own it.

Then it can be a thing that when a complaint does that we can just do “Bingo! Link”. It’s both fun and lighthearted but at the same time is a slamdown on how unoriginal the comment is.

So a bingo card normally has 24 items (middle square is “free”).
And those are normally drawn from 75 options, but obviously that is not a limiting factor.

I further suggest each bingo card entry should have a title, of the event for the bingo card.
But further, I think it would be good to have a short summary of related info (and if we make a website for it, those can be linked.)
I suggest the info should be:

  • Event
  • Example discussion
  • Link to key criticism blog posts:…
  • Link to key refutation blog posts:…
  • Examples in the wild

For example:

  • Someone bought up 1 based indexing
  • Example:
    • Critic: LOL WAT 1-based indexing?! Every language uses 0 based indexing!
    • Counter Critic: Actually lots of scientific/technical languages use 1 based indexing, like Matlab, Mathematica, Fortran and R. Mathematicians almost always use 1 based, as do people in normal language.
    • Critic No it is bad, and makes algorithms harder to write.
    • Counter Critic Some are harder, some are easier.
    • [Cont. mentiong Dijkstra, OffsetArrays.jl, FFT with indexing based on -N, TwoBasedIndexing.jl…
  • Critical blogs: [add me]
  • Refutation Blogs: [add me]
  • Examples in the wild: https://twitter.com/Sh3W0If/status/1037338085468631042

  • Someone linked Dan Luu’s 2014 post
  • Example:
    • Critic: What about https://danluu.com/julialang/ ???
    • Counter That is from 4 years ago, and was from julia 0.4, since then 0.5, 0.6, 0.7 and now 1.0 have been released… it is a pretty different world now
    • Critic: Well what has actually been fixed?!
    • Counter basically everything mentioned. The language itself is comprehansively tested, as is the ecosystem using largely the stdlib test package introduced in 0.5. Several rounds of improvements to consistancy in naming have occurred. Argument order was fixed in #19314. etc. etc. etc. it is basically all fixed.
    • Critic: Well what about people being jerks? Sounds like a really toxic community.
    • Counter: It really isn’t. Everyone has been super friendly in my experience. Come join the Slack and/or Discourse and findout for yourself.
    • Core Dev: While everyone is normally nice, we do need to be careful not to slip. If you ever have an issue or any concerns at all, please (for our sakes) contact one of the community stewards. The email is listed on the Community Standards page, feel encouraged to contact them even if you are not sure anyone has truly broken any rules.
  • Critical blogs: https://danluu.com/julialang/
  • Refutation Blogs: [add me]
  • Examples in the wild: https://news.ycombinator.com/item?id=17204967, https://news.ycombinator.com/item?id=16323084,

Feel encourages to build upon partial answers from others. Like the one above, which are incomplete.
Once we have 24+ roughly down (or maybe before),
we can move to a github page.

22 Likes

A list of some other events, (but I’m not going to write up full cards just now. Please someone who has had there fights do so).

  • Someone complains that Fibonacci benchmark uses recursion
  • Someone brings up that so many packages are broken in 1.0
  • Someone complains about precompile times
  • Someone complains that the standard library doesn’t include plotting
  • Someone complains about packages not working in windows
  • Someone complains about garbage collection
  • Someone asks why you need julia when you have Numba and Cython
  • Someone complains about status of Databases
  • Someone complains about Julia Computing being an Evil For-Profit company
  • Someone complains about breaking changes
5 Likes

My “favorite” was someone on HackerNews arguing that writing fast code in Python being less accessible was an advantage vs Julia. That it was best to contain academics and scientists to calling library code only, rather than writing any of their own.

5 Likes

Ahahah. Do not let them write real code, they will become unstop-able.

that sounds like an even more twisted version
of the argument that says:

Not being able to write for-loops, is a feature, since vectorised operations are clearer to read.

Which is itself, I think, a mis-characterization of the argument for having Einstein Sum-like notations.

1 Like

While it is good to address some commonly occurring misunderstandings, IMO it is not worth it to spend to much effort on this.

These debates tend involve people who know little about languages and compilers, and at best have invested in learning a single language and feel that they have to “defend” it, otherwise their investment is lost. Others will just try Julia and see for themselves; I think it is better to focus effort on making their lives easier than to get bogged down in these discussions.

17 Likes

I think the purpose of the bingo cards is exactly to avoid being bogged down in discussions.

7 Likes

results

see https://myfreebingocards.com/bingo-card-generator

4 Likes

While I understand the sentiment and that people want to have some good-natured fun here, if you’re gonna do this, I think you should leave references to specific people’s external blog posts off (1-based indexing etc is fair game). You just risk re-opening old wounds there. In other words, specific technical criticisms are fine, but please stay away from mentioning specific people, even, or especially, if they are vocal critics.

34 Likes

Personally, I think this sort of thing is distasteful in general. Doubly so when mentioning Dan Luu by name.

6 Likes

+10 to Keno’s comment.

The criticism about “people being jerks” and “come and join and find out for yourself” is not about how nice people here are to newbies or each other. It is about how the community reacts to internal and external criticism.

In short, I read the criticism not as “julia community consists of jerks” but rather as “julia community behaves like a cult, including preaching and excommunication/unpersoning of critics”. That’s not something people can judge by hopping on slack/discourse. And that is definitely not something that is addressed by making a bingo (or reopening old flamewars, or writing heated defenses, or angry rebuttals).

I do not concur with that criticism. But I am not a vocal critic, nor do I know any vocal critics personally, so I couldn’t tell. I can say that the hackernews threads do get a bit too heated for my taste, and we sometimes look a tad off-putting in these threads.

Julia is a really cool project and doesn’t need heated defenses because somebody is wrong on the internet, just chill and code.

13 Likes

This reminds me of an older proposal for promoting the Julia culture with Fortunes.jl. And I agree with the objection that Tamas Papp made then. That type of fun is normally fun just for the people that have been for some time in the community. For many others, specially newcomers, it rather looks like scorn, and discourages participation.

12 Likes

I think at best some of these can be FAQ items; this seems very much aligned with the purpose of a FAQ. Entries on “why does julia use 1-based indexing” and “why does the fib benchmark use recursion” seem reasonable. Much of the other stuff is best ignored.

15 Likes

Ok yeah,
I thought this would be fun.
And did not think further than that.
My bad.
Your point is well made.

Jeff’s comment that we can add more stuff to the FAQ.
Is probably a better place for it.
Rather than something flippant,
that is poking fun at the lack of originality in peoples criticisms.

4 Likes

Haha, I had do much of the “Who is Julia? That’s his girlfriend.” running gag the whole summer.