Skip to content

Guide: reference to "really smart type theorists" unnecessary #17229

New issue

Have a question about this project? # for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “#”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? # to your account

Closed
bjpop opened this issue Sep 13, 2014 · 5 comments
Closed

Guide: reference to "really smart type theorists" unnecessary #17229

bjpop opened this issue Sep 13, 2014 · 5 comments

Comments

@bjpop
Copy link

bjpop commented Sep 13, 2014

Section 5, Variable bindings, mentions "Well, Rust has this thing called "Hindley-Milner type inference", named after some really smart type theorists."

This is a minor nitpick, but I cringed when I read that sentence.

There are lots of features in Rust that are the result of work by smart people. There is no need to single out the Hindley-Milner algorithm.

Also the follow on text seems a bit patronising "If you clicked that link, don't be scared:"

Writing like that tends to perpetuate the idea that type inference is deep magic, when it is not.

I expect many people reading the Rust guide will be familiar with type inference, even the HM algorithm. It is fine to link to the Wikipedia article, but I think the Guide should not imply that it is a difficult topic or something which is scary.

@steveklabnik
Copy link
Member

I expect many people reading the Rust guide will be familiar with type inference, even the HM algorithm.

I very strongly disagree with this, but do agree that I don't have the right tone here.

I think the Guide should not imply that it is a difficult topic or something which is scary.

hindley milner

@mahkoh
Copy link
Contributor

mahkoh commented Sep 13, 2014

I'd call this fallacy "Appeal to unfamiliar notation."

@bkoropoff
Copy link
Contributor

If stylistic concerns are not compelling, there is a good technical one: Rust no longer uses H-M type inference. See the pull request and Niko's blog post that mentions abandoning H-M.

@steveklabnik
Copy link
Member

Oh, I mean, I do think that style arguments are. Didn't mean to imply otherwise. But I also forgot that we switched, so thanks for the reminder.=

@zwarich
Copy link

zwarich commented Sep 13, 2014

Rust also never really used H-M.

# for free to join this conversation on GitHub. Already have an account? # to comment
Labels
None yet
Projects
None yet
Development

Successfully merging a pull request may close this issue.

5 participants