Up-votes instead of down-votes on RFCs

[notriddle]

If you want to disagree with an RFC, you need to actually say why. Content-free comments should just be removed by the mod team as unconstructive.

[nikomatsakis]

Is there any evidence that Rust RFC proposers have gone away, never to come back, because of down votes?

I’ll be honest – I find it unpleasant to see “thumbs down” comments, even when it’s not my RFC, and even when it’s an RFC I disagree with. It just gives me a bad feeling. I like seeing and things like that, it makes me happy. I don’t want to have bad feelings when working on Rust.

[Centril]

Is there precedent for Rust requesting a feature from GitHub and getting it? Have we tried contacting them about this one?

I have no idea, but to get something we first have to decide what we want =)

[glaebhoerl]

Well yes, but it seems like it’d be worth giving that a shot before we start workarounding it by implementing bots.

I imagine we’d go to GitHub with something like “we’d like to be able to disable the -1 reaction on our repos” [and possibly the ‘confused’ one too, which comes off as more of a “meh” in practice], and then maybe they’d come back to us with something like “ok, each org/repo now gets to choose which 6 emoji are available as reactions”. (Which sounds like a reasonable approach to me - but maybe they also have other priorities like wanting the set of reactions to be consistent across their website.)

(If something like that does happen, then I’d also really want to have distinct reactions for “good comment” and/or “thanks”, versus “yes, I agree”!)

[felix.s]

I don’t want to have bad feelings when working on Rust.

People don’t want to have bad feelings generally. And yet, humanity has been suspecting that it’s impossible to have a meaningful, fulfilling life without some suffering, and without accepting that certain wants aren’t going to be met all the time, for long enough that ‘misery builds character’ has become something of a cliché.

Like @petrochenkov said, I don’t want to delude myself that my proposals are universally admired and accepted. Sure, a downvote may be less useful than an explanatory comment, but I’d still take that over pile-ons or no feedback at all. (It’s not like all comments are perfectly explanatory either.)

[glaebhoerl]

‘misery builds character’

survivorship bias

[nikomatsakis]

I don’t want to delude myself that my proposals are universally admired and accepted.

Don’t worry, we’ll let you know. And maybe we’ll even give you some constructive feedback while we’re at it!

It’s not like all comments are perfectly explanatory either

Certainly not all comments are perfectly explanatory, but a comment cannot explain, by its very nature! It is just pure negativity with nothing else.

[steveklabnik]

And yet, humanity has been suspecting that it’s impossible to have a meaningful, fulfilling life without some suffering

Rust development is (hopefully!) only part of any given life. Let’s leave the suffering to the other parts. Nowhere in that maxim is it stated that suffering must be equally distributed through every part of life.

[scottmcm]

Except that an RFC already includes a bunch of self-justification in its text, so in many cases, if someone wants the RFC to be accepted, their reasons for wanting it merged are already there.

The same isn’t usually true for dislikes.

An RFC typically addresses a number of scenarios, and it’s unclear whether a is agreement that the problem needs to be solved, about particular details of the approach, etc. And the RFC is supposed to also have a drawbacks section, so a can just as well be considered an “The drawbacks enumerated in the RFC outweigh the advantages” just as much as the opposite means the opposite.

[notriddle]

Isn’t that why GitHub has both

“good comment” and/or “thanks”

versus “yes, I agree”!

and ?

If these two seen similar, I think it’s only because the person posting the reaction doesn’t have a clear idea about the difference between them. People conflating “good point” vs “I agree” is not exclusive to Rust RFCs.

[comex]

[glaebhoerl]

@notriddle We could try to give them those meanings through coordinated effort (e.g. a bot mentioning it at the beginning of each thread; I don’t think that’d be worthwhile, fwiw), but in the absence of that, they can’t really be taken as more than “I like this” and “I love this” (or “I love you”). Anyway, we’re going off topic

[repax]

I realise that this isn’t viable, but what I’d really like to be able to give, and receive, are and on a far more granular level – such as individual sentences. It might be too much to ask of people to always respond to one’s ramblings with thought-out written comments, even when they care enough to read them.

Internet forums as a medium lack this immediate feedback that you get non-verbally in face-to-face conversations. You really don’t know if anyone at all reads what you (sometimes painstakingly) try to express. Sometimes no response at all isn’t that uplifting either.

Anyway, I’m no longer going to downvote an entire comment or rfc. I’m not going to punish someone for the virtue of taking their time to put their thoughts out there, when doing so in a constructive manner – no matter how misguided or bad the ideas are.

[ssokolow]

The good thing about reviews is that they are tied to some specific RFC text, so they are much more local. With reviews, some concerns can be raised, resolved, and then folded (which GitHub will do automatically) once the concern has been resolved. That way, it is easier to see what issues are outstanding and what issues are resolved.

To be honest, I find the visual experience of seeing (or, worse, catching up to) a wall of review comments (with collapse marks) so intimidating that, lately, I’ve been guilty of only addressing points raised in top-level comments and just hoping that all I’m missing is fine-grained bikeshedding or highly-abstract discussions that, lacking a background in type theory, I wouldn’t be able to contribute to anyway.

(Yes, I realize that my most consequential comment about the ..= syntax would fall under that heading and have never happened in that case.)

[cramertj]

Why shouldn’t I downvote if I think the RFC proposes something bad? That’s almost always a purely technical point of view.

“something bad” is never a purely technical decision, and always includes some degree of value judgement. RFCs are a place for criticism and for discussing technical downsides. They are never a place for telling someone that something they did is bad, terrible, or a waste of time.

I’m sorry but if one’s self-esteem depends on whether one gets downvotes on an RFC then one should go see a doctor.

It’s completely natural to have emotional reactions to the way that other people treat you and your ideas. Criticizing people for wanting to improve the emotional wellbeing of our community members is never acceptable.

[cramertj]

Disallowing the expression of one side of the argument is called censorship

We’re not disallowing technical arguments. You’re welcome to lay out whatever technical position you want. What isn’t acceptable is making negative value judgements that don’t contribute to the conversation (e.g. downvotes, “this is terrible”-style comments). Those aren’t arguments-- they’re namecalling.

[H2CO3]

Have we really come to the point where a simple downvote is considered name-calling? I just don’t buy that.

[MajorBreakfast]

I think @H2CO3 generally does a great job coming up with technical arguments. But I have to agree with @cramertj on this point. Downvotes aren’t helpful. I think the Discourse approach which lets us only react with the heart emoji is the better. It’s healthier. Downvoting isn’t healthy.

[H2CO3]

It’s not any less healthy than upvoting. Upvotes and downvotes are completely symmetric. Agreement-disagreement. If downvoting without an explanation were to be disallowed, then upvoting also should be disallowed. Not doing so induces a bias in everyone who views it, including those who will ultimately accept or reject the RFC. If an RFC only has upvotes, it might have any number of comments explaining why it’s a bad idea; people will just see that it has many upvotes, ie. others like it, and will think that therefore it’s probably good. Let’s be realistic: not even the responsible team’s members read all comments. Should they decide the acceptance or rejection of an RFC based on a heavily asymmetrical, biased, unfair voting system? OK, maybe a downvote without a comment is not useful. But then an upvote without a comment isn’t useful either.

[cramertj]

FWIW we don’t take emoji reactions to an RFC into consideration at all in our decision making, and I think that should continue to be the case. I use upvotes as an easy way to say “I’m excited about this RFC! Thanks for all the work you’ve put in to get it to this point.” I don’t use them to indicate that I am in favor of merging a proposal, and I don’t downvote when I’m opposed to merging an RFC.