[swift-evolution] Reflecting updated feedback was Re: [Review] Require self for accessing instance members

Erica Sadun erica at ericasadun.com
Thu Dec 17 20:20:37 CST 2015


Doug said stuff, I said stuff, and I stupidly cut off all context. I'll let you figure out who said what:

>> * I wish ideas could go through initial vetting before they appear as a formal on-list [Proposal].  
>> 
>> Ideally, it would be great to pitch the team with a [Pitch] and see if the pitch has a chance. The downside is that the Swift.team has enough on their plate to deal with. It may not be feasible to do this. That said, the sooner you can get a sense of the pitch's health the better:  Is the pitch not worth pursuing, better suited for bug report system, worth pursuing but not for 3.0, or worth pursuing?
> 
> Pushing things into bug-fixes is the easy one of these, since that skips a more heavy-weight process for a lighter one.
> 
> The thing I like about your “[Pitch]” idea is that it indicates a willingness to hear “not worth pursuing”, which is generally a tricky social problem: it is very hard to convince someone that their idea is “not worth pursuing” (even “not now” can be difficult), and the core team absolutely do not want to come across as dismissive. That’s especially true because there are a ton of ideas that we *like*, but the need to focus on Swift 3 prevents us from engaging on them.

I'm basically going for an "I have an idea, does it have legs" sanity check here because what's the point of clogging the list and people's time for something that really has no chance of ever coming to fruition. As I mentioned in a separate reply, "no" can mean "not technically feasible", "not within the right language scope", "not compelling enough a pitch", and "other reason that we're honestly not going to get into because we're not going to fight with you on this and we have to have some leadership here."

I admit there are really good badly phrased ideas, but isn't it a part of life that to make something happen you have to be able to express it and think about it? Forcing an elevator pitch as a first step would cut down on a lot of fluff and nothing I've said prevents an [RFC] from happening before a [Pitch].

>> I don't know what you'd call it, but maybe [Proposal Draft} would be a better subject for discussing and refining an idea.
> 
> There’s a general theme here of being more clear about where a proposal is in the process. Right now, we have “pre-review” and “review”, but you’re looking for something to distinguish the “pre-review” state further: an initial, more free-form “ideas” state which is more about solving a specific problem, then a later “proposal draft” state that means we’re getting closer to the “review” state.

I don't think it even has to be formal. I just don't want it to be this: http://imgur.com/azaPOPu

> The benefit of having more states is that it makes it easier for people to jump into the process at the point where they get interested: one might not have time to consider all of the “ideas”, but would want to weigh in on proposals once they hit the “proposal draft” state and things are getter more specific and more serious. Indeed, this is the reason why we have the swift-evolution-announce list for reviews, so that more casual observers can see what’s made it to the last part of the process at which point it’s time to either weigh in or accept whatever happens.

Having a vocabulary, even an informal one, helps differentiate how "real" a subject is: I'm kicking around an idea, I'm trying to formalize this idea, Is this idea worth pursuing, I have formalized this idea and need a final sanity check before starting the real proposal process, etc.

Right now there's so much stuff in my mailbox it's hard to find the wheat among the chaff. (Although I have a few biased algorithms I'm starting to follow.)

>> 
>> * You may want to test the idea's temperature before starting the formal proposal process
>> 
>> As I learned, Survey Monkey is free and is a good way to quickly test the crowd.
>> 
> 
> I’m mixed on this. It’s so very easy to “+1” a partially-formed idea and then, when confronted with specific details and trade-offs, change one’s mind. I guess that’s fine if you’re going into the survey knowing that you’re getting very imperfect data, and that the “yes”’s that come in are fairly weak and entirely nonbinding.

I think if the "require self" had been put to a survey, it would have gone down in flames without taking up so much list time. I get how passionately people who love the idea feel it will enhance safety, etc. But I suspect it would have given a pretty strong, pretty quick dousing. (I could be completely wrong however about the current vote tallies, and if so, please adjust my opinion accordingly.)

-- Erica


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.swift.org/pipermail/swift-evolution/attachments/20151217/dca83411/attachment.html>


More information about the swift-evolution mailing list