make a 'uservoice' / feedback channel for ghc-mod

[nrolland]

It seems that it would be useful to have a place to collect votes for features, as indication of what is important for the users. maybe a twitter account, now that votes are available ?

[DanielG]

I like the idea but I'm not sure twitter is the most accessible forum for this sort of thing. Ideally one would have a presence on like all of the social networks but that's a rather unfeasible amount of work.

I am a bit dissatisfied with GH issues as things stand tough at least for discussions. Generally I prefer mailing lists for this sort of thing but I'm not sure anyone would use it even if I set one up. Maybe a Discourse instance would be nice? I think it has OAuth support so we could add the option of logging in with GH to make it less annoying than signing up for a mailing list.

[hackeryarn]

Why not just make a gitter for this repo? We can still utilize it to hold discussion especially if they are topic driven, and it comes with full chat logs built in.

[nrolland]

I would be in favor of anything that makes it possible to vote. I dont know how we know what people really want or use now

[DanielG]

I'd rather not use Gitter it's just another proprietary cloud service, Github is already bad enough. Besides what exactly is wrong with IRC anyways? With the web client we link to in the README it seems to me it has a pretty low barrier, no?

[nrolland]

IRC is very nice for conversation, but it does not show the priorities, the most wanted stuff, or ways to gather feedback beyond the 27 people who happen to be connected at a particular point in time

[DanielG]

Well that's something that's solvable with a bot or just by counting votes manually. I really don't see what usecase we would have for this? If we want to do user surveys then posting them on reddit/haskell-cafe is going to give them more exposure than anything we could set up anyways so why bother?

[nrolland]

everything is solvable, especially with haskell, but it's better to not reinvent the wheel.

I find it extremely damageable to not be able to cast vote on feature. That means arbitrary choice are made which can be very detrimental to the user base at large.

For instance : How bad is it that stack gets overridden by default with cabal ?

[hackeryarn]

Trello is another option. It could be used to keep a list of desired features and track votes.

[DanielG]

Sounds pretty much exactly as bad as Gitter: proprietary API, proprietary Apps, proprietary service backend -> no thanks.

[nrolland]

trying to parse out what you say.... which characteristic associated with being proprietary makes you reject all of those ? is it just a philosophical stance ?

[DanielG]

Well there are free software alternatives so I don't see a reason to downgrade to any proprietary tool that might or might not do a better job at the cost of control just because they look fancy. So far I'm still guessing as to what we're actually looking for here apart from it having some kind of fancy support for voting which I argue is not enough incentive to use such a thing for other communications or anything really.

I'm pretty sure there's plenty of web services that implement only voting, for example this thing we used to determine what haskell-ide-enigne should be called: http://strawpoll.me/5842105 there are probably others even ones that don't rely on proprietary javascript. So why use some bloated Cloud™ Service?

[nrolland]

ok I close this issue then

[DanielG]

Just to be clear I have no problem with setting up any project infrastructure we might like on a server of mine, I'm planning to do some custom CI setup anyways. I just don't know what exactly you want?

[nrolland]

I think it will take someone else to explain it

[hackeryarn]

I think what @nrolland is looking for is a way to integrate github issues into a voting app like that. This would allow us to see what feature requests people are really interested in.

If you are thinking of hand rolling this type of solution, it could be really interesting. And I would definitely be up for helping out. Just let me know.

[DanielG]

That much I got, the bit I don't understand is why don't we use GH issues/IRC for that like for everything else? Are we trying to target users that don't want to respond to a GH issue/IRC or is it about visibility? Because in the latter case how is a voting app or self rolled solution any better than the existing communication channels?

[hackeryarn]

I think the goal is to quantify the desire for features. Not everyone that's interested in a feature has something to contribute on a GH issue/IRC, and just seeing a bunch of +1's or polling an IRC isn't a very good method to gauge the community as a whole.

[DanielG]

Yeah OK, so I would suggest just to use some single page poll thing (which ought to exist), host that somewhere and send it out a link over GH/reddit/haskell-cafe/whatever to make people aware of it.