Getting up and running with Bang is too hard

[markwhiting]

There appear to be a few ways we can simplify to make it easier for others to use bang and I think its worth having a discussion about what that might look like before we decide on a way to execute the change.

As an example, a project I'm currently work on relies on Empirica, which is amazingly easy and quick to set up. Their architecture is quite different but I think this should be about the level we should be at. And for more context here's their getting started doc.

How can we get there?

cc'ing @deliveryweb as I'd love your input.

[i-gao]

One thing that is pretty counterintuitive for running locally is that you have to add the test users yourself before you can run anything. Can we set things up such that we automatically add those users upon init?

[markwhiting]

For now, a short term solution to this is to improve documentation.

A long term solution would be shell script that gets everything set up or something like that.

[phoebexy]

@markwhiting +1 on improving documentation short term

[xehu]

Agree with all above comments. Do you all think that it's worth having a dedicated meeting at some point to sort out documentation issues and how we'd like to package up Bang?

[xehu]

@deliveryweb the only action item here is optional: if you have thoughts on how we can make bang easier to set up, that would be great!

[markwhiting]

@deliveryweb have done a lot of improvements on documentation throughout and thats great.

I think on the other side, this requires a relatively refined top down effort that will probably involve a lot of architecting. I think it's ok for us to not do this now as I suspect it will not impact the experiments we are currently running.