Talk:Discovery/Testing

Propose to split this page

I would like to see one page that documents how tests work, and then a separate page for tests themselves. Possibly even splitting the tests page into "past" and "future", but I don't have strong feelings about that. (Belated signature: --KSmith (WMF) (talk) 17:37, 29 March 2016 (UTC))

Should we announce both before a test, and after?

Dan recently added a task to the standard process to announce (via our Discovery weekly status update) when a test report is published. Should we also announce when we start a test? If so, would that be before it goes live, or would shortly after be OK? That would drive whether we should send an email or post to the status update, since the status update is batched for a week. --KSmith (WMF) (talk) 17:40, 29 March 2016 (UTC)

@KSmith (WMF): Good question. On the one hand, announcements are good, because they broadcast what's going on. On the other hand, an announcement takes time, and the time spent writing two very similar announcements may be better spent elsewhere. We also risk creating an constrictive and heavy-weight process if we try to account for every scenario. I'd prefer to keep the requirement to a single announcement afterwards with a link to the report, with the option of also doing a preliminary announcement for anything considered big enough. Thoughts? --Dan Garry, Wikimedia Foundation (talk) 21:30, 29 March 2016 (UTC)

@DGarry (WMF): I agree about not creating needless work for ourselves. How about when we deploy a test, add it to the Discovery weekly status report, so at least it will get out at some point, with a trivial amount of work. Then again when the report is published? Presumably the first announcement would be one sentence with a link to a phab task. --KSmith (WMF) (talk) 21:54, 29 March 2016 (UTC)