Let's lay out some solid tag monitoring rules

[ArcticEcho]

In v2, I attempted to guess and "monitor" the tags a user was reviewing, and then trigger an event when they finished reviewing said tag(s). This works™ to a degree, but fails when a user doesn't actually use the tag filter whilst reviewing (I could try to fix this though).

Anyway, as I'm now working on v3 I'd like to clear up a few things about this whole idea of tracking tags, namely:

• What is the end goal here? Are we trying to get the tags a user has completed, or just simply the ones they're currently reviewing?
• Should the analysis take place real-time (nicer for the end user), or at the end of each day (more accurate)?
• And finally, do we want to (attempt to) monitor all tags (3), or just 1?

[gunr2171]

(Notices this issue is on the library, not a bot).

Well, the end developer should have the choice on a per-user basis to enable tag tracking, if we do that. I'm thinking it defaults to "yes, track me", and if the person responds "no" to a "are you done?", the bot would say "Sorry, my mistake. If you would like me to no track your tags, reply with ...".

I don't know how much of that responsibility is on this library or the bot code.

The end goal is to know which tags we have completed. Originally, we would manually remove those tags once enough people completed them from the SEDE listing, so we wouldn't hit them twice. I don't think that's a feasible solution anymore. If we can use SEDE without needing to manually clear tags, I don't know of much of a reason to record tag completion.

If we do record tags, I would vote for "end of each day", and maybe ask the person at the start of the next day.

We should get @rschrieken on this: see if he knows if we really need that info anymore.

[ArcticEcho]

The end goal is to know which tags we have completed.

"we", as in the room collectively, or on a per-user basis?

[gunr2171]

Both? The point was to see if "enough" people finished a tag so that we could cross it off the list for the room. I don't know if we still need to do that.

[ArcticEcho]

Right, I see.

[rschrieken]

We keep that tag completion data to remove/suppress tags from the SEDE result. So it doesn't need to be per user as the next 5 tags command is also not per user. As such is a once per day round-up fine.

If I get rid of the suppression of tags in the SEDE query Closey can keep track of the tags to be suppressed based on completion data from the library. SEDE returns ALL reviewable tags worth their counts, Closey removes the tags that are completed and returns the next couple of tags that needs focus. That makes that the SEDE query can be stable and Closey can re-enable tags automatically after a year?

[ArcticEcho]

Sounds like a good plan.

[ArcticEcho]

So when do we consider a tag to be completed?

[rschrieken]

When 5 or more users report There are no more reviews in tag Haskell
On a single day / event