Track usage of off topic names to better distribute name usage.

[scragly]

Currently off topic names are used randomly, however randomness is pretty far from evenness. It would be nice to make use of names more evenly, to avoid recently used names being used multiple times.

One method would be to have a column tracking usage in the current iteration, a bool of True/False. used=False would be the default for new names. The pool of available names to pick from will be all tags with used=False. When a names it picked, it'll be given used=True. Once no names are returned for used=False, we just update the used column for all to be set to False again and start over again.

[ikuyarihS]

Should we instead store the number of times tags have been used and use that as the weights ( max - current ) so unused tags have more chance to be used? That way we can have a nice chart of tags and its usage as well.

[scragly]

It's possible the original issue description wasn't clear as I for some reason was saying "tags" instead of "names", which is my bad, and partially because I was talking about tags at the same time as writing this up lol.

Since this is for offtopic names, there's no nice chart to be had, since usage would be distributed evenly.

[ikuyarihS]

Ah I see, thank you for the clarification. If we only want every name to be used evenly then yes I agree, a column of True / False will be much lighter, and can easily guarantee every name will be used until reset.

[ks129]

Can I work on this?