Post
Topic
Board Meta
Merits 1 from 1 user
Re: A proposal to "hush" other users
by
PowerGlove
on 25/07/2022, 15:49:51 UTC
⭐ Merited by ETFbitcoin (1)
3. How exactly do we get P number?

@ETFbitcoin: For a new activity window, each user's "P" starts at zero. Then "P" gets incremented every time one of their posts is rendered/viewed by a logged-in user, you just have to be careful not to count views more than once per logged-in user.

It's rather simple, but i expect P number for post on first page would be rather high. Some member might only read the thread, then going to latest page.

That's true, accounts that tend to post early in a thread would presumably have a higher P value (assuming that posts on later pages are read less).

I'm sure you already understand this, but (for the sake of clarity) it's important to remember that P is not per-post but per-user (i.e. if any one of your posts has been hushed, it's as if all of them have been, but T and P only get incremented once).

So, if you didn't hush any of another user's posts (but did read at least one of them) in a given activity window then their T gets left alone and their P gets incremented.

If you did hush any of another user's posts in a given activity window then both their T and P get incremented.

Once you've hushed a user, you won't see their posts for the rest of the activity window (see OP edit) and so won't be able to hush them multiple times (i.e. more than once per activity window).

4. With current formula, H value is likely to be unstable. IMO using scaling such as log scale is needed.

@ETFbitcoin: Yup, I just put in a quick non-linear adjustment on top of a ratio without thinking about it too deeply. If you mean unstable as in "sensitive" or "changes non-smoothly" then that's by design.

It's by design? Ok, but i find it's not useful when P value is low.

I see your point, but if P is very low then very few (logged-in) users are reading any of your posts. I don't think that's likely to happen while being in a signature campaign and posting regularly.