** Clean Up proposal - Removing old, never active accounts **
For purpose of internal discussion and to gauge people's opinions, I am opening a discussion topic on this subject -- here, in Public.
I had very recently brought this up internally, in a Staff only area (essential for sensitive discussions, and common in all internet sites).
The image included in this post is a snapshot of my post there - - which looks much nicer than it could ever be here (until we get better, enhanced text format supported).
This idea was motivated by my belief this deserves attention, and by a very recent post by Paul Sutton -- asking what the meaning of 'active'users is on the server stats displayed at the front doors.
My post follows, in Text form. Look forward to your comments, suggestions and participation. Thank you.
---------------------------------------------
A pet peeve of mine.
I would have done it myself. Proposing here for discussion and possible action.
* that long term, never active accounts represent a Zero value added to instance.
* that such accounts can be dormant ones for video sharing people, as we had before
* that high number of users doesn’t really matter, this is my personal opinion.
I propose that :sparkler:
* inactive accounts, zero posts ever, older than a certain threshold, be eliminated.
A report from the admin panel could provide some stats on how many those would be
The threshold I would suggest – one year and older, inactive, Nuke applied.
Thank you.
------
Screenshots included :
1. Forum post snapshot.
2. The content of this post, as seen in a properly full formatted way. Markdown enhanced.
The counter points to this:
1) We already seperate out stats between active users and registered. So by removing old accounts it would make the registered number useless as we would effectively be reducing it to the active count. Aside from perhaps being dishonest to our users it prevents us from actually being able to know the ratio of active users to registered users, as the registered user count doesnt represent anything meaningful.
2) Users who register with no activity represent 0 value, but also represent 0 negative sides. They do nothing to decrease the quality of the instance
3) users may wish to have an account for read-only purposes. There is no harm in that
4) we can not truely delete accounts, it looks more like a suspension. This means users can not re-register later without them emailing an admin. This adds extra work load on admins, hassle for potential users, and with no gain.
5) there is no easy way to suspend accounts on the criteria you suggests. Someone (me?) would have to invest time and energy to write a script to do it.
6) many users want a server with expiernce and a history. The registered user count reflects this. For that reason many users pick their instance specifically because they have a large registered user count. so we would be doing ourselves harm
It seems to me there are quite a few down sides to doing this and not a single positive effect I can think of... What, exactly, is to be gained by doing this?
Fair points there, I was just thinking about how people may see the number of active uses vs registered users.
As an alternative we can come up with a text explanation to counter any arguments people may have for that.
I believe that a reported number of users who have :
* at least bothered to login once, in the past 12 months.
* posted a single Byte in the instance.
...is not in any way distorted.
Mastodon network user count is approaching 4 million accounts, as seen in the Stats posts in federated timeline.
Well -- how many of those are similar Zombie accounts?
How many are multiple accounts held by one single active user for some reason ? (I have about 12 atm, and increasing. With reasons)
I don't want to spend much time arguing, my logic is on the OP above, and I have other projects to spend time on -- more ideas deserving attention and effort than time, which is precious.
So carry on, let's hear the people's voices.
@design_RG @freemo
I agree here, worth discussing anyway.
I am more thinking that the critics I come across will pick up on anything,
But if you consider that Facebook has what is it now 3 billion users, how many of those users are zombie, duplicate or robot accounts.
> But if you consider that Facebook has what is it now 3 billion users, how many of those users are zombie, duplicate or robot accounts.
Lots of them -- years ago it was easy and quick to get a completely Bogus new account up and running. It's harder now, as these have been abused badly by state agents on campaigns against other countries.
See also, 2016 US election, 2016 Brexit referendum, and likely... 2020 US election, again.
@freemo
Seems this solution would be the best way to satisfy your concerns without having the downsides then, since you mostly just want better reporting of dormant but once-active accounts:
@freemo @design_RG
Maybe I am so used to people being cynical about anything other than the mainstream stuff, I just seen the registered / active users stats as another attack vector for them.
Maybe leave it as it is for now. I am making a blog post later in the week to cover alternatives to mainstream social media.
Already made a post to link to explanations for the fediverse, activity pub etc.
In what way would it be an attack vector? Do you feel adding the third statistic type I mentioned would mediate that attack?
@freemo @design_RG
What I mean is people are just cynical and see what we have here as "esoteric." for one such word.
They already say no one uses it, being no one they know. So I was saying here that if they see users vs active users they will use that to justify their argument.
Given I have, over the last few days had some really interesting discussions on the fediverse, maybe it is time to just ignore them and just promote what we have here the best way I can.
So if i understand you correctly people think the service isnt used by enough people to be useful.
If that is your main concern wouldnt eliminating registered accounts have the opposite effect? It would reduce the number of apparent users making the service look more obscure not less, no?
@freemo @design_RG Yeah, I think it is time to ignore the cynics and not really worry about what they think unless they have actually been here and used the fediverse in some way.
I generally tend to agree, but I think it is important we at least listen to the cynics. We just shouldnt jump to make changes over every criticism either.
@freemo @design_RG Good point, I think this thread highlighted an issue with the delete and redraft feature, which is a 'good' thing in that it can be looked at by the developers.
I actually like being able to go back in and make corrections though.
@freemo
> Ideally we should be able to edit posts without "deleting and redrafting"
It can be done, it's in place in both Corporate networks (FB for example) and on non-corporate ones, Friendica has it as previously mentioned.
Corporate Twitter is the model for Eugen's mastodon project, so a lot of what we see here is based on the original. He aded many extras and high value -- it's much better generally.
BUT -- never an edit we shall see, I bet. Not on Mastodon anyway.
Yea its absolutely doable in theory. Eugen was explicitly against it. So he did this on purpose, though I disagree.
@freemo
> Eugen was explicitly against it. So he did this on purpose, though I disagree.
He has a LOT of control in this project, and sadly that has BAD implications sometimes. It's best for Fediverse if other platforms grow.
Pleroma, Friendica, others all have value and LESS controlling lead developers.
Also -- NOT modelled on the twitter model.
@admin
AAAAHHH ---- LOL. thank you!!
Yes, that it is.
I was thinking about along the lines of the classic BOFH -- which I always loved reading! 😝
So his logic is, its better, but its too much work?
Seems weird then he would close it rather than just call for a PR.
Hey @Gargron just out of curiosity if someone submitted a high quality PR implementing edit in place with a change log for you, would you accept it?
@freemo @admin @zleap @design_RG There are a few performance optimizations possible because toots are immutable. If they can change, you lose them.
What if a PR could be submited and demonstrated to have minimal impact on performance, would that be acceptable to you or is a no-go regardless?
Benevolent Dictator For Life
@zleap @freemo
20200607_101431.jpg