librelist archives

« back to archive

rstat.us policy

rstat.us policy

From:
Carol Nichols
Date:
2012-08-18 @ 13:55
I decided to send this here rather than filing an issue, because this is
more a policy question about how we want to run our node at rstat.us, and
the long term answer will be something that each node operator will have to
decide for themselves.

I got a request from someone who would like a username that has been signed
up on rstat.us but hasn't ever posted any updates and the User hasn't been
updated in over a year.

My inclination is to honor this request because of the lack of statuses or
User#updated_at in over a year. I'd like to make this policy public though,
so that it's clear this may happen if you don't use your account for a year.

I'm also inclined to leave well enough alone unless we get a request like
this; i.e. I'm not going to purge inactive accounts proactively.

How does that sound? I'm up for amendments to this policy.

-Carol

Re: [rstatus] rstat.us policy

From:
Steve Klabnik
Date:
2012-08-18 @ 14:03
Sounds fine to me.

Re: [rstatus] rstat.us policy

From:
Colin Dean
Date:
2012-08-18 @ 16:17
Does the existing account have any contact information associated with 
it? If so, it would probably be prudent to attempt to contact the 
owner. If the owner is OK with it, or there is no way to contact the 
owner, I say honor the request.

Moving forward, howsabout a policy like this:

1.) At 6 months since the last post, an email is sent to the user 
asking them to interact. They at least need to login -- some people may 
have accounts and read, but never post.

2.) At 12 months since the last post, an email is sent telling the user 
that if they don't login, their account will be removed unless they 
take action and they will receive no further notification from us. It 
will also provide the standard "click here if you forgot your password" 
and "reply if you want to save your account but are unable to login or 
reclaim yourself"

3.) At 12 months + Math.log((32*user.posts.size)+64, 16)*32 hours 
without activity, purge the account. This gives users who at least at 
one time were active a little more time to save their account. Users 
with 0 posts will be deleted in 48 hours. Users with 100 posts will be 
deleted in ~4 days (~94 hours). Pretty much anyone else would be gone 
within a week. Or just expire them at 12 months + 48 hours if you hate 
variability.

I do realize this would take some coding to implement and require some 
kind of delayed job or cron job, so perhaps it's worth filing an issue.

Or, we just follow the human-intervention-required rule of at 12 months 
past the last activity, the username will be forfeited if another user 
requests it. If we choose this policy, we should make an effort to 
notify folks who may be affected, if not all users.

On Sat 18 Aug 2012 10:03:09 AM EDT, Steve Klabnik wrote:
> Sounds fine to me.

Re: [rstatus] rstat.us policy

From:
Steve Klabnik
Date:
2012-08-18 @ 16:27
Also, if we ask the person, of COURSE they're gonna say 'oh, I'll still use it.'

Re: [rstatus] rstat.us policy

From:
Colin Dean
Date:
2012-08-18 @ 16:59
On 08/18/2012 12:27 PM, Steve Klabnik wrote:
> Also, if we ask the person, of COURSE they're gonna say 'oh, I'll still use it.'
> 

They may. I know I'd be pretty unhappy with a service if I didn't use it
for a year and my account with a unique username was deleted without
notification. Some people are into namespace protection, in the event a
service gets popular or for future use.

If we don't want to consider those people, then consider my exception
handled. I only raise it because I think it's not right to take
someone's username from them without giving them a chance to object or
at least notifying them.

Re: [rstatus] rstat.us policy

From:
Steve Klabnik
Date:
2012-08-18 @ 17:00
> They may. I know I'd be pretty unhappy with a service if I didn't use it
> for a year and my account with a unique username was deleted without
> notification. Some people are into namespace protection, in the event a
> service gets popular or for future use.

I don't want to encourage squatting. This also is less relevant in a
federated system.

Re: [rstatus] rstat.us policy

From:
Dave Wilkinson
Date:
2012-08-18 @ 21:33
Who needs to do all of that work? Who needs a policy?

The safest option is to just not remove any user accounts. They may use the
account to log in and just read what others say. You don't know.

You want your own name? Create your own node. Your name is your webfinger
identity, so you can have whatever username you wish and just tie it to
that. This identity should be separate from something like rstat.us anyway.

I'm not sure, but I think 95% of people have no idea how to do that. :(

On Sat, Aug 18, 2012 at 1:00 PM, Steve Klabnik <steve@steveklabnik.com>wrote:

> > They may. I know I'd be pretty unhappy with a service if I didn't use it
> > for a year and my account with a unique username was deleted without
> > notification. Some people are into namespace protection, in the event a
> > service gets popular or for future use.
>
> I don't want to encourage squatting. This also is less relevant in a
> federated system.
>

Re: [rstatus] rstat.us policy

From:
Carol Nichols
Date:
2012-08-19 @ 14:50
> You want your own name? Create your own node.

That's all well and good for the long term, but seeing as i'm not even sure
running your own node with the rstat.us codebase works right now, what do I
do about *this* person right now?
-Carol


On Sat, Aug 18, 2012 at 5:33 PM, Dave Wilkinson <wilkie@xomb.org> wrote:

> Who needs to do all of that work? Who needs a policy?
>
> The safest option is to just not remove any user accounts. They may use
> the account to log in and just read what others say. You don't know.
>
> You want your own name? Create your own node. Your name is your webfinger
> identity, so you can have whatever username you wish and just tie it to
> that. This identity should be separate from something like rstat.usanyway.
>
> I'm not sure, but I think 95% of people have no idea how to do that. :(
>
>
> On Sat, Aug 18, 2012 at 1:00 PM, Steve Klabnik <steve@steveklabnik.com>wrote:
>
>> > They may. I know I'd be pretty unhappy with a service if I didn't use it
>> > for a year and my account with a unique username was deleted without
>> > notification. Some people are into namespace protection, in the event a
>> > service gets popular or for future use.
>>
>> I don't want to encourage squatting. This also is less relevant in a
>> federated system.
>>
>
>

Re: [rstatus] rstat.us policy

From:
Carol Nichols
Date:
2012-08-21 @ 14:36
Ok, since wilkie is actively working on making running one's own node
better and functional, I'm not going to purge any accounts from rstat.us.

Wdyt?

-Carol


On Sun, Aug 19, 2012 at 10:50 AM, Carol Nichols <carol.nichols@gmail.com>wrote:

> > You want your own name? Create your own node.
>
> That's all well and good for the long term, but seeing as i'm not even
> sure running your own node with the rstat.us codebase works right now,
> what do I do about *this* person right now?
> -Carol
>
>
>
> On Sat, Aug 18, 2012 at 5:33 PM, Dave Wilkinson <wilkie@xomb.org> wrote:
>
>> Who needs to do all of that work? Who needs a policy?
>>
>> The safest option is to just not remove any user accounts. They may use
>> the account to log in and just read what others say. You don't know.
>>
>> You want your own name? Create your own node. Your name is your webfinger
>> identity, so you can have whatever username you wish and just tie it to
>> that. This identity should be separate from something like rstat.usanyway.
>>
>> I'm not sure, but I think 95% of people have no idea how to do that. :(
>>
>>
>> On Sat, Aug 18, 2012 at 1:00 PM, Steve Klabnik <steve@steveklabnik.com>wrote:
>>
>>> > They may. I know I'd be pretty unhappy with a service if I didn't use
>>> it
>>> > for a year and my account with a unique username was deleted without
>>> > notification. Some people are into namespace protection, in the event a
>>> > service gets popular or for future use.
>>>
>>> I don't want to encourage squatting. This also is less relevant in a
>>> federated system.
>>>
>>
>>
>

Re: [rstatus] rstat.us policy

From:
Dave Wilkinson
Date:
2012-08-21 @ 18:55
That's the safest solution. Twitter's policy is based upon login access
times. If you don't log in in 6 months, they say they /can/ remove your
account and /plan/ on automating it.
https://support.twitter.com/articles/15362 (It is ambiguous whether or not
you have to post an update as well)

We don't have a policy written, so we don't really have a good way of
purging any accounts that exist now. I don't know. :(

On Tue, Aug 21, 2012 at 10:36 AM, Carol Nichols <carol.nichols@gmail.com>wrote:

> Ok, since wilkie is actively working on making running one's own node
> better and functional, I'm not going to purge any accounts from rstat.us.
>
> Wdyt?
>
> -Carol
>
>
>
> On Sun, Aug 19, 2012 at 10:50 AM, Carol Nichols <carol.nichols@gmail.com>wrote:
>
>> > You want your own name? Create your own node.
>>
>> That's all well and good for the long term, but seeing as i'm not even
>> sure running your own node with the rstat.us codebase works right now,
>> what do I do about *this* person right now?
>> -Carol
>>
>>
>>
>> On Sat, Aug 18, 2012 at 5:33 PM, Dave Wilkinson <wilkie@xomb.org> wrote:
>>
>>> Who needs to do all of that work? Who needs a policy?
>>>
>>> The safest option is to just not remove any user accounts. They may use
>>> the account to log in and just read what others say. You don't know.
>>>
>>> You want your own name? Create your own node. Your name is your
>>> webfinger identity, so you can have whatever username you wish and just tie
>>> it to that. This identity should be separate from something like
>>> rstat.us anyway.
>>>
>>> I'm not sure, but I think 95% of people have no idea how to do that. :(
>>>
>>>
>>> On Sat, Aug 18, 2012 at 1:00 PM, Steve Klabnik <steve@steveklabnik.com>wrote:
>>>
>>>> > They may. I know I'd be pretty unhappy with a service if I didn't use
>>>> it
>>>> > for a year and my account with a unique username was deleted without
>>>> > notification. Some people are into namespace protection, in the event
>>>> a
>>>> > service gets popular or for future use.
>>>>
>>>> I don't want to encourage squatting. This also is less relevant in a
>>>> federated system.
>>>>
>>>
>>>
>>
>

Re: [rstatus] rstat.us policy

From:
Mark Kremer
Date:
2012-08-22 @ 07:10
Sounds good to me, being able run one's own node is one of the core 
ideas of the project is it not? :)

On 08/21/2012 04:36 PM, Carol Nichols wrote:
> Ok, since wilkie is actively working on making running one's own node 
> better and functional, I'm not going to purge any accounts from 
> rstat.us <http://rstat.us>.
>
> Wdyt?
>
> -Carol

Re: [rstatus] rstat.us policy

From:
Steve Klabnik
Date:
2012-08-18 @ 16:27
This is like waaay too much ceremony. Do you really want to calculate
all of that?

If you don't want to lose an account, host it on your own domain. Then
it can be as dormant as you like. ;)