#musicbrainz

there's got to be a better approach

Setting a minimum week of being open delays bad edits from being voted out.

Setting some higher # of votes needed or setting some minimum age both still can be easily gamed.

BrianFreud: i'm not denying your knowledge (and frustration) with the existing system -- i'm just saying -- give it some more thought

Oh, I agree, there is a possibility for a misperception.

I just think that giving a 'no' vote is the wrong way to do what you're wanting to do

But perceptions can be altered. Text on the edit screen "Note: This edit type will have an immediate no voted by the system to prevent system abuse" does that

haha

we have plenty of notes but I'm not sure people read them

ok, if you don't read the guidelines, and you don't read the notes... aren't you exactly the person we're trying to prevent from mucking things up?

i'd rather channel them to the guidelines, i wish there were some way to automate that

what's wrong with simply making it so they stay open for a week unless there's enough unanimous 'no' votes? why does it have to be implemented by giving everyone a 'no' vote?

that pre-supposes we have an easier-to-read set of guidelines

Coming in in the middle of the conversation, I agree with nikki here.

"No" vote is a "No", no amount of noting will make it less.

nikki: you didn't say that before - you said only "require it open for a week"

the current set is to self-referential

*too

And it's even worse if "No" means different things in different edits.

BrianFreud: well, you can add all sorts of various checks, the idea is the same

if it was minimum time of a week no matter the yes count, but could still be voted down in a shorter period, that seems fine to me

could make it so that it only keeps them open if the editor joined in march or october if you fancied :P

I just really think the minimum age to vote or higher yes count required ideas are far too easy to game

Morning. :'(

hey faux

moin moin faux faux

:) /me reads BrianFreud craaaazy ideas in the scrollback, starting a merge with no, ouch. :/

[02:38] <nikki> what's wrong with simply making it so they stay open for a week unless there's enough unanimous 'no' votes? <-- that's the best solution I've seen suggested so far

From my experience of not voting on anything, unanononnonoimus no is incredibly rude? Why even allow it?

Unless you want an emergency way to hide bad data, or something.. if only we had some trusty auto-editors.

Huh?

8)

well, most people cancel once they realize why people are voting no makes sense

but if you leave it open a week without regards to the no count, it's easier to game

What's wrong with the current system?

destructive edits can be pushed through before anyone can see them simply by creating 4 accounts

How often does that happen?

7 times that I've seen this year, and I think we've been lucky so far - a 4chan group could easily make chaos if they decided to

They could make chaos with add edits as well.

Yes, but add edits create new data, and can be cleaned out simply by our identifying the editors and searching their edits.

But imagine, say, if someone forced a merge of Mozart into JS Bach.

There's simply too much data - we'd have to revert (and lose any data since that last dump)

to me, it just seems very risky to leave that the db open to serious data mangling or deletion, when all it takes to pull it off is 4 accounts and a spare 10 minutes

I'm like sorta kinda agreeing with srotta agreeing with nikki here. A 'no' is percieved as such. And they are devastating at that lousy desperate pickup-bar where everybody is so drunk that they are seeing triple, it's 5 minutes to closing late sunday morning, she's like making the ugly tree looking beautiful as a magnolia, and still you're getting a no.

.. or so Ive heard

petros: lol, what about I'm busy tonight, but here's my number, call me in a week? :D

BrianFreud: I agree, but the no vote system gets a no vote from me. 8)

If we make it a week, but allow it to still be voted down, then there's still plenty of time for it to be noticed and voted down. But if it simply stays open a week, it still becomes the yes votes vs the no votes, and

thus can still be gamed by just creating more accounts to vote yes than no votes get cast

Eh?

An edit is created. a day later, drsaunde sees it and votes no. He tells us about it, 2 of us vote no as well.

If it can be voted down by 3 nos in a week, all is good. Data didn't change.

I don't like it, I think the problem, if there is one, should be solved in other ways, not making more complex voting rules.

But if it has to stay open a week no matter what, and the current yes-or-no-whichever-is-higher still applies, then all the prankster has to do is cast a ton of yes votes, to outbalance any nos.

how about minimum edits before you get to vote. Say 50 non-auto?

worst case, if the nos were wrong, we're still at status quo.

So there's a basic rule, an exception to that rule, and an exception to that exception.

Yeah.

8)

[02:58] <srotta> So there's a basic rule, an exception to that rule, and an exception to that exception. <-- sounds a lot like our current DQ vote count policies :P

I'd rather make sure on the backend side that the destructive edits can be reverted, if it's necessary.

well, we have the data dumps, we can always restore from them.

But we'd lose any edits made in the meantime, and the server would have to go offline while it was reverted

I know I'd be rather annoyed if I just had done all the ARs for 10 CDs, and that was all lost in a revert because someone decided to have some fun...

I'd understand, but I'd still not be happy about it

I think we can agree that a restore is a last measure and should be avoided.

yes - that's why I think we do need at least some bare minimum of higher protection on destructive edits than we have now

I think the best way to avoid destructive edits is to "limit" "new" accounts.

how would you do it? If Rapidshare can

BrianFreud: That

he question :)

is the

Damn keyboard

lol, is late, accidentally hit enter :P

If Rapidshare can't block people powercycling a modem, is tracking ips really going to be effective?

But baically I would award voting after a new user had 'proven' himself through viable edits

You can edit away, but voting has to be 'earned'

I'm not against that idea either - I would worry that it would even further discourage overall voting, but then, it might perhaps raise voting by making it seem "special" that you have reached a point where you're being allowed to vote...

question then would be how many edits before we trust someone to vote intelligently, and not to try and game the server?

50 non-edits could be a starting suggestion. I doubt you'd do 200 edits just to have 4 accounts to game with.

'earned' Voting shouldn't be 'special'. Just a mean to discourage vandalism

perhaps lower, like 25? My concern would be that editors can learn from voting semi-badly, but seeing the notes from others. -- exactly. we want to block vandalism, and perhaps block totally uninformed brand new voting, but we don't want to cut off that important training we can do via example either

25 is also fine with me. I'm more curious what as to how the idea would be received by new users.

And the other editors :)

I'm open to all kinds of suggestions and ideas ... like a drunk virgin on prom night

BrianFreud: Yeah, as I said, I'd like to resolve the problem in backend. Restoring backups is not resolving it 8)

:)

argh, someone's being really sloppy and making all sorts of bogus Marley family & foo collabs :(

If I cancel an edit, do I lose all the comments made on that edit?

no

edits and edit notes last pretty much forever, unless the artist or release they're tied to is removed or merged away

natta mb :)

BrianFreud: while I do think we should change the current system, I really do think you're trying to find a solution for problems we don't really have, I still haven't seen any edit pushed through by sock puppets which weren't newly created, and I certainly haven't seen any sock puppetry used to vote down edits

I did ask you for a link to the edit you were talking about, but I still haven't received it :P

nikki: http://musicbrainz.org/show/edit/?editid=8636482

drsaunde: all those accounts were brand new

drsaunde: one suggestion was to make it so new accounts can't vote for a week, but BrianFreud claims he came across one where the accounts weren't brand new

drsaunde: but I'm still waiting for a link to it :)

?lookup artist 2c3a60f7-ebea-4001-888c-fe953ed87212.html

Please enter an MBID

?lookup artist 2c3a60f7-ebea-4001-888c-fe953ed87212

Artist: Aquarhythms (http://www.musicbrainz.org/artist/2c3a60f7-ebea...)

loi

bah dusjetid

nikki: sorry bout that..i had no idea the full story, i though you meant the edit starting the discussion, so i found and posted that

What if new users had to be "sponsored/mentored" by an established editor or autoeditor who would have to vouch for them before they could vote? </brainstorming>

You'd run out of mentors pretty fast.

too many new editors, too few autoeditors

And we have low tolerance for suffering fools ;)

MClemo: BrianFreud is taking care of that problem ;)

And any of those things will cause that there will be even less people voting.

mikemorr: hehe. Very tiresome to vote on this many elections though :)

What if new/unsponsored/etc. editors could only vote on certain (i.e. less potentially destructive) edit types?

Then I (big surprise) like my idea better

To only letting them vote after having done a certain amount of viable edit.

edits

It could be graduated: No voting at all until you've done this and this, then you can vote on the more minor edits for a while, then eventually you can vote on the big edits.

Sounds like an idea.

I still see it more as a problem that could be resolved by making sure destructive edits can be reverted, not by creating elaborate schemes to make the voting/editing system more complex.

would sure be better than the current system. And as it's not touching editing, new users wouldn't complain too much as they tend to edit more than vote

reverting would be cool too ;)

Potential idea, luks will probably wanna shoot me due to the work involved implementing, heh: Destructive edits go through vote as normal. When it's approved by vote, the people who Yes the edit are checked that they're "active" moderators (have some approved edits of their own) before the edit is actually committed. If none do, it goes to a page only AutoEditors can access. Any AutoEditor would then approve or reject the

edit. ... This keeps the current voting system, just adds a "final step" for destructive edits with no active-editor votes.