[pydotorg-www] Vandalism/spam on wiki.python.org: changes to default permissions
techtonik at gmail.com
Sun Jun 1 23:39:55 CEST 2014
On Sun, Jun 1, 2014 at 9:53 PM, Paul Boddie <paul at boddie.org.uk> wrote:
> On Sunday 1. June 2014 19.55.44 anatoly techtonik wrote:
> > On Sun, Jun 1, 2014 at 6:48 AM, Barry Warsaw <barry at python.org> wrote:
> > > On May 31, 2014, at 10:49 PM, M.-A. Lemburg wrote:
> > > >I've added a new user group called NewUsersGroup, which does
> > > >get editing rights, but we'll have to manage this manually and
> > > >new users who want to receive editing rights will have to write
> > > >to this mailing list to be added to the group.
> > >
> > > TBH, we've been doing this on the Mailman wiki for years now. Of
> > > we
> > > don't have as high a contributor rate as the Python wiki, and I wish I
> > > didn't
> > > have to take that step, but it *has* reduced the spam rate to nearly
> > > zero. I'm sure there's been a cost in drive-by contributions, but we do
> > > get occasional requests for write permissions and it's trivial to add
> > > folks to the
> > > "editors" group.
> > Do you have stats about decline in contributions thank to this measure?
> You can get the stats about a decline in "contributions" by just looking at
> how many spam edits were taking place during any given period and how many
> there will be for a corresponding period in future. But for serious
> contributions, of course there are no statistics: the measure has been in
> place for a matter of hours.
You need to count non-reverted edits before and after the certain date.
And am curious about Mailman - it is clear that for w.p.o there is no stats
I'm not entirely sure what measures Confluence had for spam prevention
> the Mailman Wiki made a similar move, but I would imagine that there
> very many given the amount of spam I saw in the site content for that wiki.
Didn't know Mailman chose proprietary software for the wiki.
> From what I can see, this didn't stop committed contributors from improving
> that site.
Being alone contributor is not fun either.
> Is the competence debt in MoinMoin is so big nowadays that there is no
> > Moderation Queue plugin for new users?
> There are people working on MoinMoin, so I suggest you ask in their IRC
> channel: #moin on Freenode, if I'm not mistaken. Alternatively, you could
> on the Moin site for extensions and find at least two for moderation, one
> which I wrote a while back and is actually in use on a site for a project
> I believe you have dealt with in the past (and maybe still have some
> with). It's entirely possible that you have no idea they're using my
> for moderation at all and yet have edited that site.
I found only one - http://moinmo.in/ActionMarket/ApproveChanges - it looks
good except that Moderation Queue will be more useful if it is global and
visible by anyone. More people will be aware of it and it could be
> But here's the disagreement: some people think that the barrier to editing
> public Internet sites should be low and yet magically repel defacement;
> think that getting serious contributors to demonstrate their commitment to
> making quality edits isn't that hard and that they actually won't mind
> so (because they are, after all, committed).
> Experience shows that magically repelling defacement of Internet sites,
> somewhat possible with some interesting measures that could be implemented
> Moin, is a bit like announcing a generously catered party for one's closest
> friends in the most public and open way possible, not assigning some rather
> "persuasive" people to the venue entrance, and then somehow avoiding random
> vandals and hooligans from inviting themselves in and trashing the place.
The cost of defacement should be higher than a cost of revert. If defacement
edits are not indexed, there is no motivation to write bots for it.
> Meanwhile, some of us have better things to do with our time than to "muck
> out" public Internet resources so that lazy people and vandals can use
> them at
> their own convenience.
I still think that the problem with wasted time is a software problem.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the pydotorg-www