[WikiEN-l] A six-day roll-back poll?

[Jussi-Ville Heiskanen]

On 1/4/08, phoebe ayers <phoebe.wiki at gmail.com> wrote:
> On Jan 3, 2008 7:13 PM, doc <doc.wikipedia at ntlworld.com> wrote:
> > Old hands will remember the perennial proposal to grant non-admins
> > rollback facilities. We polled on this for 6 months in 2006, 500 people
> > voiced an opinion and we got no consensus.
> >
> > Well, it's back
> > http://en.wikipedia.org/wiki/Wikipedia:Non_administrator_rollback
>
> ZOMG drama indeed.... but as to the substance of this proposal, I
> thought that particular perennial proposal more or less went away when
> the "undo" button was instituted on diffs? Hmm, apparently not.
>
> As far as I can tell, the differences between rollback and undo are
> that a) rollback shows up in the page history and user contribution
> histories as well as diffs; b)  rollback doesn't require you to hit
> "save" a second time; c) rollback is easier to make mistakes with
> because you can undo more than one diff at a time (i.e. everything by
> an author). Am I missing anything?

Well, it is not clear to me that rolling back only one diff at a time causes
less mistakes.

If you look at the article [[Confucius]] on the english wikipedia, there
is a long patch there where various anon editors made minor
punctuation edits and linking changes etc. which it was very hard
to see as vandalism except in aggregation. The whole thing there
was made a lot hairier by some script reversion chaps who would
only revert a single edit, but leaving the previous subtle miss-edit
by the same IP unchanged. And in the end the history of the page
became such a morass of IPs doing multiple subtle vandalism edits
in a row, and a script riding vandalism hunter only referting the very
last one of those, I in the end saw that the simplest way to correct
the whole mess was to simply revert back to my own version of
37 revisions ago. (I did look through each edit to note that none
of the edits, by either anons, or the single edit reversion scripts
added anything of even arguable value to the article).

--
Jussi-Ville Heiskanen, ~ [[User:Cimon Avaro]]