[WikiEN-l] User Michael and ways to deal with him

[Erik Moeller]

Anthere-

> It appears the performance are much better now. At
> least, there are much better for the international
> wikis, which I think have not been moved.

The disabled functionality is of pages that do not scale well, regardless  
of current sever load. Unless the underlying SQL queries are rewritten and  
additional indices added to the respective tables, these pages will  
continue to generate high load. Enabling them without optimizing the code  
and database first would be unwise, as we would quickly find ourselves in  
the molasses we just crawled out of (thanks possibly in larger part to  
Brion's programming hacks than to the new server).

> Tannin presented "Michael case" as an experiment of
> soft banning.
> IP banning and deletion on sight are hard ban tools.
> RainClouding, instant reversion and blanking are soft
> ban tools. I think an experiment has to be done
> completely or not done at all.

The question is: Why do you want to use that strategy? In order to make  
sure that Michael creates no harm, or to prove a point about banning in  
general? If it is the latter, you've lost me: I believe in hard bans and  
think they should continue to be used. I'm willing to agree on a  
compromise here, that is, to use a "soft ban" because it may be more  
effective here.

This is beacuse, if we want to prevent Michael from doing harm, we should  
employ the most effective strategy. It has been argued somewhat  
persuasively that it would not be wise to ban Michael again and again, and  
to continually revert his edits instead. This may be worth trying. But the  
"let's blank, not delete" argument does not fall into this category of  
argument. It is a more philosophical notion about the treatment of vandals  
in general which I do not agree with, at least not in this form.

Michael is a vandal. You may think that some of his pages are worth  
salvaging. Unfortunately, this seems to be more ideology speaking than  
actual experience with Michael's edits. His articles are grossly factually  
inaccurate and the last thing we need are people going around willy nilly  
and restoring pages which they "think" are OK, but which they haven't  
really bothered to check (Michael's articles usually contain a lot of  
wrong titles, dates etc.).

> The fact the pages are going to be deleted anyway is
> "your" opinion.

Michael is a banned user, this is not an opinion, it is a fact. He was  
banned for virtually all violations of our rules that are possible. He was  
never unbanned. We have merely modified the method by which we enforce  
this ban. Please do not use the willingness of sysops to go along with  
this approach to promote a general anti-ban agenda.

>> More respectful? Towards whom?
>> Michael, who has insulted
>> virtually every Wikipedian who has tried to talk to
>> him?

> You know quite well the "toward whom" is not only
> adressed to Michael.

No, I do not know that. Do you see it as disrespectful if articles by a  
known vandal are deleted?

>> I for one find it absolutely acceptable
>> to delete new Michael-
>> pages on sight and encourage sysops to do so -- but
>> only if it is certain
>> that he is really the author.

> only if it is certain, right.
> And only if the edits are bad, right.

No on the second point. We have already determined that Michael's edits  
are not worth the trouble. Will you go to the search engines and check  
every single date in a discography, every little factoid about a band's  
history? If you're not willing to do this work, you should not talk about  
"bad edits". If you are, make a promise now, and I will hold you to it.  
For every Michael edit and every claim.

> I am not worried of the rift. I am worried of errors.

Everything can be reverted. This will not be necessary in the case of a  
known, banned user, however.

> I also think you are on the verge of saying I will be
> responsible of subsequent errors.

See above. The last thing we need are more unchecked pages by Michael. It  
will take years to go through the ones he has already created and check  
them for errors.

Regards,

Erik