[WikiEN-l] Things that bother me

Erik Moeller erik_moeller at gmx.de
Sun Jan 19 13:43:00 UTC 2003


Round and round we go ..

> We're not saying "we have 100.000 complete, detailed, perfect,
> bullet-proof articles", we say "we have 100.000 articles". We even link
> directly to our definition of "article".

The major problem was the import of ~30,000 articles, but there is a valid  
point to be made that these should be counted: Many of them *are* built  
upon, either by people using "Random page" to find articles to work on, or  
by people searching for their hometown etc. Then again, there are 10- 
person-towns which are likely never to be edited.

I suggested counting only articles with a minimum number of edits, but  
that wouldn't help us much here, as Rambot updates his own articles  
occasionally to change the wording. Neither would a specific article  
length. What would help us would be flagging particular users as bots and  
excluding them from the count. [No offense to Ram-Man: A lot of work went  
into the bot's design, but it is not equivalent to 30k human-written  
articles.]

But there's the problem that we're going to cross 100k articles in a few  
days, and nobody is going to rewrite the code until then. And after that,  
drastically reducing our article count would seem silly. It's not that the  
counter is *deceptive*, it just doesn't properly reflect the human  
attention that articles have been given.

So here's what we should do: Go ahead, put out the press release when we  
reach 100K. Work on a system to properly handle bots in the article count.  
Then, suggest this revised article count as a way to measure our *human*  
contributions and list it as an additional count in the stats. If we feel  
like it, we can put out another 100K press release when we hit that count,  
and hope that nobody remembers our first one ;-)

> Vera Cruz
> Well, I never had any quarrel with him/her/it. I noticed some strange
> activity on Recent Chages, but that was it.
> On the other hand, I *am* tempted to ban "Vera Cruz" just to reduce the
> amount of incoming mails.
> If it's Lir (prove!), or harming the 'pedia (subtle or not), I say get
> rid of it.

Proof:

1) Vera Cruz appeared immediately after Lir's ban. Her edit behavior is  
virtually identical (mark everything as minor, don't use preview etc.).  
Her homepage style, listing all minor edits, is the same. She has worked  
on exactly the same articles and finished the work she started as Lir.

2) Brion checked the email address and it was Adam [name omitted for privacy reasons] (she has  
reportedly since changed it to another Hotmail account). A. [name omitted for privacy reasons] runs  
this blog: http://qwert.diaryland.com/ where she has written about her  
Wikipedia experiences as Lir. Lir has also posted to this list as "Bridget  
[name omitted for privacy reasons]", a name for which Google has no hits other than these archived  
mails.

> Images
> After DW told me some weeks ago in a rather commanding tone that some
> images I uploaded were too large and to dark (images that I had already
> downsized, filtered and dust-removed), and started to replace them with
> tiny thumbs, I found myself rather displeased with the current image
> handling, both by software and by people. Did anyone notice that lots of
> our images have no source or PD explanation given? That we have
> double/triple images of the same thing/person? That there are "x.jpg",
> "x (small).jpg", "x (large).jpg", etc. and no handling of the different
> sizes?

See wikitech-l discussion - let's go ahead and redesign the image code to  
generate thumbnails on request.

> Consensus
> Great thing. We talk about the different options, argue a little, and
> finally agree on the way to go.
> Except we don't.
> Anyone know a country or big company that is successfully run by
> consensus? I don't.
> Dictatorship or voting. Pick one.

You know I agree. The main problem I see is that we already *are* voting  
(votes for deletion, votes for NPOV, votes for www.wikipedia.org) without  
an ordered process to do so because, thanks to the anarchists who  
constantly decry voting, we are living in cognitive dissonance. This  
ordered process should actually be supported by the anarchists, because we  
could say "before any vote, we have an argument period of X days and try  
to collect all pro/con arguments on a wiki page; everyone who votes is  
encouraged to make a reasonable effort to read and understand these  
arguments". We can also set specific thresholds instead of the current
guessing what the "not quite unanimous consensus" is.

Things will only change when Jimbo decides to let us officially use voting  
in some cases. He said to me that he thinks that voting may be necessary  
in a small number of cases, so perhaps it's possible to convince him. Drop  
him a mail.

Regards,

Erik



More information about the WikiEN-l mailing list