Hmmm... here's a simple idea. If ten (or some other
arbitrary number) of logged in users have read the
article since the last update, then it is stable. We
could also require that articles have 1000 pageviews
or something as well. Simple, no hassle, in the wiki
spirit. :)
Chuck
=====
Homepage: http://amuzulo.babil.komputilo.org/
Taglibro: http://amuzulo.livejournal.com/
====
Enciklopedio en Esperanto, 2250+ artikoloj!
http://eo.wikipedia.com/
__________________________________________________________________
Gesendet von Yahoo! Mail - http://mail.yahoo.de
Yahoo! präsentiert als offizieller Sponsor das Fußball-Highlight des
Jahres: - http://www.FIFAworldcup.com
Hi maveric and all
Just wanted to make it clear that I was not criticising your comments re Max
Weisman in any way! I'm quite sorry if you got that impression. I simply
started with what you said, but then went off on a tangent completely.
I hoped I had made that distinction clear in my original post, and if not, let
me make it clear now. I thought you had every right to question the validity of
the article, although if JHK says he has some merit then I certainly won't
argue.
Cheers
Manning
>I don't know anything about Max and so I don't know whether he
>deserves an article, but the underlying question is important and
>interesting: should anything that somebody bothered to write up be
>accepted into Wikipedia, assuming it is NPOV?
>
>I say no. The item has to be of relevance in some way...
I was inclined to agree, and I definitely think there are topics
that don't belong in an encyclopedia. But if we do implement some
approval system, especially a multiple-valued one such as I suggested
earlier today, I like Jimbo's suggestion of making relvance one of
the approval criteria. If that's done, then there's really little
reason not to allow reasonable articles on anything and anyone, and
simply consider them potential articles (which might get promoted
to "relevant" status if, say, Jimbo's mom gets elected to congress or
something).
One possible downside is that many of them might become orphans,
making database maintenance harder, because even if we allow such
articles, we should not necessarily allow links to them from relevant
articles.
0
Hi Anthere,
> While trying to bring new people to internat.
> Wikipedia, I have been asked what was the validation
> process in terms of reliability for the article...
>
> I don't exactly know where to find an approach to this
> question (there's much to be said) and I don't know
> either where to ask the question :-)
Here's the short answer. Wikipedia works on a similar principle to open source software development. Anyone can pop in and write or change an article and it is immediately posted. Thus, there are no technological mechanisms to prevent people from posting nonsense. However, most Wikipedia participants are intelligent people dedicated to making a high quality encyclopedia. There is a page that shows all the recent changes made in the database, and at any given time there are many pairs of eyes checking those additions. When nonsense or vandalism is found it is removed and/or edited immediately.
This sounds like an unreliable system to many people, but it is surprising how well it works. Ifyour friend is skeptical, have him look around Wikipedia to get a feel for the quality of the articles.
-- Stephen Gilbert
________________________________
Wikipedia: The Free Encyclopedia
http://www.wikipedia.com
--
_______________________________________________
Download the free Opera browser at http://www.opera.com/
Powered by Outblaze
On Tuesday 28 May 2002 05:02 pm, Lee wrote:
> One possible downside is that many of them might become orphans,
> making database maintenance harder, because even if we allow such
> articles, we should not necessarily allow links to them from relevant
> articles.
I can think of another downside - new contributors and visitors become
increasingly clueless as to what the project is about and the difference
between personal posts about their friends and family and actual encyclopedia
content. That is, unless we somehow hide the "non-qualified" articles from
search engines or visitors to the site -- but then what use would they be?
Who is going to be the one to tell new contributors <not> to do things the
natural wiki way and link their personal biographies to other pages? New
contributors generally don't read our policy pages before contributing and it
is already a struggle to get them to up-to speed on naming conventions and
NPOV -- let alone having to explain the difference between the two
incompatible, yet, side-by-side types of articles. What is described is
another project altogether (Wikicities anyone?).
For example: Under this setup, if I were a new contributor and just wrote a
bio on myself after seeing many other personal bios in RecentChanges, It
would seem pretty logical for me to place [[Daniel Lee Mayer]] on [[1975]]
(my birthyear) and on [[December 31]] (my birthday). It would also seem
natural to add my name to [[Biologists]] since I have a degree in
biology.....
Minden' the shop is difficult enough as it is. I for one don't want to have
to constantly check to make sure links actually link to encyclopedia content
and not personal bios. But then, that's just me.
Cheers!
--maveric149
OK, Jimbo's DOS-resistant password-mailing feature wasn't too
difficult, so I did it. Unfortunately I also accidentally deleted the
accounts of the 10 or so people who had previously logged in, so you
folks will have to log in again. Sorry about that.
Looks pretty robust now, and no more unencrypted passwords anywhere.
0
On Tuesday 28 May 2002 05:02 pm, Jimbo wrote:
> One of the things that a future "qualification process" might do is
> flag or rate articles based on both quality and "importance", so that
> reproductions in media with space constraints might find some guidance
> as to what to leave out.
What an great idea! That way we can fully integrate many of the ideas of
"beta/stable" and eventaully be able to create a printed Wikipedia on the
fly. Excellent and most useful!
--maveric149
> The slashdot software does something similar to this.
> Here's something to think about...
> The point of these steps is to prevent a certain kind of
> denial-of-service attack.
Thanks for the idea. I just checked in the new code for
encrypted passwords, but I'll save your message and consider
adding that as well. As I was writing the code, that particular
DOS even occurred to me, but I didn't think it was likely
enough to justify the extra code and a solution didn't leap
immediately to mind--I'm glad to know there is one.
> ...this email should identify the ip number of the person
> clicking on 'send a new password'.
Another thing to consider. Thanks.
0
I and others have expressed dismay at the fact that the wikipedia
database contains plain text user passwords. This would be very
easy to fix in the new codebase, but there are several options on how
to handle it, some of which are feature changes that ought to be
discussed by the list at large.
The intended functionality (and what should be the current
functionality of http://www.piclab.com/newwiki/wiki.phtml) is this:
new users are created by looking up the name; if the name doesn't
already exist, the user is created with the password specified. The
user can then log in with that name and password at any time. Logged
in status is maintained thoughout a session. If the user has
the "remember my password" option turned on, he is logged in even
after closing the browser and restarting later. The login form has a
button for those who have forgotten their password. If they have
entered an email address for their account, their password can be
sent to that address (in plain text). This "email password" function
is tricky to do without plain text passwords in the database. A user
can change his own password at anytime.
As I see it there are three options:
(1) Eliminate the "email my password" feature. This would require
people who forgot their passwords either to create new user accounts,
or aking an administrator to clear their password (which will be made
an option available to administrators). Passwords in the database
will then be encrypted.
(2) Change the feature to "email me a new password". Passwords will
be encrypted, and when the user forgets his, the system will create a
new random password and email _that_. The user will then log and
change his password (hopefully to something he can remember).
Plaintext passwords won't exist anywhwere with options 1 or 2--this
is the most secure.
(3) Keep the full existing functionality by creating a "shadow"
password table with plain text passwords that not even administrators
can access (but that a few folks like Jimbo can). This complicates
administration, and also might add some latency to functions like
creating accounts and changing passwords, but probably not to the
basic login. This is the simplest from the user's point of view.
So let me ask the group: is the functionality of sending a plaintext
password over email useful enough to justify some extra
administration? Is security that important? (There's nothing
sensitive in the user table, but people do tend to re-use passwords
for several things, so having plain text passwords lying around is
probably not a good idea). Which option do you think is best?
0