Here's what happened: I asked Jimbo or Jason to create dumps
of the current database so that I could load it into the new
server for testing the transition. I expected to have them
copied directly to the new server as Jim had done before (and
to which only the Bomis folks, me, and Axel have access at the
moment), but Jason put them in the upload directory instead.
I have since moved them to the new server already, so they can
be deleted--but I don't have access to the old server, so Jason
will have to do that.
I raised the security issue of plaintext passwords in the
database a few months ago; it was generally agreed that it was
not a good thing (any sysop can already read the database,
passwords and all), but that changing it in the existing code
was a pain. The new code encrypts passwords in the database,
and has a "mail me a new password" function.
TBC, I removed your password for you. I suspect the problem
with your passphrase may have been the single quote character.
Try a long passphrase without it.
0
On Saturday 29 June 2002 12:01 pm, Jan Hidders wrote:
> Apologies for being a bit too simplistic here. I fully agree with you and
> Axel that there should not be an exact number for this and should be judged
> separately each time. On the other hand, it wouldn't hurt if we would have
> some policy statement that says that these page are not supposed to be
> exhaustive plus a rough estimate of what would be considered "too large"
> for which periods in time. Just to given an idea of what would still be
> considered "readably short" and how important something has to be in order
> to be on the list.
>
> -- Jan Hidders
No apology needed. Rough guidelines are fine with me, so long as some
exceptions are allowed for years that are just super-full of history. The
amount of listings in an of itself should also (eventually) give a person a
visual idea of how much history was made in that year. We can discuss the
specifics at the year pages (can't say where exactly, because wikipedia.com
is not accessible right now).
--maveric149
During a routine check for uploaded junk, I saw two gzipped SQL files. I
downloaded one (the other was huge) and found that it contained my password,
as well as other user names. This means that any of our junk uploaders (and
junk was uploaded) could have found this file, read it, and log in as anyone,
including a sysop (the file also tells who is a sysop). Please change your
passwords immediately to prevent this.
phma
I hit the send button too fast, here is the rest:
Parenthetical disambiguation should also be avoided whenever possible.
It would be idiotic to have a non-article list with a disambiguation notice
at [[Paul Simon]] with links to [[Paul Simon (musician)]] and [[Paul Simon
(politician)]] just because these two people share the same name. Of course,
the musician (who is hands down the most widely known person by the name
"Paul Simon") should be at [[Paul Simon]] and the politician should be at an
article that has his middle initial (Like [[George W. Bush]]) and if that
doesn't work then the hideous [[Paul Simon (politician)]] would have to be
used. Notice, the decision to use a parenthetical title here was a last
resort and was only used for the less well-known Paul Simon.
In addition; if one thing is known only by a one word name, and another thing
is usually called by the same one word name in its native context but is
often also called by a naturally disambiguated two word name, there is NO
need to have a disambiguation page at the one word page title (because the
context here is an online hyper-linked encyclopedia, not the native context
of the two word term). We can reasonably assume a person of average or
greater intelligence making a link intended for the two word term, will make
a link to the two word term and not the one word term because this is an
encyclopedia and we should assume that person has enough sense to naturally
disambiguate the link.
For example: It would be equally dumb to turn [[worm]] into a non-article
disambiguation page just because there is also such things as computer worms
and candy worms.
It is my opinion that one of the reasons why wikipedia has been as
preposterously successful as it has, is due the ease with which links can be
made from one article to the next. Unfortunately, as wikipedia grows,
ambiguities will increasingly pop up and resolving those ambiguities might
make linking more and more tedious as time goes by -- if we are not careful.
In order to dramatically slow this, I propose that we should be very careful
when deciding to make disambiguation pages and only make them when true
ambiguities exist (also, to avoid parenthetical disambiguation whenever
possible). No one of average or greater intelligence would link to [[Paris]]
in the context of a hyper-linked encyclopedia and reasonably expect that link
to go to an article about Paris, Texas. I also proposse to make it a policy
that whoever makes these pages should have to fix each and every misdirected
link so that they point directly to where they intend.
--maveric149
On Friday 28 June 2002 04:12 am, Jan Hidders wrote:
> How about the following:
>
> - We make a guideline that says that there should be no more than, say, 10
> births / deaths / events on the page and that these should be the most
> important ones in that year.
I don't think that would be a good idea. Some years are just loaded with many
notable people who were either born, did something really important, or died
during that year that continue to be relevant to us in the present. We
shouldn't place any type of artificial rule on the number of things listed.
After the page gets too long on a readability standpoint weighed against the
relevance that year has to us in the present (1940 being more relevant to us
than say 1840 -- therefore 1940 should naturally be longer no matter what),
it can be broken up and the less notable stuff placed in a "List of .." page.
> - ....
>
> - If someone comes along and thinks a very important event is missing but
> the list is already full then he or she has to remove the least important
> one.
Again, this wouldn't be desirable with such an artificially short and
contraining list -- what is important to a geneticist is going to be
different than what is important to a painter and these two different
contributors would cull very different things when having to choose such a
"top ten list" (esp. for years that both geneticists and painters were doing
astounding things during the same year). And when the time comes for a
break-up of the article, less contraining rules should prevail so as not to
impugn any truly notable events just because those events happened to occur
in a year in which even more notable events occurred in.
There is also the issue that the 20th century has a HUGE amount of history
(more people, more things happening) than any other century in history (both
in terms in sheer amount and more importantly, in terms of what is relevant
to us today). If there are to be any rules to keep these list-type articles
readably short, then much of that should be directed to these pages and not
say, many of the pre-Renaissance AD or BC entries which have one or two
things listed if anything at all. Besides, if the event weren't at least a
little important, chances are very good that any written history from those
earlier times wouldn't have even survived to us today. So this issue might be
self correcting for these earlier year pages.
--maveric149
On Thursday 27 June 2002 12:01 pm, Pierre Abbat wrote:
> Of these risings the first ([December 1848]?â"[July 1849]?) took place in
> Mázandarán?, at the ruined shrine of [Shaykh TabarsI]?, near Bflrfur6sh?,
> where the BábIs, led by Mullâ Muhammad â~All of Bârfurfish and Mullfi
> Husayn of Bushrawayh (â the first who believed â), defied the shahâs
> troops for seven months before they were finally subdued and put to death.
>
> I think that there should be a warning against inserting text with
> illegible scanos in it. Where do they come from? Does anyone have a
> hardcopy of the 1911 encyclopedia so that he can fix them?
I do hate it when people just dump, unwikified, unchecked, incorrectly named
and OCR error ridden 1911 text into wikipedia expecting others to tidy things
after them. It is oftentimes far more difficult to update all this AND
modernize pedandic Victorian, often wrong and ultra-biased prose than it is
to do some research on the topic and create an article from scratch. Dumping
of unimproved 1911 text should be highly discouraged.
--maveric
Whoever has access to the upload page, please add the following in addition
to the previous warnings:
Please do not upload uncompressed sound or picture formats, such as BMP or
WAV, except small (<10KB) files used as examples of the file format. All
large files in uncompressed formats will be summarily deleted.
phma
> I attempted to log in with my username and password from the
> real site and got the message "There is no user by the name
> \"PierreAbbat\".". A username/password check should not reveal
> whether it is the username or the password that is wrong.
I'm not sure that really applies here, where anyone can get a
complete user list anytime anyway (of course it doesn't have
any real information like email addresses or anything).
Is there any real security reason to hide our user list?0
The code running at:
http://www.piclab.com/newwiki/wiki.phtml
is now in a state where I believe all the features I intended to have
working for release are working. I'd like to have as many
Wikipedians as possible pound on it for a few days while I work on a
transition plan for installing it on the new server.
The code is labelled and frozen and will not change during this test
period. Please use the Sourceforge pages (there are links on the
site) to report bugs. At the end of the test period, those bugs that
I think can be fixed without destabilizing anything will be fixed and
tested by me, and that version will be the first one installed on the
new server (which will also get some testing before migrating the
live database).
Feel free to make new feature requests too, but I will not be adding
any significant new features to this release--only bug fixes that are
unlikely to destabilize anything.
Known issues:
Not all quickbar options work well on all browsers. It just isn't
possible. "Fixed left" should work for everyone, and is the
default. "Fixed right" works just about everywhere except Netscape
4.x. "Floating left" can produce odd displays on some pages (notably
those with wide images or tables) on I.E., but otherwise works pretty
well on all browsers I've tried. I will consider fixing bugs in this
area only if (1) "Fixed left" fails for some browser, or (2) Some
common browser has a display so messed up that it's hard to switch
back to the default. "None" should work for everyone as well.
The "StarTrek" skin is gone. Good riddance, as far as I'm concerned.
It doesn't work in the present code anyway, and very few users have
the setting--likely none of them are active. "Standard" is the
default, but I'm open to the idea of making "Cologne Blue" the
default instead if people think that's the right thing to do. I
think "Standard" will work well for more people, though, because of
the shorter sidebar, and it's more likely to resemble the user's own
browser default settings.
The size of the edit area is controlled by your user settings. It
now defaults to 80 columns by 20 rows, which you can change. I know
the current code fills the browser window with the box automatically,
but it's difficult to get that to work well in all browsers, so I
left out that feature.
0