I just uploaded a DVD ISO image (bzip2, ca. 1GB; ~1.3BG uncompressed) to
the wikipedia server. It contains a snapshot of the German wikipedia,
including images, for offline browsing under Window$. It can be found at
http://download.wikimedia.org/wp_de_2004_05.iso.bz2
It comes with a self-written installer (hereby under GPL, will upload
source soon) for the webserver software. You can choose if you want to
copy the database and/or the images to the hard drive as well, or leave
them on DVD. For the fulltext search to work, I highly recommend to copy
the database to HDD.
Get it to work:
1. Decrompress the ISO image
2. Burn the ISO image to DVD
3. Run the Installer
4. Go to the installation directory and run "start.bat" (twice if it
doesn't show the homepage right away)
5. To shut down the local web server, run "stop.bat"
As the pages are generated on-the-fly from a sqlite database, and that
generation is done by a relatively new C++ software, many little bugs
are to be expected. Math display doesn't work at all.
Note that when copying the database to HDD, one can actually edit,
preview and save pages!
I consider this more a proof-of-principle rather than a product
distributable on a large scale, but I strongly believe it is an
important step in the right direction.
Some things I'd suggest for further development:
* Auto-update to current online version via internet
* Sumbitting offline edits to the online version
* Another web server that runs directly from CD/DVD (no need for
installation)
* Linux and Mac software, running on the same data
BTW: Does anyone know a filesystem compression software that runs under
Windows and doesn't require manual installation? If we could reduce
database/image size by 50%, it would all fit on a CD again (for now).
I'll be on vacation for two weeks, starting tomorrow, so I might not
answer to all suggestions/criticism right away.
Magnus
P.S.: An alternative would probably be a static HTML dump and
HouseSpider (
http://freshmeat.net/projects/housespider/?branch_id=28546&release_id=161379
), but it requires Java.
I have managed to make changes to code which enable users to choose their user
interface language, feature sought at Wikisource, Wiktionary etc. To whom
should I send the code?
I'm going to assume wikibooks is one of the smaller
projects that's been updated to 1.3. Well, some
strange things started happening there tonight,
mainly, the project namespace has moved from
wikibooks: to wikipedia:. Interwiki links don't seem
to be working, and a few other minor problems. I put
some I've noticed at
http://wikibooks.org/wiki/Wikipedia:Staff_lounge
Thanks in advance for looking into this,
John Pozniak (gentgeen)
These have probably been noticed by somebody, but in case not, there's a
few issues I've immediately noticed upon the upgrade:
1) The inline linking to other wikis on meta: no longer works, so
instead of [[en:blah]] showing up as a link to
en.wikipedia.org/wiki/blah, it shows up at the top of the page as the
"English" link amongst the interwiki links.
2) The "retreived from" printed-version footer is generated for all
pages, not just printed versions.
That's it for now I think. =]
-Mark
Please forgive if this is the incorrect list to post my question. Also,
I am not currently a member of this list, so will require a direct
reply.
Basically, I'm curious to know if Wikipedia currently supports, or has
plans to support video files in the future.
I am thinking about doing PhD work on "online biographical video
learning objects"...and wondered if this maps at all w/ Wikipedia (past,
present or future).
Also, please feel free to educate me on how I could have found this
information out without spamming everyone.
Thanks!
John
On Tomk32's request I've blocked www.themensuche.de (217.172.181.98)
from access to our servers. It was "hosting" German Wikipedia pages by
sending every request on to de.wikipedia.org and slurping the content,
but using it solely for the purpose of bringing in search hits. The text
is never displayed; instead every hit uses a JavaScript redirect through
an intermediary or two to amazon.de.
Pages looked like this:
<html>
<head>
<title>Der Jäger aus Kurpfalz</title>
<script language="javascript">
what = 'site www themensuche de aus';
</script>
<script language="javascript" src="/daten/dat.js" type="text/javascript">
</script>
<link rel="stylesheet" type="text/css" href="/daten/lay.css">
</head>
<body bgcolor="white">
<div id="dat" name="dat">
Friedrich Wilhelm Utsch Friedrich Wilhelm Utsch wurde
[actual content from wiki snipped for brevity]
<center>
<a href="index.html">HOME</a> |
<a href="De.htm">INDEX</a> |
<a href="mailto:webmaster@themensuche.de">MAIL</a>
<hr>
<p><a href="http://www.google.de/search?q=Der Jäger aus Kurpfalz">SUCHE
BEI GOOGLE</a> | <a href="http://search.msn.com/results.aspx?q=Der Jäger
aus Kurpfalz">SUCHE BEI MSN</a></p>
<p>
History:<br>
Copyright (c) 2004
<D-E.W-I-K-I-P-E-D-I-A.O-R-G></wiki/Der_Jäger_aus_Kurpfalz><br>
Permission is granted to copy, distribute and/or modify this document
under the terms of the GNU Free Documentation License, Version 1.2
or any later version published by the Free Software Foundation.
A copy of the license is included in the section entitled<br>
<a href="gnu_fdl.txt">"GNU Free Documentation License"</a>.
</p>
</center>
</div>
</body>
</html>
Where the script brought up at the top contains only:
top.location.replace('http://www.steine.de/partnersys/index.html?swi=' +
what)
which if you go there, sends you on to the amazon.de search page. This
script is executed automatically before any content is shown (and the
markup is invalid and nothing at all shows in some browsers even if JS
is disabled). There is no way to read Wikipedia content at that site
short of turning off JavaScript and using 'view source'.
This is not even *vaguely* legitimate. I've cut them off in the IP
firewall on coronelli and browne, so they're no longer able to steal
bandwidth on every page hit just to promote their referrer links.
-- brion vibber (brion @ pobox.com)
As reported on wikipedia-l, zh was the subject of a large-scale
automated attack. Approximately 3000 pages were deleted over the last
two days. I just wanted to point out that this could have been prevented
by the use of an appropriately configured proxy scanner.
Scanning for proxies may require some administrative overhead in terms
of replying to automatically generated intrusion detection messages, and
in obtaining an understanding from Verio and any other upstream network
service providers. But it would certainly have its benefits.
Our response to this in human terms was less than ideal, but I'll post
my thoughts on that to wikipedia-l.
-- Tim Starling
Is it possible to have a couple of wikipedia logos (several language) in
high resolution, for display at a larger size that the current one on
Wikipedia ?
Are they directly available somewhere, or could Nohat tell me how I
could have these ?
thanks
Ant
I keep getting "Konnte keine Verbindung zur Datenbank auf $1 herstellen"
(can't connect to database on $1) on the German wp, and the "technical
difficulties" message on en. Then it works again for a few minutes. I
know it is olny a glitch in the matrix ;-) but today it appears quite
often, which becomes annoying.
On a (hopefully not) related matter, I am in the process of rsync-ing
the images of the German wp, for an offline (CD/DVD) edition. Is rsync
stressing the server too much, or can I continue with this?
Magnus
After the recent vandal attack on zh:, I have written a bot to enable
faster deletion of pages. Basically, it shows each page in a list, and
asks one whether to delete it. Because of the contentious nature of the
bot, I prefer not to give out the code freely (although it is actually
very simple), but trusted users with some experience in large-scale
vandalism repair can ask for it.
Note that this bot is meant to be used ONLY in cases of large-scale
vandalism, NOT for normal deletion.
Also note that Tim is fully correct in stating that a developer could
do things even easier.
Andre Engels