I haven't seen this mentioned here, so I figured I'd post it.
http://researchweb.watson.ibm.com/history/
Fascinating visualization of the history of Wikipedia articles,
contributions, contributors, etc.
~ESP
--
Evan Prodromou <evan(a)wikitravel.org>
Wikitravel - http://www.wikitravel.org/
The free, complete, up-to-date and reliable world-wide travel guide
Evan Prodromou wrote:
>So, isn't the whole point of Wikipedia to create and distribute free
>information? I mean, yes, downstream publishers should -- no, *must*
>-- comply with the GFDL, but besides that, shouldn't we be _happy_
>that someone's redistributing the encyclopedia?
The GNU FDL has a strict author credit requirement. We make things /very/ easy
of downstream users by stating that a link-back to the particular article
copied fullfills this requirement. Given this, and since Wikipedia is a
collective work, it is not much to also ask for a mention of our project's
name.
>We need to make it _easier_ for people to re-publish the encyclopedia
>in a way that complies with the GFDL -- not punish them for doing so.
OK, how does that relate to what I said? I'm all for making it easy to use our
content. But I also think that we should be properly cited since we do not
ask that any /individuals/ get author credit. Again since Wikipedia articles
are collective works, I think that since individuals are not credited in
third party copies, that the project should.
Otherwise downstream users can state only '[This article] is licensed under
the [GNU Free Documentation License]' with [This article] being a link to the
Wikipedia article and [GNU Free Documentation License] being a link to the
GNU FDL (with no underlines under [This article] or [GNU Free Documentation
License] to indicate that those are hyperlinks).
I've seen this on at least one website and I don't like it - it does not give
readable credit to the collective work and is not a proper citation - it
gives the impression (just by reading the message) that the article was
written at that website and that that website somehow owns the copyright.
>That said, it's probably a good idea to ask for links back to
>xx.wikipedia.org. Not to promote the site (hell, it's plenty popular),
>but to make it easier for readers to contribute and fix errors in
>articles.
??? OK, what about the author requirement? The only alternative is for them to
list 5 authors of every article.
-- mav
A month, by month summary and links to each month's detailed PayPal
transactions is now on Meta. For privacy purposes all names, email addresses,
and other personal info has been removed from the data.
If I have blotted out your name and/or user name in a comment field and you
want to have your name and/or user name restored then just email me, drop a
message on my Meta talk page or fix the page yourself. :-)
NOTE: This page and its subpages will probably be protected in a day or two.
http://meta.wikipedia.org/wiki/Wikimedia_PayPal_donations_for_2003
If my presentation of the data is unclear, then please say so on the talk
page.
Source file (OpenOffice Calc format - feel free to add graphs or better
formating/analysis):
http://meta.wikipedia.org/wiki/Image:Wikimedia_PayPal_History_(Aug-Dec_2003…
-- Daniel Mayer (aka mav)
PS Ray Saintonge has also developed a daily summary but I haven't had time to
double check the data he used to obtain those numbers (there were several
errors in the data I gave him).
Scanning my logs, I discovered few sites using Wikipedia articles.
http://www.infosearchpoint.com/display/Chinese_poetryhttp://sciencedaily.com/encyclopedia/I_Chinghttp://nationmaster.com/encyclopedia/Yi-Jing
Is there a page in Wikipedia where to build this list, so developers can
check if those "mirrors" aren't harmful for servers ?
- Infosearch seems to be simply a variant of WP.
- Sciencedaily seems to be a commercial website for students
- Nationmaster seems to be a big container for data and info on countries,
but I found weird results on religion pages (I hope it's a bug).
Delirium wrote:
>Oh, it's certainly a good idea, and doesn't hurt anything,
>to mention Wikipedia as the source. But we can't *require*
>people to do so, and write them nasty letters if they don't,
>unless it actually is a legal requirement, which I don't think
>it is.
We may be able to argue that it is since we explicitly give people the option
of providing a link-back in lieu of the more strict author requirement of the
GNU FDL. Given our liberal interpretation of the GNU FDL I don't think that a
mention of our project's name is much to ask (even if it is just a request,
and not a mandate). As I said before this may even go beyond the GNU FDL and
get into proper referencing practices.
-- Daniel Mayer (aka mav)
A Dutch ISP is mirroring the Dutch Wikipedia in a way which
makes it VERY up to date. (Method yet unknown to us.)
Apparently it is beta at this time - it says so on a logo of
this ISP which is displayed on every page of the mirror.
Currently it is not mentioned on their homepage. One of the
Dutch wikipedians noticed the existence of this mirror in a
logfile. We are wondering what to make of it.
This ISP is using a Wikipedia-logo of their own, and on all
pages the title-tag says '[ISP-name] partner Wikipedia'.
There is no link to the GNU-license, and the pages are not
stating they were copied from Wikipedia. However they have
a link to the original page on the Wiki-server.
For reasons unknown to me, the Dutch version of Wikipedia is
not carrying a subclass title like 'From Wikipedia, the
free encyclopedia'. As a result, this mirror isn't either,
which I feel is a pity.
I informed Jimmy Swales, who encouraged me to post here the
questions I asked him. I am relatively new to Wikipedia, so
please feel not offended if I made a mistake, or am asking an
answer to familiar questions. Also please excuse my English.
Q u e s t i o n s
-----------------
I would like to receive an answer to the following questions:
* Is this ISP entitled to keep an integral and up to date copy
of the Dutch database? (I guess it is.)
* Is this ISP entitled to use a 'Wikipedia'-logo they made
themselves by combining elements of the official Dutch logo?
(I guess they are not.)
* Is 'Wikipedia' somehow registered as a trademark? (Would
make our position stronger I guess, if we want them to stop
using their homemade logo.)
* Is this ISP entitled to call themselves 'partner' of
Wikipedia without our consent? (I guess they are not.)
O b l i g a t i o n s f o r a m i r r o r
-----------------------------------------------
I got the impression the following is mandatory for a mirror:
* every copied page contains the original titletag
* every copied page contains a link to a local kept version
of the GNU-license
* every copied page contains a byline - contained in the source
as a subclass item - mentioning the page is originally from
Wikipedia. (This item is sometimes referred to as 'titlepage'.)
* every copied page contains a link to the original page on
the Wikipedia-server
Am I mistaken?
Is there anything missing on this list?
H o w w e p l a n t o p r o c e e d
-------------------------------------------
First we will contact this ISP, and ask them to provide the
emailaddress of the administrator of this Wikipedia-mirror.
Our next step will be to inform this administrator about the
obligations in case of a mirror, and ask the administrator
whether the ISP is crawling the pages - politely discouraging
them from doing so if this is the case.
Next steps will be to make sure the ISP complies, or to have
the site taken down, if they won't.
Any comment is welcome.
Best regards,
Erik vdMb
Erik van den Muijzenberg wrote:
>Brion just informed me it seems to be *not* a mirror.
>Looks like it just fetches the requested page from our server,
>reformats it to their own liking, and displays it.
That's not very nice at all.
>As their server is not in production it hardly takes bandwith
>or CPU-capacity.
Not yet. But when they get a lot of hits that is going to slow our servers
down. In a sense our servers will be feeding content to a website that may
put adverts all around it and /not/ give us the correct credit.
That's just not nice. They need to use a backup dump and host the content
themselves while following our liberal interpretation of what the GNU FDL
requires.
-- Daniel Mayer (aka mav)
Erik van den Muijzenberg wrote:
>Yes, but they are not interested in the French version. :-)
>It's a Dutch ISP and they are copying the Dutch database.
>(Apparently via a Perl-script.)
D'oh! Dutch, French, what's the difference? Uh, actually a lot.
Sorry about that. :)
--mav
Delirium wrote:
>I'm not entirely sure one of those points is necessary:
>the statement that the article was retrieved from Wikipedia.
Well, that is what the printable page says. So what if it goes over the bounds
of what what is minimally needed for GNU FDL compliance? Another purpose of
the message is for citation purposes. In those cases the name of the source
and when the article was last edited are needed.
-- Daniel Mayer (aka mav)