I just saw a page [[Ethymology]], which resulted from a misspelled link on
[[Science]]. I edited it, replaced it with a redirect to [[Etymology]], and
got an edit conflict on [[Ethymology]] in which the left side was the
redirect and the right side the original text. I then erased the original
text, pasted the redirect, and hit submit. This resulted in replacing
[[Etymology]] with a redirect to itself. Can anyone figure out how this
happened?
phma
The statistics page for Wikipedia says "There have been a total of
2445066 page views, and 49482 page edits since the software was
upgraded (July 20, 2002)."
Will these counters be reset the next time the software is upgraded?
Or will they ever be reset or reduced? What do we really learn from
these numbers? Could they be more useful if calculated differently?
Once the new software is launched for all languages, can we compare
the German Wikipedia against the Polish, using these numbers?
In addition to the accumulated value, I think it would be useful to
have a value that indicates the current rate of activity, either by
beginning a new counter each month (or week), or by having a decaying
average (i.e. an accumulated value filtered through a hi-pass filter).
Either way, it should use the same definition for different sites, so
we can compare values.
--
Lars Aronsson (lars(a)aronsson.se)
tel +46-70-7891609
http://aronsson.se/http://elektrosmog.nu/http://susning.nu/
http://www.wikipedia.org/tarballs/ now contains only intermittent
backups of the English wiki, I can't find any trace non-English tarballs
on the new server, and the "old.wikipedia.com" get-at-the-old-server
trick no longer seems to work.
It's a little worrying.
-- brion vibber (brion @ pobox.com)
I was thinking about creating a static HTML tree from the database
dump. The only thing I can't figure out is how to provide a decent
search function. Are there any ready-made solutions for this problem?
The possibilities I see are
* ship a minimal web server, and use some sort of htdig cgi search script.
* hack together some javascript code which doesn't need a server.
(Something like this exists already at
http://www.javascript.freeuk.com/oldindex.html, but they recommend
it only for "up to 50 pages")
* Use some format other than HTML which provides search capabilities,
such as MS help files or pdf.
I don't like either option.
Axel
> SVG graphics support was suggested on wikipedia-l. I like the
> idea; that's why I tried it for Nupedia once, but... :-(
> Proposal: [[svg:foobar]] is just plain-text editable, but
> displays as an image; A link to [[svg:foobar]] is displayed as
> an image as well. Optional, after saving an edit to
> [[svg:foobar]], an image (foobar.png) could be created via
> svg2png (never tried that, though). The image would then be
> displayed instead of the SVG, and "real" SVG display could be
> turned on in the user options (until most browsers support
> native SVG).
That's pretty much what I had in mind as well, but SVG isn't
mature enough yet to think about. For now, people can upload
the "source file" for a drawing in SVG, as well as uploading a
PNG version, and just link to the PNG in the article.
SVG graphics support was suggested on wikipedia-l. I like the idea;
that's why I tried it for Nupedia once, but... :-(
Proposal:
* [[svg:foobar]] is just plain-text editable, but displays as an image
* A link to [[svg:foobar]] is displayed as an image as well
Optional, after saving an edit to [[svg:foobar]], an image (foobar.png)
could be created via svg2png (never tried that, though). The image would
then be displayed instead of the SVG, and "real" SVG display could be
turned on in the user options (until most browsers support native SVG).
I didn't make this a feature request yet.
Magnus
daniwo59(a)aol.com wrote:
> To me, a major part of the problem is that the material is so out of date. It
> fails to take into account the past hundred years of archeological research,
> which is essential. Furthermore, the statistics it gives about places are
> hopelessly outdated. For example, Anatoth, currently 'Anata, is a fair sized
I think I have a working solution to this sort of problem:
Old texts should be scanned (in facsimile if possible) and put on a
read-only website which allows deep linking. For example, the article
on the Electric Telegraph from a 19th century Swedish encyclopedia is
available on the URL http://www.lysator.liu.se/runeberg/nfad/0192.html
(have a look, nice pictures, all public domain).
The URL up to ...runeberg/ is the name of the website and "nf" is for
the encyclopedia and "nfad" is the 4th volume of the 1st edition.
Then in the wiki, a rule is added so the shorthand "nf:ad0192" is
automatically recognized and converted into a hypertext link, in a
fashion similar to ISBN numbers. The example is found on the wiki
page http://susning.nu/Telegraf
where the wiki text "nf:ad0192" is converted into (my translation)
See [http://www.lysator.liu.se/runeberg/nfad/0192.html the article]
in the 1st edition of [[Nordisk familjebok]], volume 4, 1881.
and then into HTML.
So, what you need is a stable and deep-linkable read-only website with
the old contents that you want to use, and a shorthand linking scheme
in the wiki software. You do not want old text copied into the wiki.
Easton's Bible Dictionary is available in a deep-linkable, stable,
read-only website, the Christian Classics Ethereal Library,
starting on http://www.ccel.org/e/easton/ebd/
For example, the article on Anatoth is available on the URL
http://www.ccel.org/e/easton/ebd/ebd/T0000200.html#T0000233
apparently with 100 articles per HTML page, and this is article 233.
If this is a work that you often want to refer to, add the following
pattern rule to the wikipedia source code for the English Wikipedia,
ebd:([0-9]+) e.g. ebd:233
translated into
'See [http://www.ccel.org/e/easton/ebd/ebd/' +
sprintf("T%05d00.html#T%07d", $1/100, $1) +
' the article] in [[Easton's Bible Dictionary]] (1897)
Adding this "ebd:" rule to the wikipedia software doesn't hurt
anybody, since 99.99% of all articles will not contain the ebd:
pattern. But as soon as anybody, who knows EBD and this rule, starts
to use it, it saves a lot of time and effort in creating links instead
of copying useless text into the wiki.
--
Lars Aronsson (lars(a)aronsson.se)
Aronsson Datateknik
Teknikringen 1e, SE-583 30 Linuxköping, Sweden
tel +46-70-7891609
http://aronsson.se/http://elektrosmog.nu/http://susning.nu/
I just deleted Asianlove.wma at 8:33 according to the server. Jeroen Heijmans
previously deleted it at 16:53, before which I deleted it at 7:45.
1. How can 16:53 be before 8:33 on the same day?
2. How can a file be deleted thrice if it was uploaded only twice?
3. Why is the server not on UTC?
phma
Perhaps this could be interesting for the Wikipedia as well...
Inspired by Wikipedia's "What links here" function, I implemented a
in susning.nu, my Swedish wiki website, a MySQL table that contains a
line for each link, with a column for the source (the page that
contains the link) and one for the destination (the page the link
leads to). Fine.
But I also added a counter to each row of that table, and that counter
is updated on each access (no sweat, it works fine, no performance
degradation), based on the referer URL. And if the referer URL isn't
already in the table, a new record is created. This means also
external referer URLs will be recorded. And if they belong to Google
or some other wellknown search engine, the script extracts the search
expression. This is a way to find out who or what links to wiki page
and how often those links are used, without access to the server logs.
If you want to check it out, my version of the "what links here" is
called "Länkar" and is available in the menu bar at the top and bottom
of each page. The displayed page lists links to and links from each
page, ordered by the frequency of their use.
For example, the page on Ericsson T68, http://susning.nu/Ericsson_T68
has a Link page on the URL
http://susning.nu/susning.fcgi?action=links&id=Ericsson_T68
Since the function was introduced four days ago, that page has seen
this access pattern (my translation):
Most frequently used links to [[Ericsson T68]]
6 - HomePage
5 - external link(-) - external search(t68)
2 - external search(ericsson+t68)
1 - external search(t68+and+t68i+difference) -
external search(wap+settings+for+t68+i)
0 - Ericsson - Xpedio
A total of 8 links, that have been used 20 times.
Most frequently used links from [[Ericsson T68]]
2 - RecentChanges
1 - Buzzword - GPRS - SyncML
0 - Bluetooth - Ericsson - Nokia 8210 - POP3 - SMTP - WAP
A total of 10 links, that have been used 5 times.
The links with zero count are links that exist between wiki pages, but
haven't been used. External referer URLs never get a zero count.
>From the statistics we can learn that people who found the page were
searching for "wap settings" and "differences between t68 and t68i",
so maybe that should be explained by the page.
--
Lars Aronsson (lars(a)aronsson.se)
Aronsson Datateknik
Teknikringen 1e, SE-583 30 Linuxköping, Sweden
tel +46-70-7891609
http://aronsson.se/http://elektrosmog.nu/http://susning.nu/
>Tomasz Wegrzanowski wrote:
>
>>"Show "s and "List "s should be removed.
>>They are only distracting.
>
>I was just thinking the same thing earlier today.
Must have been about the time I was doing it. I love easy fixes.