Beware, shameless Propaganda is on its way. It's about the Interwiki-Link-Checker (http://tools.wikimedia.de/~flacus/IWLC/start.php) by de:user:Flacus, a brilliant tool which has facilitated the generation of >100.000 Interlanguage-Links within the last six month.
There are still lots of articles in en:Wikipedia which have the same name as articles in other Wikipedias, but are *not* linked by an Inter-Wiki-Link (e.g. en-de: ~6300 articles, en-fr: ~6500 articles and en-es ~3100 articles, as of the dumps from February 2006).
This articles have been listed for many language combinations by de:user:SirJective. But not all articles with identical names are about the same subject, so human work is needed to decide whether an Interwiki-Link can be automatically inserted by a bot. In order to make this decision easy and comfortable, there is a brilliant tool called Interwiki-Link-Checker written by de:Benutzer:Flacus.
You might want to check the FAQ at http://de.wikipedia.org/wiki/Benutzer:Flacus/Wikipedia_Interwiki-Link-Checke... or just try it at http://tools.wikimedia.de/~flacus/IWLC/start.php Thanks for your time and (maybe even) help, I can just invite you to try the tool, it's fun and you're reading (and editing) quite a lot of stuff you wouldn't stumble upon otherwise.
If you do like the tool and want to spread the word, visit http://de.wikipedia.org/wiki/Benutzer:Flacus/Wikipedia_Interwiki-Link-Checke..., help and new ideas are welcome.
Cheers, de:Benutzer:Mdangers
On 10/03/06, Marc Dangers marcdangers@gmx.de wrote:
This articles have been listed for many language combinations by de:user:SirJective. But not all articles with identical names are about the same subject, so human work is needed to decide whether an Interwiki-Link can be automatically inserted by a bot. In order to make this decision easy and comfortable, there is a brilliant tool called Interwiki-Link-Checker written by de:Benutzer:Flacus.
This is pretty fun (and surprisingly easy to do, even if you speak neither language - it's amazing how quickly you can guess if two articles are on the same topic, assuming you have even a passing grasp of the alphabet or vocabulary...). But dear goodness, it's *slow*...
-- - Andrew Gray andrew.gray@dunelm.org.uk
On 3/20/06, Andrew Gray shimgray@gmail.com wrote:
This is pretty fun (and surprisingly easy to do, even if you speak neither language - it's amazing how quickly you can guess if two articles are on the same topic, assuming you have even a passing grasp of the alphabet or vocabulary...). But dear goodness, it's *slow*...
I just wish there was a way that interwiki links would not show up on article histories...
Steve
Andrew Gray schrieb:
On 10/03/06, Marc Dangers marcdangers@gmx.de wrote:
This articles have been listed for many language combinations by de:user:SirJective. But not all articles with identical names are about the same subject, so human work is needed to decide whether an Interwiki-Link can be automatically inserted by a bot. In order to make this decision easy and comfortable, there is a brilliant tool called Interwiki-Link-Checker written by de:Benutzer:Flacus.
This is pretty fun (and surprisingly easy to do, even if you speak neither language - it's amazing how quickly you can guess if two articles are on the same topic, assuming you have even a passing grasp of the alphabet or vocabulary...). But dear goodness, it's *slow*...
--
- Andrew Gray andrew.gray@dunelm.org.uk
WikiEN-l mailing list WikiEN-l@Wikipedia.org To unsubscribe from this mailing list, visit: http://mail.wikipedia.org/mailman/listinfo/wikien-l
Regarding the slowness, in general the tool itself appears to be fast enough for me, most of the time. I actually use it without being logged in, so I suppose the Dutch Proxies are my friends and I get lots of cached pages. "Guessing" unknown languages is lots of fun, if you know at least one germanic and one romanic language, you can grasp quite a lot of meaning. My ability to understand swedish and norwegian is probably comparable to a 1.6 year old native speaker, that's enough for me ;-). Thanks for trying it, Marc Dangers de:Mdangers
On 20/03/06, mdangers marcdangers@gmx.de wrote:
Regarding the slowness, in general the tool itself appears to be fast enough for me, most of the time. I actually use it without being logged in, so I suppose the Dutch Proxies are my friends and I get lots of cached pages.
Having experimented further, it seems to come in batches - it'll be immensely slow one hour (never loading anything past the first pair of pages), and happily zip through them the next.
"Guessing" unknown languages is lots of fun, if you know at least one germanic and one romanic language, you can grasp quite a lot of meaning. My ability to understand swedish and norwegian is probably comparable to a 1.6 year old native speaker, that's enough for me ;-).
I spent last night interwiki-ing Norwegian and Catalan, of all things...
-- - Andrew Gray andrew.gray@dunelm.org.uk
On 3/20/06, Andrew Gray shimgray@gmail.com wrote:
I spent last night interwiki-ing Norwegian and Catalan, of all things...
I haven't tried this tool, but is it smart enough to know that if en.foo links to fr.foo, and fr.foo links to ca.foo, then en.foo should link to ca.foo? Otherwise this problem seems like a massive amount of manual intervention...
Steve
"Steve Bennett" stevage@gmail.com wrote in message news:f1c3529e0603200617x7c05c5f2n9671e14bfa4c753a@mail.gmail.com...
On 3/20/06, Andrew Gray shimgray@gmail.com wrote:
I spent last night interwiki-ing Norwegian and Catalan, of all things...
I haven't tried this tool, but is it smart enough to know that if en.foo links to fr.foo, and fr.foo links to ca.foo, then en.foo should link to ca.foo? Otherwise this problem seems like a massive amount of manual intervention...
That's a later stage of the process.
This particular tool is to allow human intervention to ensure that [[en:foo]] and [[fr:foo]] are actually about the same "foo", and similarly for [[ca:foo]].
I'm doing the "de-en" set at the moment, and there are quite a few cases where the two articles, despite having the same name, are radically different.
The usual reason is that there are a good many things called "foo", all of which probably appear on [[foo (disambiguation)]] or the equivalent. However, different languages have their own ideas about which of those many instances of "foo" should get the main article of that name.
HTH HAND
On 3/20/06, Phil Boswell phil.boswell@gmail.com wrote:
This particular tool is to allow human intervention to ensure that [[en:foo]] and [[fr:foo]] are actually about the same "foo", and similarly for [[ca:foo]].
I'm doing the "de-en" set at the moment, and there are quite a few cases where the two articles, despite having the same name, are radically different.
Ok, I can see how that could happen... fictitious example: En: [The Foo King] is about an anglophone TV series, with a paragraph at the end to the original book [Le Roi Fou]. Links to.. Fr: [Le Roi Fou], an article about the book. It was written by a Czech author, who is well known in France precisely for this single book. So information about the author makes up at least half the article. It links to... Cz: [Fou Fouzcy]...about the author, somewhat known in Czech for all his other great works. No mention at all of that particular book...
You're right, it wouldn't make much sense to link [The Foo King] to [Fou Fouzcy] in this case...
(ok that example was probably a bit farfetched...but anyway...)
Steve
Steve Bennett schrieb:
On 3/20/06, Andrew Gray shimgray@gmail.com wrote:
I spent last night interwiki-ing Norwegian and Catalan, of all things...
I haven't tried this tool, but is it smart enough to know that if en.foo links to fr.foo, and fr.foo links to ca.foo, then en.foo should link to ca.foo? Otherwise this problem seems like a massive amount of manual intervention...
Steve _______________________________________________ WikiEN-l mailing list WikiEN-l@Wikipedia.org To unsubscribe from this mailing list, visit: http://mail.wikipedia.org/mailman/listinfo/wikien-l
Actually the different bots based on the pywikipediabot network are smart enough to do this and are responsible for insertion of vast amounts of Interwiki links (IWL). The tool is as yet the only means I know of to at least semi-automatically insert IWLs into articles which are *not* linked to any other language.
Since FlaBot, which is inserting most/lots of the verified IWLs, is also pywikipediabot-based, it is inserting extra IWLs along these lines, e.g. if en: is verified to be the same in de:, extra links to fr: on en: are also inserted in de: (hope this is readable/understandable, you know what I mean, anyway, I guess).
I already tried to persuade Flacus earlier to primarily include articles without any IWL. This would make sense in language pairs with a huge backlog, at least. I came too late, anyway, because just then the February dump data was already included into his database. From personal experience, I believe that this would not make a very big difference, anyway, maybe 10-20% less article pairs.
marc