Hi, I found really good the Wikimedia Outreach Project, which tries to improve the relationships between sister projects and Commons. The it.wiki community has talked long about this.
The description of the pictures are almost always in english (anyway in a single language: the language of the uploader) and this makes things not easy for those who explore commons and want to understand what the descriptions say. I was wondering "what about a requiring-language template?"
For example we could add in the information template - in the description field - the main language-templates:
{{Information |Description = {{de|...}} {{en|...}} {{es|...}} {{fr|...}} {{it|...}}
When a language-template is not filled it could appear
en: description of XXX fr: *description in french language required* it: descrizione di XXX ecc..
When a language is missing, the picture could be automatically put in a category like [[Category: French description required]].
On 8/30/07, Andrew bouncey2k@gmail.com wrote:
I was wondering "what about a requiring-language template?"
For example we could add in the information template - in the description field - the main language-templates:
{{Information |Description = {{de|...}} {{en|...}} {{es|...}} {{fr|...}} {{it|...}}
When a language-template is not filled it could appear
en: description of XXX fr: *description in french language required* it: descrizione di XXX ecc..
When a language is missing, the picture could be automatically put in a category like [[Category: French description required]].
The problem with this is that nearly every image on Commons needs to be translated into some language. The only images that are consistently translated into more than 5 are the featured pictures and pictures of the day, but considering how many files Commons has that is certainly a small percentage.
Categories such as you describe may seem like a good idea, but they will contain about (very very vague guess here) 95% of all the files we currently have; at 1,828,506 http://commons.wikimedia.org/wiki/Special:Statisticsfiles and counting, that is a lot of files to have in a category. The only category that will be less full would be the "English translation required" category, but that one will have quite a number of files as well.
I think a better way to do it is probably to increase awareness about translation efforts, ask for volunteers to translate images as they come across them, or even let other-language wikis know that their contributors are more than welcome to come to Commons and translate the description of a file they are using. If every person who comes from it. or fr. or de. or ru. or zh. (etc. etc.) wiki to find a picture for a new article on their wikipedia translated the descriptions of the 5 - 10 files they come across in their search, that would be a huge help. If there are 200 people from other-language wikis coming to Commons every day to look for images, each look at at 3 images and translate the descriptions, that's 600 new translations per day.
Ayelie ayelie.at.large@gmail.com wrote on Thu, 30 Aug 2007 09:40:05 -0400:
I think a better way to do it is probably to increase awareness about translation efforts, ask for volunteers to translate images as they come across them, or even let other-language wikis know that their contributors are more than welcome to come to Commons and translate the description of a file they are using. If every person who comes from it. or fr. or de. or ru. or zh. (etc. etc.) wiki to find a picture for a new article on their wikipedia translated the descriptions of the 5 - 10 files they come across in their search, that would be a huge help. If there are 200 people from other-language wikis coming to Commons every day to look for images, each look at at 3 images and translate the descriptions, that's 600 new translations per day.
What about a javascript thingie that matches the language templates of the currently viewed image and the language code of the browser and asks for translation? Magnus?
regards,
Flo
I am aware of the fact the Commons has got over 1,800,000 files. But if we never fix this, the problem will always remain. Better late than never.
On 8/30/07, Andrew bouncey2k@gmail.com wrote:
I am aware of the fact the Commons has got over 1,800,000 files. But if we never fix this, the problem will always remain. Better late than never.
Oh I wasn't saying the issue doesn't have to be fixed. I just meant that other means may be more likely to get the translations we are looking for, since adding them all to categories would lead to very overpopulated categories that people are scared to look at (trust me, same thing happens with CAT:IFC ... gives me nightmares now that I'm back from vacation and get to haul through all the new stuff). Working at things a bit at a time and getting outside help or bringing the situation to the attention of the community at large and requesting people help out may be a more reasonable way than just adding every image to "translate to x language" categories.
I certainly agree with your opinion on the state of translations, we do need to deal with it. There is also the issue of untranslated categories and articles, which tend to get slightly more translations but are still lacking.
Hi list,
---Selon Ayelie ayelie.at.large@gmail.com:
There is also the issue of untranslated categories and articles, which tend to get slightly more translations but are still lacking.
Could someone run a bot to add a [[Category:No language templates]] to articles that contain none of the currently defined language templates? This would be a help to start the process, even if the category may be overpopulated :)
While writing this, I realize that: that bot may be configured to only add 200 (or 100, or 500) items to the category, and would only run again when less than 20 (or 50, or 100) articles remain in the category... Then the category population would be less scary for the volunteers like me :)
Best regards from France,
On 8/30/07, Alexandre NOUVEL alexandre.nouvel@alnoprods.net wrote:
While writing this, I realize that: that bot may be configured to only add 200 (or 100, or 500) items to the category, and would only run again when less than 20 (or 50, or 100) articles remain in the category... Then the category population would be less scary for the volunteers like me :)
Sounds doable. Maybe tomorrow if nobody has caught it before.
Bryan
If we run a bot to add a [[Category:No language templates]], we are at the same point IMHO.
A picture with an {{en}} template would be skipped. Instead we should mark it with the following categories:
[[Category: no french template]] [[Category: no german template]] [[Category: no spanish template]] [[Category: no italian template]]
etc..
This way people can work. French people work on french templates, italian on italian, and so on. If we have just a category ([[Category:No language templates]]), this will be overfilled.
Andrew wrote:
If we run a bot to add a [[Category:No language templates]], we are at the same point IMHO.
A picture with an {{en}} template would be skipped. Instead we should mark it with the following categories:
[[Category: no french template]] [[Category: no german template]] [[Category: no spanish template]] [[Category: no italian template]]
etc..
This way people can work. French people work on french templates, italian on italian, and so on. If we have just a category ([[Category:No language templates]]), this will be overfilled.
Better do it at the toolserver. So you can ask to a page: give me some pages without X translation which have in Y or Z languages.
On 31/08/2007, Platonides Platonides@gmail.com wrote:
Andrew wrote:
If we run a bot to add a [[Category:No language templates]], we are at the same point IMHO.
A picture with an {{en}} template would be skipped. Instead we should mark it with the following categories:
[[Category: no french template]] [[Category: no german template]] [[Category: no spanish template]] [[Category: no italian template]]
etc..
This way people can work. French people work on french templates, italian on italian, and so on. If we have just a category ([[Category:No language templates]]), this will be overfilled.
Better do it at the toolserver. So you can ask to a page: give me some pages without X translation which have in Y or Z languages.
I strongly agree: use the toolserver, not categories. We do need some way of really encouraging people to add translations, and figure out which ones need them (esp. starting with POTD/FP and QI) but I don't think categories are the way to go. Possibly invisible templates, but a toolserver thing that can just figure it out is even better.
Bouncey2k, thanks for starting the itwikipedia page. I hope it takes off!
cheers Brianna
Platonides Platonides@gmail.com wrote on Fri, 31 Aug 2007 00:05:14 +0200:
Andrew wrote:
If we run a bot to add a [[Category:No language templates]], we are at the same point IMHO.
A picture with an {{en}} template would be skipped. Instead we should mark it with the following categories:
[[Category: no french template]] [[Category: no german template]] [[Category: no spanish template]] [[Category: no italian template]]
etc..
This way people can work. French people work on french templates, italian on italian, and so on. If we have just a category ([[Category:No language templates]]), this will be overfilled.
Better do it at the toolserver. So you can ask to a page: give me some pages without X translation which have in Y or Z languages.
Wasn't there some bot which used to copy captions from local Wikipedias?
Regards,
Flo
WANTED: SQL expert ;)
On 8/30/07, Bryan Tong Minh bryan.tongminh@gmail.com wrote:
Sounds doable. Maybe tomorrow if nobody has caught it before.
Bryan
Can't figure out how to do it :(
Bryan
Explain your problem :)
2007/9/1, Bryan Tong Minh bryan.tongminh@gmail.com:
WANTED: SQL expert ;)
On 8/30/07, Bryan Tong Minh bryan.tongminh@gmail.com wrote:
Sounds doable. Maybe tomorrow if nobody has caught it before.
Bryan
Can't figure out how to do it :(
Bryan
Commons-l mailing list Commons-l@lists.wikimedia.org http://lists.wikimedia.org/mailman/listinfo/commons-l
On 9/1/07, Andrew bouncey2k@gmail.com wrote:
Explain your problem :)
Already got something: http://tools.wikimedia.de/~bryan/unI18N.php. Needs more features though and the random page generator is not excactly working.
Already got something: http://tools.wikimedia.de/~bryan/unI18N.php. Needs more features though and the random page generator is not excactly working.
what's that for?
Andrew bouncey2k@gmail.com wrote on Sat, 1 Sep 2007 16:58:13 +0200:
Already got something: http://tools.wikimedia.de/~bryan/unI18N.php. Needs more features though and the random page generator is not excactly working.
what's that for?
I assume it is for finding images without language tags like {{en}}.
On Friday I created a js function to insert links that help you by asking you for a translation or by putting {{lang|}} around a selected text (firefox only).
Maybe you want to take a look at the functions show_infobox() and insert_translation() at http://commons.wikimedia.org/wiki/User:Flominator/flos_functions.js
Someone, I believe it was Magnus, wrote SumItUp which collects the first sentences of an article in all available languages. Maybe someone could write a tool, that uses CheckUsage in order to retrieve image descriptions from wikipedia articles.
Regards,
Flo
On 03/09/07, Florian Straub Flominator@gmx.net wrote:
Someone, I believe it was Magnus, wrote SumItUp which collects the first sentences of an article in all available languages. Maybe someone could write a tool, that uses CheckUsage in order to retrieve image descriptions from wikipedia articles.
There used to be a bot that did that... it copied any caption the image was given in any Wikipedia to its Commons page. Haven't noticed it on my watchlist for a while so probably it's having a bot sleep. Let's make sure we don't re-invent the wheel, folks :)
To Bryan, if people want to provide captions they would probably prefer to be able to pick a topic to do it for rather than just random. So get people to supply a keyword, find the closest category via CommonSense and just go through all the images in that category.
cheers Brianna
"Brianna Laugher" brianna.laugher@gmail.com wrote on Monday, September 03, 2007 8:41 AM:
On 03/09/07, Florian Straub Flominator@gmx.net wrote:
Someone, I believe it was Magnus, wrote SumItUp which collects the first sentences of an article in all available languages. Maybe someone could write a tool, that uses CheckUsage in order to retrieve image descriptions from wikipedia articles.
There used to be a bot that did that... it copied any caption the image was given in any Wikipedia to its Commons page. Haven't noticed it on my watchlist for a while so probably it's having a bot sleep. Let's make sure we don't re-invent the wheel, folks :)
Any idea how to find him again? I don't think that it will be much trouble reinventing this thing, since some people already have decent frameworks at hand ...
Anyone willing to search this guy or to code?
Regards,
Flo
On 10/1/07, Florian Straub flominator@gmx.net wrote:
"Brianna Laugher" brianna.laugher@gmail.com wrote on Monday, September 03, 2007 8:41 AM:
On 03/09/07, Florian Straub Flominator@gmx.net wrote:
Someone, I believe it was Magnus, wrote SumItUp which collects the first sentences of an article in all available languages. Maybe someone could write a tool, that uses CheckUsage in order to retrieve image descriptions from wikipedia articles.
There used to be a bot that did that... it copied any caption the image was given in any Wikipedia to its Commons page. Haven't noticed it on my watchlist for a while so probably it's having a bot sleep. Let's make sure we don't re-invent the wheel, folks :)
Any idea how to find him again? I don't think that it will be much trouble reinventing this thing, since some people already have decent frameworks at hand ...
Anyone willing to search this guy or to code?
If the mystery author is not found, I'll commence with wheel re-invention tonight :-)
Magnus
On 10/2/07, Magnus Manske magnusmanske@googlemail.com wrote:
On 10/1/07, Florian Straub flominator@gmx.net wrote:
"Brianna Laugher" brianna.laugher@gmail.com wrote on Monday, September 03, 2007 8:41 AM:
On 03/09/07, Florian Straub Flominator@gmx.net wrote:
Someone, I believe it was Magnus, wrote SumItUp which collects the first sentences of an article in all available languages. Maybe someone could write a tool, that uses CheckUsage in order to retrieve image descriptions from wikipedia articles.
There used to be a bot that did that... it copied any caption the image was given in any Wikipedia to its Commons page. Haven't noticed it on my watchlist for a while so probably it's having a bot sleep. Let's make sure we don't re-invent the wheel, folks :)
Any idea how to find him again? I don't think that it will be much trouble reinventing this thing, since some people already have decent frameworks at hand ...
Anyone willing to search this guy or to code?
If the mystery author is not found, I'll commence with wheel re-invention tonight :-)
OK, the new, improved, reinvented wheel: http://tools.wikimedia.de/~magnus/whatisthat.php
Example: http://tools.wikimedia.de/~magnus/whatisthat.php?image=Beatriz_Paredes.jpg
This works by screenscraping HTML. This was the only way I could find around the potential template hell :-( Works nice, though, except for image captions that are not given as thumbnail desc.
If there are several captions in a language (like for en in the example), it will simply append them.
Cheers, Magnus
On 10/2/07, Magnus Manske magnusmanske@googlemail.com wrote:
This works by screenscraping HTML. This was the only way I could find around the potential template hell :-( Works nice, though, except for image captions that are not given as thumbnail desc.
Are you aware of the rvexpandtemplates function of the api? That would kill all the template hell :)
Looks very nice anyway!
On 10/3/07, Bryan Tong Minh bryan.tongminh@gmail.com wrote:
On 10/2/07, Magnus Manske magnusmanske@googlemail.com wrote:
This works by screenscraping HTML. This was the only way I could find around the potential template hell :-( Works nice, though, except for image captions that are not given as thumbnail desc.
Are you aware of the rvexpandtemplates function of the api? That would kill all the template hell :)
Cool! I'll look into it!
Thanks, Magnus
"Magnus Manske" magnusmanske@googlemail.com wrote on Wednesday, October 03, 2007 8:15 PM:
On 10/3/07, Bryan Tong Minh bryan.tongminh@gmail.com wrote:
On 10/2/07, Magnus Manske magnusmanske@googlemail.com wrote:
This works by screenscraping HTML. This was the only way I could find around the potential template hell :-( Works nice, though, except for image captions that are not given as thumbnail desc.
Are you aware of the rvexpandtemplates function of the api? That would kill all the template hell :)
Cool! I'll look into it!
Awesome, simply awesome. Thank you very much.
The question is: What goes wrong at http://tools.wikimedia.de/~magnus/whatisthat.php?image=Image:Clive.jpg&d... ?
Another small suggestion: Please insert $image in the box when $doit is set.
Regards and thank you very much,
Flo
Florian Straub wrote:
The question is: What goes wrong at http://tools.wikimedia.de/~magnus/whatisthat.php?image=Image:Clive.jpg&d... ?
Try use http://tools.wikimedia.de/~magnus/whatisthat.php?image=Clive.jpg&doit=Do...
Leaving out the "image:" works for me.
Cheers Matt
On 10/4/07, Florian Straub flominator@gmx.net wrote:
Awesome, simply awesome. Thank you very much.
The question is: What goes wrong at http://tools.wikimedia.de/~magnus/whatisthat.php?image=Image:Clive.jpg&d... ?
As Matthias pointed out, it's the "Image:". I now filter that.
Another small suggestion: Please insert $image in the box when $doit is set.
Done.
Cheers, Magnus
: "Magnus Manske" magnusmanske@googlemail.com wrote on Thursday, October 04, 2007 3:02 PM:
On 10/4/07, Florian Straub flominator@gmx.net wrote:
The question is: What goes wrong at http://tools.wikimedia.de/~magnus/whatisthat.php?image=Image:Clive.jpg&d... ?
As Matthias pointed out, it's the "Image:". I now filter that.
Thanks to both of you.
Another small suggestion: Please insert $image in the box when $doit is set.
Done.
Thx.
What about combining this one with SumItUp somehow?
Regards,
Flo
On 10/4/07, Florian Straub flominator@gmx.net wrote:
What about combining this one with SumItUp somehow?
IMHO the aims are quite different.
* SumItUp Knows that it should start at the beginning of the article, but has to skip templates, tables, floating images etc. to find the first paragraph, fix links, markup etc.
* WhatIsThat just takes brief plain text; the challange is to find it on the page.
While the output is similar, the acquisition method differs a lot. What would be the advantage of combining them?
Magnus
"Magnus Manske" magnusmanske@googlemail.com wrote on Thursday, October 04, 2007 5:39 PM
On 10/4/07, Florian Straub flominator@gmx.net wrote:
What about combining this one with SumItUp somehow?
IMHO the aims are quite different.
- SumItUp Knows that it should start at the beginning of the article,
but has to skip templates, tables, floating images etc. to find the first paragraph, fix links, markup etc.
- WhatIsThat just takes brief plain text; the challange is to find it
on the page.
While the output is similar, the acquisition method differs a lot. What would be the advantage of combining them?
One way of starting them: If it is an image, start whatisthat, else start sumitup.
Regards,
Flo
On Friday I created a js function to insert links that help you by asking you for a translation or by putting {{lang|}} around a selected text (firefox only).
Maybe you want to take a look at the functions show_infobox() and insert_translation() at http://commons.wikimedia.org/wiki/User:Flominator/flos_functions.js
How does it work?
Andrew bouncey2k@gmail.com wrote on Mon, 3 Sep 2007 12:48:43 +0200:
Flo wrote:
On Friday I created a js function to insert links that help you by
asking you for a translation or by putting {{lang|}} around a selected text (firefox only).
Maybe you want to take a look at the functions show_infobox() and
insert_translation() at
http://commons.wikimedia.org/wiki/User:Flominator/flos_functions.js
How does it work?
Put the following stuff into your monobook.js:
includePage('User:Flominator/flos functions.js'); document.addEventListener("DOMContentLoaded", function() {
insert_link_before("li", "pt-mycontris", '<a href="http://tools.wikimedia.de/~magnus/mynoinfo.php?user=Flominator">Info-Tag</a>'); show_infobox(); }, false);
This is a version only for de and en but if you use a modified function show_infobox you can customize it.
I'm sorry for the lack in usability but I didn't know how to make it better.
Regards,
Flo
Put the following stuff into your monobook.js:
includePage('User:Flominator/flos functions.js'); document.addEventListener("DOMContentLoaded", function() {
insert_link_before("li", "pt-mycontris", '<a href="http://tools.wikimedia.de/~magnus/mynoinfo.php?user=Flominator">Info-Tag</a>'); show_infobox(); }, false);
done. then, how does it work?
Andrew bouncey2k@gmail.com wrote on Mon, 3 Sep 2007 18:59:03 +0200:
Put the following stuff into your monobook.js:
includePage('User:Flominator/flos functions.js'); document.addEventListener("DOMContentLoaded", function() {
insert_link_before("li", "pt-mycontris", '<a
href="http://tools.wikimedia.de/~magnus/mynoinfo.php?user=Flominator%22%3EInfo-Tag</a>');
show_infobox(); }, false);
done. then, how does it work?
Empty your cache and you should see some icons on the left (if you use firefox ;)
Regards,
Flo
On 8/30/07, Florian Straub Flominator@gmx.net wrote:
What about a javascript thingie that matches the language templates of the currently viewed image and the language code of the browser and asks for translation? Magnus?
Sorry, I was on vacation :-) There ya go: http://commons.wikimedia.org/wiki/MediaWiki:MyLangNotify.js
Try it with an image that has no language tags: http://commons.wikimedia.org/wiki/Image:Ljustorpsfiskare.jpg?withJS=MediaWik...
And one that has {{de}} and {{en}}: http://commons.wikimedia.org/wiki/Image:Lab_bench.jpg?withJS=MediaWiki:MyLan...
If your user language is in the description, it will emphasize your language by increasing the font size. If not, it will (not so) kindly ask you to add a description in your language (see the page heading). This works only on image pages.
Cheers, Magnus
"Magnus Manske" magnusmanske@googlemail.com wrote on Mon, 3 Sep 2007 13:18:31 +0100:
On 8/30/07, Florian Straub Flominator@gmx.net wrote:
What about a javascript thingie that matches the language templates of
the currently viewed image and the language code of the browser and asks for translation? Magnus?
Sorry, I was on vacation :-) There ya go: http://commons.wikimedia.org/wiki/MediaWiki:MyLangNotify.js
Try it with an image that has no language tags: http://commons.wikimedia.org/wiki/Image:Ljustorpsfiskare.jpg?withJS=MediaWik...
And one that has {{de}} and {{en}}: http://commons.wikimedia.org/wiki/Image:Lab_bench.jpg?withJS=MediaWiki:MyLan...
If your user language is in the description, it will emphasize your language by increasing the font size. If not, it will (not so) kindly ask you to add a description in your language (see the page heading). This works only on image pages.
Awesome as always :)
What about reading the language of the browser?
Also I would like to have an even more aggressive alert()-message. Would that work as well?
regards,
Flo
On 9/3/07, Magnus Manske magnusmanske@googlemail.com wrote:
Sorry, I was on vacation :-) There ya go: http://commons.wikimedia.org/wiki/MediaWiki:MyLangNotify.js
Try it with an image that has no language tags:
http://commons.wikimedia.org/wiki/Image:Ljustorpsfiskare.jpg?withJS=MediaWik...
Oh, dear, thank you for the laugh. Good start to the day :)
Very nice job, though it could be improved aesthetically perhaps; but definitely a good start and base. Great work again, Magnus, kudos. :D
Magnus Manske wrote:
Sorry, I was on vacation :-) There ya go: http://commons.wikimedia.org/wiki/MediaWiki:MyLangNotify.js
Try it with an image that has no language tags: http://commons.wikimedia.org/wiki/Image:Ljustorpsfiskare.jpg?withJS=MediaWik...
And one that has {{de}} and {{en}}: http://commons.wikimedia.org/wiki/Image:Lab_bench.jpg?withJS=MediaWiki:MyLan...
If your user language is in the description, it will emphasize your language by increasing the font size. If not, it will (not so) kindly ask you to add a description in your language (see the page heading). This works only on image pages.
Cheers, Magnus
Oh, i don't understand the description language. Also, there's some red text below the image name. If i could just understand it... ;-)