Ttrimm
1
Is anyone doing images these days?? http://www.wikihow.com/Do-Pin-up-/-Rockabilly-Make-Up
Pretty sure every single one is a copyvio
Maniac
2
I already talked to Ttrimm about this, but I figured it would be a helpful explanation for the entire community.
Nowadays, images are checked during an automated process coming from two fronts. The first part of the script that runs, checks all of the images that are uploaded with no license at all. When a person chooses the “I just found this on the Internet. I don’t know what the license is” option when s/he is uploading an image, the image is placed in the “Unlicensed Images” category. This category is checked periodically (once every hour or two) by an automated script which, in turn, notifies the uploader that their image is unlicensed. The uploader then has 7 days to respond to the talk page license notification indicating where s/he got the image from. If a response is not received within a week, then the image is deleted, and any instance of the image on wikiHow is removed as well. This entire process is automated, except for when we have to talk to the people who uploaded the unlicensed images. There is also a second part to the same script. Say that a person uploads an image s/he found on Google, and chooses some random license - or s/he says that they created the image themselves. Every once in a while, the automated script looks at the 100 most recently uploaded images or so and cross-checks them against online databases, for possible copyright violations. If multiple sources are found for the image (which is often the case for “stock” images) then the script tags the image and the image is placed into the “Copyright Issues” category. This is all that the script does; a human being must check the images in this category to see if the image is truly a copyright violation or not. Since scripts can make mistakes, it ONLY tags the image for a real person to verify…it does not go around deleting images that it thinks are copyright violations, because there is always a possibility that the bot could be wrong. About 75% of this process is automated, and it has been around for at least 6-8 months now. We have only had one issue with it, which was resolved quickly, and that was months ago. It seems to be a very efficient way to deal with unlicensed images and copyright violations, and it saves human time for things that only humans can do, like writing articles. Kudos to @SudoKing
for his time and effort in developing and testing this script. Also a big thanks to the people who help make it run and respond to the uploaders who receive licensing notifications. In the specific case listed above, because the images were uploaded over 6 months ago, the script did not have the opportunity to check them for copyvios. I am pretty sure that if these same images were uploaded today, the licensing bot would pull them out and place them into the Copyright Issues category for human review.
system
3
I would heartily give thanks to @SudoKing
for this and many other scripts around wikiHow, such as the related wikiHow links and the kiwiHow bot. Being the image overlooker for years by myself (before any bots were created) … this has got to be the best thing eva. It does calculate precisely one week, it does give people the proper talk page message, along with specific image links, it uses a real person’s account to remove them from the server after that specified time, etc.
Ttrimm
4
Without a doubt! @SudoKing
ROCKS!
@SudoKing
now that’s what I call some real serious volunteer work!
And when all of you get to know him we all know him to be only a teenager(at this time), and with that much experience to produce these things… That’s one brilliant user!
Thanks @Maniac
, very interesting/useful.
system
8
@krystle
: that isn’t always true. My one cleaning program cleans my computer blindly. When it does this, the computers cache and all saved changes are reverted and this change was one of them it undid. (CCleaner is my cleaning program)