I've been asked by some of the editors on dawiki of the possibilities to create a set of questions that can lead to the correct image license(s) for image uploads. Furthermore, there is a request for enforcing a name for the photographer, so proper crediting can be done.
Looking at SpecialUpload.php, there are no hooks that is really suitable for this, and also I'm a little unsure about any effects of beginning to send people to new pages mid-upload, which would happen if added to any of the two existing hooks.
So I propose to add two hooks: One for adding fields to the initial form, and one for validating the contents before proceeding with the upload. Is there any reason that validation of license information should not be done in this way? Also, if a more suitable way of "quizzing" people exists, I'd be happy to hear about that.
On 29 Jun 2006 14:21:46 +0200, Anders Wegge Jakobsen wegge@wegge.dk wrote:
I've been asked by some of the editors on dawiki of the possibilities to create a set of questions that can lead to the correct image license(s) for image uploads. Furthermore, there is a request for enforcing a name for the photographer, so proper crediting can be done.
Looking at SpecialUpload.php, there are no hooks that is really suitable for this, and also I'm a little unsure about any effects of beginning to send people to new pages mid-upload, which would happen if added to any of the two existing hooks.
So I propose to add two hooks: One for adding fields to the initial form, and one for validating the contents before proceeding with the upload. Is there any reason that validation of license information should not be done in this way? Also, if a more suitable way of "quizzing" people exists, I'd be happy to hear about that.
-- // Wegge
I think it's a good idea, but IMHO the "other" answer should stay. People who want to add unauthorized copyrighted pictures use it most of the time... and that makes it very easy after to sort and remove. If the "other" or "I don't know" is removed, I think those people will try to add it selecting a wrong licence, very difficult to check.
Plyd
On 29 Jun 2006 14:21:46 +0200, Anders Wegge Jakobsen wegge@wegge.dk wrote:
I've been asked by some of the editors on dawiki of the possibilities to create a set of questions that can lead to the correct image license(s) for image uploads.
Well, damn me if I wasn't thinking of that just today. I was even working out a flowchart. In fact, I started a sort of rudimentary flowchart at http://en.wikipedia.org/wiki/Wikipedia:Can_I_use_this_work%3F a while back. Of course, Danish law probably differs significantly in the specifics.
On 6/29/06, Plyd wiki.vincent@amplyd.com wrote:
I think it's a good idea, but IMHO the "other" answer should stay. People who want to add unauthorized copyrighted pictures use it most of the time... and that makes it very easy after to sort and remove. If the "other" or "I don't know" is removed, I think those people will try to add it selecting a wrong licence, very difficult to check.
I was thinking of having that exist but lead to a "copyrighted" answer (i.e., if you aren't sure we just assume the worst every time). If we're talking a software-level implementation, though, a abortive upload should be logged: "User:Whatever attempted to upload a file named XXX.jpg, but stopped after 4 steps. [some brief gibberish here to allow those in the know to see what he answered to the questions]". If the user initially gets to a "copyrighted" result and then backtracks and claims he took it, you'll know something's up.
On 6/30/06, Simetrical Simetrical+wikitech@gmail.com wrote:
I was thinking of having that exist but lead to a "copyrighted" answer (i.e., if you aren't sure we just assume the worst every time). If we're talking a software-level implementation, though, a abortive upload should be logged: "User:Whatever attempted to upload a file named XXX.jpg, but stopped after 4 steps. [some brief gibberish here to allow those in the know to see what he answered to the questions]". If the user initially gets to a "copyrighted" result and then backtracks and claims he took it, you'll know something's up.
I don't like it. It reeks of spying, and it's not the sort of thing we would be proud of telling everyone about.
Steve
On 30/06/06, Steve Bennett stevage@gmail.com wrote:
I don't like it. It reeks of spying, and it's not the sort of thing we would be proud of telling everyone about.
Concur. People fuck up all the time, god knows I've done it enough. We don't need to keep a log of their "almost" fuck-ups as well.
Rob Church
On 6/30/06, Rob Church robchur@gmail.com wrote:
On 30/06/06, Steve Bennett stevage@gmail.com wrote:
I don't like it. It reeks of spying, and it's not the sort of thing we would be proud of telling everyone about.
Concur. People fuck up all the time, god knows I've done it enough. We don't need to keep a log of their "almost" fuck-ups as well.
Better turn off the ability to view revisions then...
The ability to scour image page histories for people switching licenses after the fact has been instrumental to detecting dishonestly tagged material on enwiki (people uploading content found on the web, tagging some website, seeing the notice, changing to non-commercial, seeing the notice then finally selecting pd-self, for example). If we are to provide any 'quiz' to users about the copyright status of their images, we should store the intermediate results so that searching for the right answer.
If there is concern that people will see your mistakes and criticize you for them thats a social problem.
On 30/06/06, Gregory Maxwell gmaxwell@gmail.com wrote:
If there is concern that people will see your mistakes and criticize you for them thats a social problem.
I'm talking about not storing information which crosses the borderline from needful retention to smelling off, to me.
If someone almost uploads something which would infringe copyright, but doesn't, do we *need* to keep it? I'm inclined to state not, unless it becomes a frequent occurrence. And who would have access to this information, and under what conditions and policies?
Rob Church
On 6/30/06, Rob Church robchur@gmail.com wrote:
On 30/06/06, Gregory Maxwell gmaxwell@gmail.com wrote:
If there is concern that people will see your mistakes and criticize you for them thats a social problem.
I'm talking about not storing information which crosses the borderline from needful retention to smelling off, to me.
If someone almost uploads something which would infringe copyright, but doesn't, do we *need* to keep it? I'm inclined to state not, unless it becomes a frequent occurrence. And who would have access to this information, and under what conditions and policies?
Eh, I see it the same as "If someone tags an image with {{forbiddenlicense}} and then changes it, why should we keep a record of their initial 'mistake'?"
The initial comment that started this subthread was justifyed with "If the user initially gets to a 'copyrighted' result and then backtracks and claims he took it, you'll know something's up.".
Any solution which could be equally used to teach a new uploader what lies to tell to keep his image kept really should write a record. Would you like to see some examples of deleted copyright violations on enwiki which would have been undetected if not for the ability to discover where an uploader has cycled through license options?
Perhaps an alternative would be a system which doesn't tell the user the results of their quiz until after the image is uploaded, then we have no situation where a change of mind before the upload results in a record.
On 6/30/06, Gregory Maxwell gmaxwell@gmail.com wrote:
Perhaps an alternative would be a system which doesn't tell the user the results of their quiz until after the image is uploaded, then we have no situation where a change of mind before the upload results in a record.
That would be fine too. It would probably be the easiest choice, actually; I think I was semiconsciously imagining something Ajax-y, where you had to click through dialogs and fetch stuff from the server, but unless the list of questions is *really* long, it would probably be more efficient to just send all of them at once. In that case, I can see why people would be weirded out by the idea of clicking the wrong radio button for one second and then changing it back before submitting, only to have their screw-up be recorded forever.
On 6/30/06, Gregory Maxwell gmaxwell@gmail.com wrote:
Better turn off the ability to view revisions then...
No. We are totally upfront about the fact that when you press "save" your history is being recorded. Tracking what clicks someone makes before they press save is, well, wrong, deceitful, pointless, ....
Anyway this is the wrong direction to be heading. There are countless ways we can massively reduce the addition of good-faith copyright violations, by restricting uploads to registered users etc etc etc. Focussing on the few bad eggs who deliberately upload copyrighted images is wasting our time...
Steve
On 30/06/06, Steve Bennett stevage@gmail.com wrote:
Anyway this is the wrong direction to be heading. There are countless ways we can massively reduce the addition of good-faith copyright violations, by restricting uploads to registered users etc etc etc.
Education, education, education. But with less Blairite incompetence and more up-front, honest, "look dude, we can't accept this, and here's why".
Focussing on the few bad eggs who deliberately upload copyrighted images is wasting our time...
Well, it's not a waste of our time to stop them. Frequent, persistent uploading of copyright infringing material is blockable, no? If it's not, then why the hell not?
Rob Church
On 6/30/06, Rob Church robchur@gmail.com wrote:
Well, it's not a waste of our time to stop them. Frequent, persistent uploading of copyright infringing material is blockable, no? If it's not, then why the hell not?
Oh, stopping them, no. But trying to sneakily trick them into leaving a clue which helps us catch them - when they amount to such a small percentage of all the bad uploads - we're not at that point yet.
Steve
Rob Church wrote:
On 30/06/06, Steve Bennett stevage@gmail.com wrote:
Anyway this is the wrong direction to be heading. There are countless ways we can massively reduce the addition of good-faith copyright violations, by restricting uploads to registered users etc etc etc.
Education, education, education. But with less Blairite incompetence and more up-front, honest, "look dude, we can't accept this, and here's why".
Focussing on the few bad eggs who deliberately upload copyrighted images is wasting our time...
Well, it's not a waste of our time to stop them. Frequent, persistent uploading of copyright infringing material is blockable, no? If it's not, then why the hell not?
Rob Church
Hoi, I would not be surprised that a fair percentage of dubious content comes from a group of people who persist in their habit of seeing everything as fair game. Blocking these people after the appropriate first steps leading up to it makes sense, it would make better sense if this were not only true for Commons but also for their "home" project. This is where they would really feel the pain of withdrawal.
In a similar way, people who troll on other projects should at some stage feel the pain in their own back yard. Thanks, GerardM
On 6/30/06, Steve Bennett stevage@gmail.com wrote:
On 6/30/06, Gregory Maxwell gmaxwell@gmail.com wrote:
Better turn off the ability to view revisions then...
No. We are totally upfront about the fact that when you press "save" your history is being recorded. Tracking what clicks someone makes before they press save is, well, wrong, deceitful, pointless, ....
Eh, then make sure the quiz doesn't let you know what you did wrong until after you save.
Anyway this is the wrong direction to be heading. There are countless ways we can massively reduce the addition of good-faith copyright violations, by restricting uploads to registered users etc etc etc. Focussing on the few bad eggs who deliberately upload copyrighted images is wasting our time...
You're taking a very black and white look at this... it's not as simple as good-faith vs bad eggs.
It would appear that there is substantial cognitive dissonance between the innate ethics of Joe Sixpack and the requirements of copyright law.
In particular, it appears very common for people to be quite convinced that it's okay to upload anything they were able to obtain off the web at no cost, and also fairly common for people to be convinced that they have complete ownership of things like screenshots they made of movies. This is complicated by the fact that numerous violations and borderline cases exist, so Joe can go point out other cases where images have not yet been deleted which are similar to his.
So you end up with people who would be described as honest and trustworthy who think nothing of iterating through a set of options in order to find one which doesn't cause the computer to say that their image will be deleted.
It's a complicated problem. Personally I'd like to make new users be dropped into a live chat with someone experienced at the time of their first few uploads to consult with them on the image... but scaling that would be a challenge, and I suspect the idea would go over like a lead balloon. "Oh no! it reduces openness!".
Oh, and if you weren't aware. Uploads are already restricted to registered users.
wikitech-l@lists.wikimedia.org