On 2/13/07, Platonides Platonides@gmail.com wrote:

Has the look of a commercial engine. Not sure if it's good or not, but seems 'profesional'. Instead of Copyright... Copyleft? ;)

But copyleft wouldn't be accurate for much of the media. I think that when this is 'done' the best thing would be to display license with the image... in due time, in due time.

On 2/13/07, Husky huskyr@gmail.com wrote: [snip]

• Having the license beneath the image would be very handy. Even more handy

would be an option to do an 'advanced search' for a specific license (e.g. only PD images)

PD! Sure lets just use the template links table to find all the images that transclude a template that starts with PD- ... Such as... {{PD-CAGov}} (http://commons.wikimedia.org/wiki/Template:PD-CAGov)

Hard problem. :(

Even our most machine readable metadata isn't very machine readable. :(

Brilliant solutions welcome.

On 2/14/07, Husky huskyr@gmail.com wrote:

I'm not a brilliant programmer, but isn't it simply a case of

1. having an array with all possible (or at least, the most popular)

licenses and their template tags

This is hard because you have to first make the list.. And you have to deal with examples like the {{PD-CAGov that}} I listed, and {{GFDL-joeblow}} (personal license templates). The list of license templates are growing every day.

2. parsing all {{template}} tags

Thats not enough because people often use licenses through wrapper templates..

{{self|GFDL|cc-by-sa-2.0}} {{self2|GFDL|cc-by-sa-2.0}} {{template2|..

There are a whole bunch, they change too.. and you have to track them down.

3. matching the template tags with the licenses array and using that to

display the license in the Mayflower results screen?

Because there are so many different minor permutations, even if you could extract all the licenses, figuring out the mapping to something small enough to put in the interface is a challenge alone.

I don't think you have to start with *all* licenses (that would indeed be a very long list). Simply starting with some of the most popular ones you could capture a major part of the collection. Licenses like * {{PD-old}} * {{PD-Art-YorckProject}} * {{PD Art}} * {{self2|GFDL|cc-by-2.5}} * {{self2|GFDL|cc-by-sa-2.5,2.0,1.0}} * {{PD-self}} Are used by many pictures throughout the whole project. If you could only match these you could already display the license under a lot of the images.

And for more licenses you could take one of the main license categories such as http://commons.wikimedia.org/wiki/Category:Public_domain And extract all categories from there and use it for the 'license array'.

However easy it might seem, i'm not sure if this will work out. As i said before, i'm not a programmer :)

-- Hay Kranen / [[User:Husky]]

On 2/15/07, Samuel Klein meta.sj@gmail.com wrote:

On Wed, 14 Feb 2007, Gregory Maxwell wrote:

...

...

3. matching the template tags with the licenses array and using that to

display the license in the Mayflower results screen?

Because there are so many different minor permutations, even if you could extract all the licenses, figuring out the mapping to something small enough to put in the interface is a challenge alone.

This is a resolvable problem; the more minor the permutation, the smaller its effect on the image that shows up in the interface...

SJ

---

foundation-l mailing list foundation-l@lists.wikimedia.org http://lists.wikimedia.org/mailman/listinfo/foundation-l

On 14/02/07, Husky huskyr@gmail.com wrote:

Wow, this is really good! I can finally find something on Commons without having to click trough a list of images or browsing categories :)

Just some small recommendations:

• I would give the 'x results found' a colourful background (maybe purple)

to separate the search box from the thumbnail view. You could also move the search box to the right of the logo and use the dotted grey bar as a separator.

• Having the license beneath the image would be very handy. Even more handy

would be an option to do an 'advanced search' for a specific license (e.g. only PD images)

• Maybe the colours should be a little more Wikimedia-like (green, blue,

red) and the imagery could be more related to Wikimedia Commons to refer to the connection between the two. I attached an example logo (just an idea).

Anyway, great job! With some more work Mayflower could be a great replacement for the current Commons search.

I suspect it would already be vastly more useful to our average user than our default search :P

Yeah, I would put a by-line that says something like 'Search a database of over one million <b>freely licensed</b> media files'. This can be a great way to introduce non-Wikimedia people to Commons.

As a couple of people have commented, distinguishing between images/sound/documents would be useful too.

Anyway -- fantastic work, Tangotango! It looks fabulous and we will push it hard at Commons. And when you want some help developing that MediaWiki extension, well, you know where to find me :)

cheers, Brianna

On 14/02/07, Gerard Meijssen gerard.meijssen@gmail.com wrote:

Hoi, There are quite a few sound files and there will be many more given that some great functionality is developed in the "Shtooka" software. The notion that these files are not desperately useful for some.. is also true the other way around. For some the images are not desperately useful and it is the sound files that are of interest.

Hey, I didn't say "stop displaying them", I said "make it disableable" ;-)

[Most people doing a commons search are likely looking for images only; I suspect a large fraction of the rest are looking for non-images only. I wonder how many are intentionally looking for a mixture...]

Making this more fine-grained would be interesting, too - restrict by "image", and then provide the option to restrict by filetype. Hmmm.