having Google help Wikimedia Commons

classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

having Google help Wikimedia Commons

David Monniaux
Howdie,

As you probably know, administrators and other volunteers on commons
(and other wikis) wage a constant battle against people who upload
photos from web sites under bogus claims ("I took it myself!" etc.).

One thing that would be nice would be a tool to check whether and where
an image file is available somewhere else on the web.

I think we can get Google to help us in that matter.

Google Images, in order to build its thumbnail database, has to download
the files and compute thumbnails. They can, at the same time, compute a
hash of the file (SHA, MD5 or similar). Perhaps they already do.

If the hash is stored into the database, they can essentially answer our
problem. They already offer a SOAP programmatic interface (which does
not offer this feature); conceivably they could offer this "look for
identical files" feature, perhaps to selected partner sites.

Since they have offered us a hand in the past...

Regards,
DM

_______________________________________________
Wikitech-l mailing list
[hidden email]
http://mail.wikipedia.org/mailman/listinfo/wikitech-l
Reply | Threaded
Open this post in threaded view
|

Re: having Google help Wikimedia Commons

Jake Nelson-2
David Monniaux wrote:
[...]
> One thing that would be nice would be a tool to check whether and where
> an image file is available somewhere else on the web.
>
> I think we can get Google to help us in that matter.
[...]

This is a very good idea... I can see "Search for this file on the web"
being extremely useful to Google's userbase in general (watching the
spread of certain "viral" images, seeing where people are using images
you authored, also finding copyright violations). This should definitely
be proposed to them.

-- Jake Nelson
_______________________________________________
Wikitech-l mailing list
[hidden email]
http://mail.wikipedia.org/mailman/listinfo/wikitech-l
Reply | Threaded
Open this post in threaded view
|

Re: having Google help Wikimedia Commons

Evan Martin-2
In reply to this post by David Monniaux
On 1/13/06, David Monniaux <[hidden email]> wrote:
> One thing that would be nice would be a tool to check whether and where
> an image file is available somewhere else on the web.
>
> I think we can get Google to help us in that matter.

I'll look into it.  (No promises, but the idea seems sound...)
_______________________________________________
Wikitech-l mailing list
[hidden email]
http://mail.wikipedia.org/mailman/listinfo/wikitech-l
Reply | Threaded
Open this post in threaded view
|

Re: having Google help Wikimedia Commons

Tels
-----BEGIN PGP SIGNED MESSAGE-----

Moin,

On Saturday 14 January 2006 03:03, Evan Martin wrote:
> On 1/13/06, David Monniaux <[hidden email]> wrote:
> > One thing that would be nice would be a tool to check whether and
> > where an image file is available somewhere else on the web.
> >
> > I think we can get Google to help us in that matter.
>
> I'll look into it.  (No promises, but the idea seems sound...)

While the bit-hash can easily defeated (just alter one bit), the idea is
interesting - a lot of images are very probably copied without any
alteration at all.

Computing a watermark or something alike that is probably to
CPU-intensive, but it would allow loose matching (e.g. someone adds his
own (C) on an image, but leaves 99% unaltered, etc).

Best wishes,

Tels

- --
 Signed on Sat Jan 14 10:54:53 2006 with key 0x93B84C15.
 Visit my photo gallery at http://bloodgate.com/photos/
 PGP key on http://bloodgate.com/tels.asc or per email.

 Marketing lesson #1: The synergy of the result driven leverage can
 *never* incentivize a paradigm shift. --  Walterk (124748) on 2004-01-16
 at /.

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iQEVAwUBQ8jK53cLPEOTuEwVAQHp5wf9G/duqnyiAToRCbv58mLF4nEXplg8RkmV
7M+iPJyg6MOZU1MnzToX/p7yyyM4a/XKLWXMRE0TorggfeEOx1vvHYAMSsS81rT/
i6XLIlKUwaK+rIEhxyzIAQJfPYXcuOCAwJ6c0wUmsmSNFKY5VP9t7WFP4ftS6Edf
74TWOQotUW5CujeJRYShe8w15RlXhAuTHzzc1ENmBliuOEBbTy21DEDYoP3OgJwb
EwAzbTHlUPI0wSbDkXRd80p0/fsFReq9l8NeT0YbFin7RnYPOWTgVM+PAv0j8YEL
v/Mxq7IVXuAA9dKhXBRUQK7IMbL2b2/b2yTqym2DDT5obF8yfIDKmg==
=APAg
-----END PGP SIGNATURE-----
_______________________________________________
Wikitech-l mailing list
[hidden email]
http://mail.wikipedia.org/mailman/listinfo/wikitech-l