New toy! Upon loading an image page, it will scan the categories of
that image, and, for each category, show the last and next one in the
order of "addition to the category" as a thumbnail. Sounds
complicated? Have a screenshot:
http://www.magnusmanske.de/wikipedia/gallerypreview.png
I put it under [[MediaWiki talk:Gallerypreview.js]]. Installation
instructions on the page:
http://commons.wikimedia.org/wiki/MediaWiki_talk:Gallerypreview.js
(Now that I read it, "gallerypreview" is not exactly correct; catchy
name, anyone?)
Besides the "pretty" factor, it has some practical uses:
* Shows number of images in the categories (spot one-image-only categories)
* Complains if there's no/only "trivial" (licensing etc) categories
The query.php interface won't give more than 500 results per category,
so last/next information on larger categories /will be wrong/. If you
find any more "trivial" categories, please let me know (or add them to
the script, it's simple, really!).
Have fun,
Magnus
> Date: Thu, 17 May 2007 13:07:59 +1000
> From: "Brianna Laugher" <brianna.laugher(a)gmail.com>
> Subject: Re: [Commons-l] Two new tools for Flickr
> To: "Wikimedia Commons Discussion List"
> <commons-l(a)lists.wikimedia.org>
> Message-ID:
> <d20d84ea0705162007r4cd29ba0kd38820d055c2b5b0(a)mail.gmail.com>
> Content-Type: text/plain; charset=ISO-8859-1; format=flowed
>
> On 17/05/07, Bryan Tong Minh <bryan.tongminh(a)gmail.com> wrote:
> > As you might, or might not know, I have been quite busy with Flickr
> > lately, especially with [[User:FlickreviewR]]. I have written two
> > tools (or actually, one tool with two functions) with helps Commons
> > users with images for Flickr.
> >
> > The first is a database of all images reviewed images from Flickr:
> > http://tools.wikimedia.de/~bryan/flickr/browse You can
> search on nsid,
> > username, photo_id, link, and Commons image. The database contains
> > over 28,000 images, which is over 70% of the total number of Flickr
> > images on Commons :) [1]
>
> Whose username, Flickr or Commons? And if Commons is that
> reviewer or uploader or what? What is nsid? No search I tried
> actually returned any results. Bit more help, please?
>
> > Now the second tool is really handy (imho ;P). It allows
> you to easily
> > upload images from Flickr:
>
> If you find any security bug in
> > the upload part, the bot that performs the uploads and to
> be blocked
> > is Flickr_upload_bot.
>
> Magnus had a similar idea, a bot that performed transfers
> from (eg) Wikipedia to Commons. I asked him to disable it...
>
> I kind of have a problem with this is in that it allows
> essentially anonymous uploads. At least in this case they
> are restricted to images from flickr with suitable licenses,
> that is better than totally anonymous, but still. What stops
> me putting the username 'Bryan' in and putting up whatever
> irrelevant, offensive, invasive, stupid images I can find on
> Flickr? oh... nothing.
>
> I think there's a good reason MediaWiki requires users to be
> logged in before uploading, and I don't think we should use
> bots that circumvent that requirement.
>
> At the very least I think there should be a bot approval
> thing for this bot, where we can discuss as a community if we
> want to allow this kind of thing to happen.
>
> cheers
> Brianna
> user:pfctdayelise
The idea of a tool to help bring images over from flickr and put them on
commons, properly tagged and with nice {{information}} templates filled in,
is very attractive. However toolserver restrictions prevent tools from
asking for passwords. That seems a bit of a conflict that will require some
thought to get round. I agree with Briana that such uploads, even when done
with tools, really ought to be verifiably associated with the real user, not
with a bot ID.
I think we might want to try to look for ways to do this
AWB for instance gets around this by being client side, and requiring the
user to log in. It's also heavily intertwined with windows machines, as it
relys on .net and on IE services...
I wonder if there are other ways to do this that are platform independent
but that do not require the user to give their commons password to a
toolserver hosted tool...
Larry Pieniazek
Work mail: lpieniaz at us.ibm.com
Hobby mail: lar at miltontrainworks.com
I can only agree
On 5/15/07, Larry Pieniazek <lar(a)miltontrainworks.com> wrote:
> Just needed to point that out.
>
> Larry Pieniazek
> Work mail: lpieniaz at us.ibm.com
> Hobby mail: lar at miltontrainworks.com
>
>
>
>
>
>
> _______________________________________________
> Commons-l mailing list
> Commons-l(a)lists.wikimedia.org
> http://lists.wikimedia.org/mailman/listinfo/commons-l
>
Hi all,
I'm slowly getting settled in at Cambridge, and as a warm-up I hacked
a little JavaScript that might be an alternative to the CheckUsage
script. Demo at:
http://tools.wikimedia.de/~magnus/commonsusage.html
It runs through a combination of project types (currently wikipedia,
wikibooks, wikinews) and languages (currently en, de, es, fr, pl, it,
sv), and queries the usage of an image through the API, which is
always up-to-date. Enter an image name and click "Search", and you can
see the list getting shorter as queries are completed (or longer if
there are many hits); IMHO this cuts down on the "subjective waiting
time" ;-)
Two drawbacks to this:
* It hits the API, which might slow things down if used large-scale
* Due to JavaScript security limitations, I have to go through a Perl
script on the toolserver for each query (ugh!)
However, it might be an in-place alternative to Duesentrieb's script
if installed on commons. Note that the perl script would also have to
be on commons.
Magnus
Hello all,
I was thinking about how we could test Commons' coverage compared to
other media sources such as Getty Images and Flickr. One way would be
to make a list of topics that we should aim to cover. Periodically we
can compare our coverage to other resources to see how we are
progressing.
Of course we accept that in some areas such as logos we will always be
deficient...
There are also several ways we can measure. e.g.
-Binary: Yes/No we have something (anything) on that topic.
-Quantity: Number of photos in the relevant gallery and/or category
-Quality: Number of QI/FP in the relevant gallery and/or category -
maybe something also about general believability of other licenses
(own work of trusted Wikimedians = great, {{copyrighted free use}} =
not so good)
Here are some topics I've thought of so far. What else? What might you
go to a media resource trying to find? (I've just kept it to images so
far, not audio or video.)
==Nature==
Animals
Plants
Weather and geography stuff (oceans, volcanoes, lightning)
==Science==
Planets/celestial bodies
...
==People==
Modern celebrities
Historical figures
==Stock photography==
Generic objects
Generic people
Generic activities (verbs)
Wallpapers
...
==Places==
World cities
World landmarks
==Misc==
Historical events
Types of food
Works of art
Symbols
Maps
Diagrams
Flags
So how would you evaluate it? Of course if we pick 10 specific animal
species, for example, to evaluate against, there is a danger that we
would just fix up any missing ones right after the first evaluation,
because of course that's the wiki way. So although some of these are
fixed (eg. World cities and landmarks), with 'animals' we could each
time pick a different random subset of 20 or so to evaluate against.
For the 'fixed' ones, well, we can hope quality and quantity measures
improve.
So as to whether topics should be covered, things that we could
potentially get images of should be included, even if it seems
unlikely, e.g. people. Things we could not host in accordance with
Commons:Licensing should not be included (eg maybe Eiffel Tower,
Ausralian Aboriginal flag, works of art that are still in copyright).
Basically it would be good to have this kind of evaluation, because
then we can say "look our coverage of this broad range of topics is
like this". It also gives us an idea of coverage gaps, where we need
to improve.
So ideas for topics we should aim to cover, and other resources
besides Getty and Flickr that we should compare against, would be
welcome.
I read that Lulu.com (vanity press/prints books on demand) has made a
licensing partnership with Getty images -
http://www.lulu.com/partners/getty/ "Lulu.com is working with Getty
Images to enable authors, creators and organizations to use licensed
content to enhance and embellish their original works, while
protecting and compensating the copyright holders. Portions of three
extensive collections — Stockbyte, Digital Vision and Photodisc — from
Getty Images' vast library of photos and illustrations will be
available to Lulu.com users." I wonder if this agreement is
exclusive...
Do you think Commons could pursue an agreement with a company like
this? Do you think Commons should? (I would put up a subset, like
QI+FP, to avoid bad licensing problems.)
Lulu does not really get the open content idea, I think (I had an
argument in their forum once with a guy who said freely licensed work
puts pro photographers out of business, or similar nonsense). So maybe
WMF could still set up an agreement with them and get some small
profit out of each use :) or maybe just a good example of the variety
of applications of open content.
Also there is a competitor http://www.blurb.com/ and AFAIK they don't
have any such agreement yet :)
cheers
Brianna
user:pfctdayelise
On 12/05/07, Casey Brown <cbrown1023(a)comcast.net> wrote:
> I think is a very good idea that we should evaluate our coverage and also
> see that there are many ways we could try do this! Binary/Quantity would
> probably be easiest, but I like your idea of "quality". It could be like a
> point system. However, the quality system might be hard to get other users
> who are looking at our data to understand.
Well, I think 'quality' is important because if we're comparing
ourselves to professional organisations like Getty, probably 100% of
their work is Quality. If we have 10 photos of something but they're
all amatuerish and blurry, that's not necessarily useful (but it does
depend on the end user's purpose - it's not to say they're useless,
especially if we're comparing against Flickr). Other measures of
quality might be image resolution.
David Gerard said:
>But, we have plenty of pro photographers who put images up on Commons
>as GFDL and also license commercial use. Can we collect a pile of
>them? Write a page on the subject?
Sure, I think that's a good idea too. [[Commons:Meet our
photographers]] (Perhaps another one [[Commons:Meet our creators]] or
[[Commons:Meet our vector artists]] - think of LadyofHats for example
- amazing work.) Have mini bios and photos and get them to write a
paragraph about why they license their work freely. That would be a
really cool way to reward the superb contributors who donate their
fabulous work for us. A) We say "look, we have professionals here" and
B) we promote them as well. if someone saw a Wikimedian's bio this way
and decided to hire them for some work, I would consider that quite
awesome.
We would need some criterion to keep it to the truly professional -
perhaps something like 5 or 10 (or 15?) FPs total at either Commons on
English Wikipedia?
Can someone run a database query over [[category:featured pictures]]
&& [[Category:Self-published work]], and rank the image uploaders?
then we will have an idea of a reasonable threshold I guess.
cheers
Brianna
