[Foundation-l] Whishlist from the National Library in Norway

John at Darkstar vacuum at jeb.no
Tue Jul 14 16:40:51 UTC 2009

I got an email from National Library in Norway and it has some
interesting points. My comments are after the bulleted points. The
bulleted points are my writeup from their comments, the original email
was in Norwegian. Hopefully it is understandable.. :D

* Backlinks to the museums themselves is important, especially so the
museums can serve high resolution images or alternate images.

Note that in this context there could be images with extremely high
resolution, or processed images that could be part of some continued
work. Imagine a 2D image of a viking helmet that really are a 3D scan,
the 2D visualization is really a simplification and the 3D scan can be
refined with new processing. The viking helmet is an example and is
located at Kulturhistorisk museum.[1] Medium resolution images are by
one source described as images in the range 800-1000px across longest
axis. High resolution images could be 25Mpx and higher, one source was
talking about 150Mpx. A 3D building scan combined with photographic
textures could be very much more than this.[5]

This is also important for us, how can we tell our users that the
museums can provide additional services? To post a template isn't very
dynamic, yet it somehow solves the problem. What would be very
interesting is to make some kind of API that makes it possible to get
additional information directly from external sites. Probably something
like this should go through a white list of some kind or be sufficiently
laundered to make it safe in our environment.

* Send a message to a contact if a specific image is used on Wikipedia,
probably also containing metadata.

This kind of service probably should be some kind of RSS feed with an
additional option of en email notice. Probably it should be possible to
follow a RSS feed for a whole category, like "Images from National
Library of Norway" or "Photos by Axel Lindahl".[2] Such a RSS feed
should probably be available in a daily or weekly digest mode. But what
if there is some sub category, what should then be sent as a message -
"This category and 1, 2, 3 levels below?"

* Geotag from Wikipedia should be available through Commons, and
additional tools for adding geotags are important.

Probably geotags from Wikipedia should be available through some kind of
API for Commons, but this should not be confused with geotags added to
the actual image itself. Changes to the Geotags should be available
through some kind of RSS feed too. It seems like tools to geotag an
image through features in the image is important, but I don't know if it
is feasible to do that today. I know about a few algorithms that can do
this, at least if they have some clue about where the photo comes from.
A wild guess "Norway" will most likely fail, but it isn't unlikely that
a general area of a municipality and a few additional features like the
location of a church and a couple of hilltops is sufficient. I know
Riksantikvarsembetet in Sweden and ABM-utvikling in Norway have talked
about this, so some interest definitely exist outside our community.

* Users should be able to give comments about an image or tag it, and
this should be reported back to the owner. This should be independent of
the use of the image at Wikipedia.

This I guess is two different thing. One thing is use of the image in
mashups, use in blogs, etc. For now you has to transfer the image to be
able to tag it on a blog, but what they say its interesting to get the
actual tagging on Commons and then the image should be mashable. The
other thing they talk about are comments on the image itself, which I
guess is simply an RSS feed from the talk page. They make an example
about Flckr.com, but I guess Expono.com is a better example. They said
that the total information from a wiki are probably more interesting for
them than Flickr.

Automatic tagging due to reuse is very interesting. How can we do this?
It works like trackback in blogs, but would probably mean that we allow
reuse of images through mashups. Imagine buttons like "post this on
flckr" and similar sites, and make the information about where the image
is reused available.

* They wonder how and if uncategorized images could be utilized somehow.
They make an example of several thousand photos taken by a German
soldier during WW2 in Norway, and this is about the only thing they know
about the images.

Perhaps we need some statistics on usage of the uploaded archives from
the German museums? Some of the images did not have sufficient
information for localizing them, yet they were later located. One
example is a photo from Mehavn. [3][4] I've been wondering if a service
like Expono is better suited and that we can transfer images when there
are something about them that makes them interesting for us. As long as
we have no information it is difficult for us to utilize the image, but
when we do utilize them we need to add a lot of information and that
isn't easy to do in an external interface.

As a side note, what if we make a dirt simple interface for adding
locations to images and then let the collaborative effort filter whats
usable information? Its something like "add the name of the place" and
only after more than one user adds the same name it shows up in a list
of possible places, perhaps overruled by a list of known locations. Note
that this should be done such that it doesn't create additional workload
for administrators, and that more evolved comments could be placed on
the discussion page.

* They wonder if an involved user would be interested in looking through
the images.

I believe it is better to make the images available to the community
because someone has the knowledge about the images. It is more likely
that the user with knowledge about the images will find them, than for
us to find the user with the knowledge. But if someone find a single
photo of interest, how can we use that information to find other images
of interest? I think it could be interesting to upload a bunch of images
where we have very little prior knowledge and then do some research on
how the images are utilized. Right now we have very little knowledge
about how such images are used. A project that uploads images without
any information and then let the community tag and use them will
probably take several years before any usefull information is gained, -
but even if the process is slow it will persist and there will be
available information over time.

John Erling Blad /jeblad


More information about the foundation-l mailing list