[Foundation-l] GLAM-WIKI report

[Tim Starling]

I thought I'd better write up a report about the conference I went to
last week, to justify the time I spent there. I'll give some general
observations followed by some technical ones.

GLAM-WIKI was a two-day conference billed as a meeting between
Australia's GLAM sector (galleries, libraries, archives and museums) and
Wikimedians. GLAM representatives outnumbered Wikimedians, but we had
enough people there to make sure our point of view was heard both inside
and outside of the formal program. Many of the talks were from people in
the GLAM sector who were already converted to our way of thinking, and
who endeavoured to convert the rest of the GLAM audience by speaking in
their language.

The GLAM representatives were generally very receptive. When dissenting
questions came up, they were often answered in our favour by another
GLAM representative. I asked one of the delegates about this favourable
mood, and he said that the delegates were generally self-selected people
who had a favourable opinion of Wikimedia and free content, and that the
skeptics did not attend. However, the discussions had at the conference
would provide valuable ammunition against those skeptics back in the office.

As far as I know, only one speaker expressed a completely contrary
opinion to the general mood of the conference, and that was Ian
MacDonald of the Australian Copyright Council. He said, in essence, that
institutions need to prevent reuse or modification of the content they
hold in order to preserve its purity, which risks sullied by the
cumulative distortions of the general public. This was passionately
countered by Jessica Coates during question time, with some success
judging by nearby whisperings. MacDonald also warned the audience about
evil Wikimedians like the one who "hacked into" the NPG (UK) website and
stole a million pounds worth of images. The factual errors in this
statement were briefly addressed during question time.

I tried to get a feeling for what sort of hard drive capacity we would
need if the institutions in the room decided they wanted to share large
amounts of content with us. Many of them have tens or hundreds of
terabytes of data storage, in tape and hard drives. However, the bulk of
this is in restoration-quality images (e.g. TIFFs tens of thousands of
pixels wide), which they would not be willing to share with us even if
we wanted them. Liam Wyatt proposed as a business model or compromise
with management, the idea of sharing images of a 1000-2000 pixel width
and charging a fee for access to the full resolution images. That seems
like the most likely arrangement, and if so, it wouldn't need a
significant change to our current capacity planning for file storage.

A GLAM delegate expressed an opinion in question time that they would be
reluctant to have us mirror their collection, since they've spent a
large amount of money setting up their data storage, so mirroring would
seem like a waste. Brianna Laugher was receptive to the idea of having
Wikimedia projects hotlink or cache images from galleries. I kept quiet,
the significant technical challenges with that approach were not discussed.

There is a need for bulk upload tools to be better advertised and more
readily accessible. One of the institutions reported paying students to
upload hundreds of photos to commons via the usual web-based UI, but
found it to be too time-consuming and expensive to consider on a large
scale.

Special:BookSources came up a couple of times. The libraries would love
to see software improvements, such as geolocation giving the ability to
present the nearest few libraries at the top of the page, without the
user having to click on the world map. Liam mentioned the geolocation
projects based on detecting nearby 802.11 access points. I think
MaxMind's GeoIP City would be a better as a software development
starting point.

Delegates from the National Library of Australia reported that they have
an ongoing project to collate collection metadata from all libraries in
Australia. It may be possible to replicate this data to Wikimedia
servers, or otherwise make it available. This would enable a feature
whereby the user is told which libraries have the book being searched
for, in the requested edition or a different edition. It may even be
possible to report whether the book is on the shelf or not.

-- Tim Starling