Hi Laura,

You have the idea. There are a number of access points we'd like humans to
add based on space/time/location/use/visual elements in the photos
unrelated to the actual subject matter. There are a variety of approaches
that could be taken, and I've received helpful ideas offline on how to
proceed.

I'm not a fan of CYA policies, but I'm averse to adding the sort of tags
you removed because I cannot imagine how such tags wouldn't put our library
and institution in a very bad light while undermining organizational
priorities. However, the need for providing this sort of access is real so
we need to do something. The suggestions I received are mostly based on
restricting or obfuscating some metadata, and a solution along those lines
will probably be the ticket.

One particular idea that intrigued me was classification codes in a
specialized field. This provides a lot of display and search options as
what is displayed can be very different than what is stored/searched --
i.e. people could search in plain English and the results would appear
without it being obvious why the search works (and hopefully they wouldn't
wonder). The other thing I like about it is it's easy to eliminate once the
need for it disappears.

Less sensitive stuff is more straightforward. My gut reaction is that
regular tags stored as multiword non-tokenized strings (to prevent
pollution of search results and the subject index) might be a good
approach. But since many libraries have needs similar to ours, I thought
I'd ask as I'm sure a lot of people have given this issue more thought than
I have.

kyle



On Fri, Jan 8, 2016 at 8:44 AM, Laura Buchholz <[log in to unmask]>
wrote:

> Kyle, I don't know if I'm understanding your question correctly, but I
> think this is something I was just reviewing. I removed "Diversity" as a
> subject term (we're a little loose here in applying subjects terms that
> aren't directly in the photo) from some photos that were of, for example, a
> single student studying on the lawn or in the coffee shop. The diversity in
> the photo was that the student was of color. When there is an image of a
> white student, we wouldn't put "homogeneity" or something like that, so I
> took off "diversity". But, as you say, users do want to be able to search
> for these concepts, and I think it is important not to erase differences
> just because it is difficult to represent that without being essentialist
> in metadata.
>
Are you trying to automate this process, or are humans doing this? If
> automated, watch out for what Google Photos did:
>
> http://www.usatoday.com/story/tech/2015/07/01/google-apologizes-after-photos-identify-black-people-as-gorillas/29567465/
>
>