[Makers] Cameras in the space? Privacy rules?

Adam Thomas adam.lloyd at gmail.com
Mon Jan 30 17:38:19 EST 2012


On 30 January 2012 16:35, Lachlan Horne <wil5oncle7us at gmail.com> wrote:
> I put up some cameras around the space and intend to add more - some of you
> may have seen the timelapse video of Australia Day hacking at
> http://www.youtube.com/watch?v=1xEO85lgGEg

This is a really good example of why it's cool to have cameras around the place.

> We need to decide how these cameras are to be used. At the moment, the
> cameras are focused on narrow areas (the welding area and an electronics
> bench) and images are recorded in such a way that only I have access to
> them. This isn't going to work with more cameras looking at wider areas,
> from a privacy perspective.
> So here's what we've come up with so far:
> - Signs should be up around the space, with lights which come on when
> cameras are recording

Agreed. It needs to be really obvious which areas of the space are
being recorded at any given time.

> - Cameras should be recording when the space is closed for security purposes

Sounds good.

> - Cameras should not be recording when the space is open, except when we
> explicitly set them up to do so for special events

I think having all the cameras recording at a time could and should be
used sparingly with a notice being posted that an event is happening
where that would be taking place.

I think it would be really neat for someone to be able to tell the
server to record a video or an image at a (tunable) frequency (like
you have for the timelapse) to record builds. That camera would have
it's corresponding lights on the info board illuminate so people know
that if they go within its view they will be recorded.

> - Anyone not willing to be recorded should have the option to stop the
> recording - one idea is a "kill switch" next to the sign in sheet, a big
> red button which stops all the cameras (and probably should delete the last
> few seconds of images to be sure?).

While I'm the one who originally proposed this, on reflection I don't
think it would actually be a good idea, at least not a 'kill all the
cameras' button. If someone had been working on a time lapse of a
build they'd be pretty miffed if someone cut it off, especially if
that concerned person never went in to the field of view of that
camera. As long as, at any time outside of the global "we're recording
everything", there is somewhere people can work and hang out without
being recorded I think it'll be fine.

> We should probably not run live publicly-accessible streams. If we do, they
> should only be on when recording is on as per the above rules.
> Not sure how access to recordings should be controlled. Committee should be
> able to look at them for security. We should also be able to go back
> through them to compile promos like the above timelapse. People who are
> recorded should be able to access the recordings as well. Should we just
> make these recordings publicly available on request?

I'd make the recordings from when the space is closed available in a
general share on Morphia and then when a user/member requests a
video/interval they should go in their home directory (under some user
configurable location [yes, I'm volunteering to do that bit]).

Recording all cameras for the whole time the space is closed isn't
feasible, we just don't have that much disk space. We can how ever use
packages like Motion[0] to trigger recordings when there's movement.
Telling Motion to stop monitoring the cameras is an exercise in bike
shedding so let's build that bridge when we need it.

[0]: http://www.lavrsen.dk/foswiki/bin/view/Motion/WebHome

> There is a lot we can do with derived data. How we handle that depends on
> the specific form of the data, I guess. Here is an example of derived data
> which I see no problem with having live and public:
> http://apps.pumpingstationone.org/space-activity/
> That is, using camera feeds to derive a numeric "activity level" for the
> space. The only info recorded is "stuff is going on at time x". More
> identifying info should be more restricted.

The way that PS:one is anonymising the data is good.

With the 3d from 2d tech you've been playing with (or just some
birds-eye-view placements) I think it would be possible to generate a
heat map of the space which would be utterly cool and something I'd be
interested in helping with.

> Anyway, this is where the discussion stands now. Share your thoughts!

Shared. :)


> _______________________________________________
> Makers mailing list
> Makers at makehackvoid.com
> http://www.makehackvoid.com/mailman/listinfo/makers

More information about the Makers mailing list