how private should private be?

[Ken-ichi]

Currently when you set the geoprivacy of an observation to "private"
we don't show the observation on maps, we don't show the coordinates,
and we don't show the text location, but we do use the coordinates to
populate checklists if the observation becomes Research Grade, and we
do allow people to use the Identotron, which lets you infer the
general location.

I'm personally in favor of "private" meaning absolutely private, e.g.
only the observer can see their own location when the geoprivacy is
private, as well as project curators they allow to see them, but no
one else, but others on the team would prefer for "private" to mean
"not on the map" and to still show some vague sense of location even
if we don't show coordinates, e.g. county / state / country, so at
least people who are trying to identify things can see roughly where
it was observed.

I don't think either side feels that strongly, but what we really want
is consistency. What do you guys think "private" should mean?

[Charlie Hohn]

I think state level. Maybe county. And I definitely think we should remove the thing that emails the location of private observations to people who follow you. 

--
You received this message because you are subscribed to the Google Groups "iNaturalist" group.
To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist...@googlegroups.com.
To post to this group, send email to inatu...@googlegroups.com.
Visit this group at https://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

--
Sent from Gmail Mobile

[Ken-ichi]

Just to nip any alarm in the bud, we were *not* emailing actual
coordinates, just the location description, and nothing we weren't
already showing publicly on the calendar, BUT, we're not doing any of
those things any more.

If we do adopt the partial privacy approach, it'll be more like the
political body that contains the publicly visible accuracy circle,
e.g. if your obs is private, your coordinates are in Burlingame, VT,
but your positional accuracy is an area larger than Vermont, the
location will be displayed as "United States" or more likely "North
America".

[Charlie Hohn]

right... sorry, that wasn't a descriptive enough post. some of the location descriptions had road names and such. and I'm glad to hear it isn't happening any more! 

============================
Charlie Hohn
Montpelier, Vermont

[Christopher Tracey]

I like the idea of a vaguely sense of location (state) attached to the private observations. 

[James Bailey]

I don't usually use this particular tick box on my observations, but as an identifier seeing the country at least is really a must.

[David K]

This is a regular point of discussion with potential new iNat users in my area, as many herp and butterfly observers are concerned about collectors (and sloppy observers who may ruin habitat). There is a strong bias with the more serious naturalists to keep local descriptions and map coordinates private for threatened/endangered etc species.  I think that country and state/province should be a minimum disclosure level in a public database, and I believe that county-level may be appropriate.  Providing county locations in Ontario is common but Ontario counties tend to be very large and stating that a Blandings turtle was seen in a specific county would hardly help a collector here. However, identifying the specific body of water associated with that turtle would be problematic.

The follow-on question is what does GBIF do with these observations?  Its not much good if iNat obscures something but then the data is freely available on GBIF.  But I am a big fan of GBIF and I think that providing GBIF with data on threatened species is positive, I just don't know if it is a weak link in terms of the privacy issue.

David

[BJ Stacey]

Personally, I don't identify anything that isn't at least to county level.  I don't see a problem with having something obscured set to that level.  It is quite easy to find that information on other places on the internet if you try.

One problem I do see is that people will add fields to an obscured observation so that they can share the sighting with a project.  Some of these projects require a field that is site specific, so even though the accuracy is set to 1000 acc I can tell that it was on a certain river.  Fairly easy to figure out from there.  Should these fields also be hidden?  Where does it end?

So, I think that county level at a minimum should be kept.

BJ Stacey

Follow my outings on iNaturalist.org
http://www.inaturalist.org/people/finatic

--

[Charlie Hohn]

Complicating factors is that some counties out West are larger than some eastern states, other counties are tiny. I'd say that county level makes sense in California, but in Vermont state level is fine (I don't have a problem with county either)

[Ken-ichi]

Let's not get bogged down in what rank in the hierarchy of political
boundaries we show. As I tried to explain above, we will probably just
choose a set of place names for places that include the observation's
area of uncertainty, so that might include a county and it might not.

I'm actually more interested in the perspective of observers than of
identifiers, since it's their privacy that's at stake here.

We don't send anything to GBIF that isn't publicly visible on the
website, so that's not going to change. If you can't see the
coordinates on iNat when signed out, you should not be able to see
them on GBIF.

[Charlie Hohn]

Thanks Ken-Ichi, I know this is a difficult topic and i am among the most annoying people when it comes to this too. Usually when I am asking for something weird it's because I'm trying to expand the extent to which I can use the site for more in depth scientific stuff. Which I think is awesome but I know can be mission creep too.

[Greg Lasley]

I'll just add a late comment as I have not been part of this discussion previously. I come across a number of iNat submissions of fauna where knowing just a hint of where it was taken will support me making an ID. Some species pairs such as Desert and Duckweed Firetail (and there are many others) are virtually identical in photos, but if I know it was taken in Arizona it is not Duckweed, and I know if it was taken in Georgia it is not Desert. I bump into this all the time. I will usually post a comment asking for just a hint of where the image was taken. Sometimes folks reply, other times they do not. I would certainly love to at least know the part of a state or perhaps county where an image was taken if I am going to contribute an ID. Anyway, my 2cents.

[James Bailey]

Given the lack of observers posting here that may be affected by changes in these features so far, can we reasonably assume that country level is acceptable for the time being? Unless you are a fugitive trying to completely conceal your location I really don't see any benefits to hiding the country. Privacy issues such as private land, sensitive areas, and such only seem to be at risk in terms of local or county designations. Saying your very endangered plant was found in North America is really not going to hurt anyone.

Of course we can't make too many changes to "Private" as then it'll become too similar to "Obscured".

[James Bailey]

On Friday, July 15, 2016 at 3:17:43 PM UTC-7, James Bailey wrote:

Given the lack of observers posting here that may be affected by changes in these features so far, can we reasonably assume that country level is acceptable for the time being? Unless you are a fugitive trying to completely conceal your location I really don't see any benefits to hiding the country. Privacy issues such as private land, sensitive areas, and such only seem to be at risk in terms of local or county designations. Saying your very endangered plant was found in North America is really not going to hurt anyone.

Of course we can't make too many changes to "Private" as then it'll become too similar to "Obscured".

Oh uh, let's all please forgive me for including a continent in my example of a "Country". But you know what I mean.

[Jessica Moreno]

In the Madrean Sky Islands we can occasionally have records (usually tracks) for jaguar, ocelot and Mexican gray wolf. Poaching a huge issue with these species, so we are very careful about location privacy (for the most part, these records come to us outside of iNat anyway). In addition, we often conduct BioBlitzes on private ranches in Sonora, Mexico whose owners wish to keep some things obscure/, and when the property is 10,000 ha or more, that can be challenging. County and state level information should be OK. 

[Cullen Hanks]

I vote for private to county or similar geographic unit.

Cullen

Sent from my iPhone

[Cullen Hanks]

Yes, and some of our counties in Texas are bigger than some states...

Sent from my iPhone

[Charlie Hohn]

That's going to be really hard to generalize. There are some counties in Vermont where it really is too small an area for it to work well for private observations. Conversely though, going to state level is absurd in places like Texas or California.

[Cullen Hanks]

If you wanted to be consistent, perhaps you could draw the line at a specific land area, say 200 sq miles or 500 sq kilometers.  This would allow iNat to be consistent across the board. 

For perspective, the average size of a county in the United States is 997 sq miles or 2,584 sq kilometers.  However, in Road Island, the average size of a county is 206 sq miles.

https://en.wikipedia.org/wiki/County_statistics_of_the_United_States

I'm still for sticking with counties, but I wonder how this would work in other countries with different administrative units.  Having a set geographic size might provide understandable consistency.

Best,

Cullen

On Monday, July 11, 2016 at 5:32:40 PM UTC-5, Ken-ichi Ueda wrote: