[ckan-dev] mandatory user registration?

Ian Ward ian at excess.org
Mon Jun 16 22:36:03 UTC 2014


On Mon, Jun 16, 2014 at 5:33 PM, Aaron McGlinchy
<McGlinchyA at landcareresearch.co.nz> wrote:
> RE: " I would push back on this requirement. "Public" but not unless you log in doesn't really make sense for open data. But maybe I'm missing something.
> What's this for?"
>
> A couple of reasons for thinking about this, not to try and make the data less open.
>
> - some of the data we have could if it was used inappropriately e.g. applied at a spatial scale smaller or larger than it was collected for/at, or if the level of precision in the data is over interpreted... then it could have significant (financial) consequences for a user.  This sort of thing makes our managers nervous so of course we would have disclaimers/terms of use stating such things as we are not responsible for any loss they may suffer through misuse/misinterpretation of data...  By enforcing a registration we could explicitly require users to acknowledge they have seen and agree to those terms, vs casual browsers missing the Terms statement and just going straight to a dataset.

Yes, if your users read the click-through agreement.

I think your statement is true about the datasets offered on many
other open data portals as well. A different solution is to include
all the necessary disclaimers and granularity information with the
datasets they apply to.

> - Also, we have to report on the 'impact' of our research to our funding masters.  Having a little more information about who is visiting/using our data would be helpful.  Of course google analytics may tell us a lot of the story(?) but possibly having a name and email address could be more informative.

Would someone be allowed to harvest and re-publish or share the data
further? If so, the click through agreement disappears the after the
first user; if not, is your open data really "open"?

I know I've led the discussion away from the original technical
question toward a political one, and the ckan-dev list isn't the right
place for that. Nigel's original answer was correct, it would be quite
a bit of work to make CKAN's authorization work the way you have
requested.

Ian



More information about the ckan-dev mailing list