[ckan-dev] Who is in charge of the Colorado instance?

[Seb Bacon]

On 9 May 2011 10:18, William Waites <ww at styx.org> wrote:
> * [2011-05-09 10:12:21 +0100] Seb Bacon <seb.bacon at okfn.org> écrit:
>
> ] > Do we really need to embed API calls and write CREPs here?
> ]
> ] Too late ;)  CC'ing this to ckan-dev.
> ]
> ] But yes, I think so.  Since I've started at OKF this conversation has
> ] come up several times, either in "who is the maintainer of site X"
> ] (sysadmin-type questions) or "how many instances do we have of CKAN"
> ] (marketing-type questions).
>
> I agree that this is important, I just think your propsed solution is
> much more complicated than it needs to be.
>
> I also don't like systems that call home and would be very upset if a
> piece of software tried to contact some entity on the Internet without
> my explicit configuration or approval. It's creepy and orwellian and
> a very bad thing for free software to do.

I know what you mean.

OK, instead of the check-in stuff, how about this: we add an opaque
string to the headers that makes it easy for us to get a reasonably
complete list of CKAN instances with a google search?

> ] On the point of writing CEPs -- well, maybe it doesn't need a full
> ] one, but it needs discussion IMO: it's something *I* think should be
> ] built into CKAN, and the current (pre-CEP) way of proceeding would be
> ] for me to mention it to James or Rufus and if they don't have a
> ] problem, just going ahead and doing it.
>
> I agree the "private conversation" route is bad and keeps people in
> the dark. But I think that the whole formal CEP thing is overkill,
> simply writing to the list as you have done is sufficient to keep
> people in the loop and get their feedback.

Fair enough, as per other threads I don't think everything would need
formally writing up.  In this case, it seems unlikely.

Seb