[ckan-dev] Distributed changeset model for CKAN

Sat May 14 16:33:39 UTC 2011

Good point, also worth noting that the already implemented CKAN to CKAN harvesting provides another approach for a distributed network of CKANs. 

So as not to over engineer, or risk delaying revisioning implementation lets try to keep revisioning and distribution as separte discussions for now?

Cheers,

James

----- Reply message -----
From: "Rufus Pollock" <rufus.pollock at okfn.org>
Date: Sat, May 14, 2011 17:06
Subject: Distributed changeset model for CKAN
To: <james.gardner at okfn.org>, "David Raznick" <kindly at gmail.com>
Cc: "CKAN Development Discussions" <ckan-dev at lists.okfn.org>

NB: this is not a suggestion that we go down this route at the moment
but a useful piece of background for current discussion about pending
changes.

Over a year ago we did extensive work on a 'distributed'
revision/changeset model and its relation to CKAN and even had working
implementation (though it never went production)!

This is a very general model that would allow the (meta)data in CKANs
to be pushed and pulled like source code in git or mercurial. while is
probably too general for what we want I did find the insights like the
distinction between your changeset and your "Working Copy" very useful
(point is that in a distributed model because of changeset dependency
you have to pull all changesets in a given branch but you may choose
*not* to apply all of them to your working copy i.e. the set of domain
objects / tables that are what is shown to users).

Here is overview (not much on implementation -- see wiki links from
that doc for more):

<http://packages.python.org/ckan/distributed.html>

Implementation with good docstring:

<https://bitbucket.org/okfn/ckan/src/1f72a83087ab/ckan/model/changeset.py>

Rufus
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.okfn.org/pipermail/ckan-dev/attachments/20110514/4f81a1a5/attachment-0001.html>