[OSM-talk] License plan
osm.list at randomjunk.co.uk
Tue Mar 3 16:30:54 GMT 2009
2009/3/3 Pieren <pieren3 at gmail.com>:
> On Tue, Mar 3, 2009 at 2:13 PM, OJ W <ojwlists at googlemail.com> wrote:
>> On Tue, Mar 3, 2009 at 9:11 AM, Richard Fairhurst <richard at systemed.net> wrote:
>>> 2. People who don't like ODbL and withdraw their data. _Assuming_ we can get
>>> the bugs sorted in ODbL, and we can't take that for granted yet, this
>>> percentage should be very small.
> It's very confusing now about who, how and what is deleted with the
> license change. I would appreciate if someone could answer the
> following questions:
As has been said, a lot of this is up for discussion of various
kinds... here's my brief attempt at answering... all responses are
just my interpretation... feel free to say I'm wrong :-)
> - do you delete only data from contributors who explicitly say 'no' to
> the new licence or also if you have no response ? what is the argument
> to consider an absence of response to be a 'yes' or 'no' ?
No response == no... but they might change their mind later and ask
for their data to be reintegrated which really is /fun/. See next q
> - do you delete data from big contributors only or also all small or
> single contributions ?
YMMV on this one. For cleanest DB you delete everything, for most data
kept we run the risk with small "uncopyrightable" contributions. Also
we may treat no response differently to no for this.
>From now on I'm assuming a cleanest DB scenario...
> - if you decide to delete contributions and those contributions are
> only part of the history of objects, do you rollback to a previous
> version of these objects ?
> remove completely the objects if the
> contributor is the creator or the last modifier ?
yes for creator, revert for modifier
> only if the
> contributor is the single contributor on the whole history of the
> object ?
> - if the objects you delete are part of a relation, do you keep the
> relation at the end even if all members have to be deleted ?
or do you revert the relation to the point before the object was added
to the relation, or even to the point before the object was edited (as
otherwise your remaining relation maybe "derived" from the object).
Personally I think you're probably OK removing the object.
Does an empty, unreferenced relation serve any purpose? And if it
doesn't do we care?
> or you
> also delete the relation in this case ? what happen if another
> contributor (who accepted the new license) added/changed properties of
> a relation where all members have to be deleted ?
relations are like any other object -- revert to the relation state
before the person edited, then start removing things from it.
> - if someone says 'no' to the new license and wrote a bot, do you also
> delete the bot contributions ?
we can't tell the difference, so yes. But we may be able to mark most
of the edits as trivial and not remove them.
> - after deletion, do you keep the trace in the history of other
> related objects ? will it be possible for someone else to revert the
> deletion through Potlatch for instance ?
Nasty question :-)
Really the history should be deleted. You can leave a trace that
something happened, but details shouldn't be available, neither should
revert. We don't currently have a way to do that.
More information about the talk