[OSM-talk] Import guidelines & OSMF/DWG governance

Pieren pieren3 at gmail.com
Wed Sep 19 10:25:42 BST 2012


On Wed, Sep 19, 2012 at 9:37 AM, Frederik Ramm <frederik at remote.org> wrote:

> Are you saying that we should have had a vote on the wiki, or what? Who
> would have been eligible to vote? And are you at the same time saying that
> changing a policy on the wiki is not "clearly published"?

To progress a little bit in the debat and clear up some
misunderstandings.. It seems that nobody is able to say why the
separate user account became suddently mandatory in november when it
was earlier a recommendation. Never mind, it was not a problem for us
until someone was blocked just for this particular reason, not because
it was a good or bad, small or big import. No, just because he did not
use a separate account.
Some people say "just obey to the DWG telling you to follow the
guidelines". We say "we don't agree with the last change in the
guidelines because it does not fit our local practices" (some are also
saying that the DWG is working beyond his mandate but that's another
story).
What we explained is that we defined in France our own policy for this
particular data source (also on the wiki). We started it years ago.
The size of the whole French cadastre dataset is huge. We could upload
it in a single mass import with a bot using a seperate user account as
we did for the Corine Land Cover. We could follow 100% of the import
guidelines. Trust me, we have all the capacities in humans,
competences and computers to do it. But instead, we decided that
buildings have to be better integrated with the existing data and
better controlled by simple, average contributors in smaller chunks.
We also decided to exclude major parts of the dataset because it's not
usefull for OSM like the parcels or couldn't be well integrated
automatically like the roads, street names and address house numbers.
We decided limit the import to the railsways, buildings and waterways,
we decided to do it at the size of the dataset is itself published,
means at the municipality size; it can be a 2 millions inhabitants
city or a 50 houses village. We also learned from the previous
imports. The French community size is also big enough to manage this
kind of "crowdsourced" import itself. We are not so big as the Germans
(but hey, who is ?), we are 2 or 3 times smaller but just the second
or third biggest community in OSM. We have servers, we have a local
chapter, we have quality assurance tools, we have developers, we have
many eyes watching the map and reporting issues to the group, we have
one of the most active local mailing list. And we have our own policy
to import this dataset where finally the only main difference with the
standard guidelines is the separate user account. We are not against
it, we can even promote it. We are against making it mandatory.
Because we think that all the good reasons provided for this
requirement do not apply here. Even some DWG members admit that the
separate user account will not be checked for small imports. They are
just worry when they detect some stange behaviours or very active
users. Themselves, they cannot say when exactly the special account
becomes a reason to block someone after how many uploads, changsets or
edits. It's only when the contributor enters in their radar tools and
after some arbitrary decision.
What we ask is not much. We ask that the DWG is taking into account
the local communities and their local import policies when it is done
with all the good will and sincerity. But we also have our "black
sheeps". We also need the DWG for blocking users when it is necessary.
But let the local community decides when and who. And for that, we
need to contact people in their speaking language, not in English,
either through a local representative or e.g. standard messages
previously translated. Then check with the local community if or what
goes wrong with the person and only at the end, suspend his account.

Pieren



More information about the talk mailing list