[OSM-talk] Alternative to the 2nd account requirement for imports
sly (sylvain letuffe)
liste at letuffe.org
Sun Oct 21 17:25:20 BST 2012
Hi,
( Concerning the current discussion about french cadastre import, but also
could apply to other types of imports, and following Richard's proposition. )
In order to try reaching a more peacefull atmosphere about the french cadastre
import, I'd like to suggest an alternative to the current 2nd account
requirement for imports.
As I see it, one of the problem we face is not understanding each others :
1) We don't, or I don't, at least, understand the goal achieved by the 2nd
mandatory account.
2) And you don't, or it seams you don't, understand why it is a pain for us,
and why I/we think this is not a good solution.
Recent talks here and there have maybe increased my ability to understand why
1) , but still, I might be missing something so here is what I understand
about it :
"Using a second account for imports is a very convenient way for the DWG, and
hopefully for others to tel appart import contributions from other
contributions. Which leads to easier reverts, monitoring, contacts of those
imports owner, which, statistically, tends to be more problematic than the
others. "
Have I understood all goals ?
I no, could you clarify what I'm missing.
If yes, may I introduce an alternative, wich, in fact, was an alternative
allready proposed by Richard Fairhurst here :
http://lists.openstreetmap.org/pipermail/talk/2012-September/064482.html
In a quick sumary :"No 2nd account required, but specific tags on the
changeset when doing imports required" (which tags might be another topic)
I don't really remember why I, and others of the DWG didn't looked on that
idea a little closer.
Presumably we (french) thought this wouldn't change much as tags to set in
JOSM are still to be added to every import, and easily forgotten, badly
spelt, and imports over 20'000 nodes would still require a 2nd account.
And you (DWG), presumably thought, as F. Ramm pointed it out today :
"Now of course you (and Alex Barth) have a point when you say: This could
all be solved by proper source attribution in the changeset!"
http://lists.openstreetmap.org/pipermail/talk/2012-October/064880.html
At french mappers side, we discussed this idea today, and quickly found very
good ways of solving our pain :
http://josm.openstreetmap.de/ticket/6742
Not only does this idea pleases french contributors (those who expressed) but
it also pleases our tools maintainers for cadastre import watching.
What's left is now your opinion, and issues raised by F. Ramm compared to
advantages such a solution has.
I won't expose those advantages, most of you beeing developpers, I think you
allready get them.
But I can expose disadvantages of the current 2nd account requirement related
to the "what I suppose goal DWG is try to achieve with it" compared to the
changeset approach.
(I won't mention the french angryness lowered as it is not technical)
- People will still make errors like uploading non import changeset with their
import account, and in case of revert make it hard to find out.
- Risks of unability to contact the users if the link to is real account isn't
there and people did lost his 2nd email
--- Allowing a 2nd account with the same email as F. Ramm proposed will permit
new type of unrelated problems spammeurs might take advantages of.
- France buildings are now 50% imported, the 2nd account technic will hardly
solve those 50% cases, while the changeset one will, if we are allowed to add
tags to past changsets (I know that it isn't gone be easy at all, but at
least less destructive for users than fake their user account into a new one)
- It won't polute the number of active users statistics (ok, this one isn't a
show stopers) but you get the idea, an account should better be a bijection
to a real user. (Several funny things might happen when you use something for
not what it was created for : Now that I have 2 active accounts, do I have 2
votes for changing licence ? are my futur account preferences merged ? à
I think there are others but you get the point.
What's left is : are F. Ramm listed issues not acceptable for what we gain,
and are they not doable ?
I'm taking F. Ramm's words :
"we should not invent mechanisms that make life harder for mappers"
(of which I truncated "But I have less patience for mass importers;")
That might be the only point we both agree on, but my opinion is that we
should try that.
Postponing that to the future is ok, but like we all know, future starts in
now+ε
--
sly
qui suis-je : http://sly.letuffe.org
email perso : sylvain chez letuffe un point org
More information about the talk
mailing list