On 07.03.2014 23:38, Tom Hughes wrote:
On 07/03/14 22:12, Stephan Knauss wrote:
I noticed that in the full history dump the user names associated with a
specific version does not match with the username which was active at
that time.
Well as we have no record of what the username was at the time of the
edit I'm not sure what else you would expect us to do?
Looks like Ilya missed similar information. He recreated it from the
changesets:
https://lists.openstreetmap.org/pipermail/talk/2013-January/065897.html
What makes me wonder: The API .../history call also returns only the
current user name.
This is in contrast to the problem Ilya described.
I'm also certain that before the license change I had code to deal with
changed usernames when parsing planet files (regular, not full history).
So it sounds a bit someone thought about it and it's an intentional
decision to treat the display_name volatile.
Was the API internals changed to use the uid more often instead of the
display_name?
Is the problem described by Ilya maybe solved already? Will all data
coming from the API and in current dumps have only the latest
display_name set, so uid to name is a 1:1 relation?
Then I agree, having a history of display_names has little practical use.
Stephan
_______________________________________________
dev mailing list
[email protected]
https://lists.openstreetmap.org/listinfo/dev
