[OSM-dev] [OSM-talk] Trouble in Rangoon

[Marcus Wolschon]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Frederik Ramm schrieb:
| The only thing I'd ask of editors is to support "create_changeset"
| and then use the ID returned on subsequent edits. This, I believe, is
| a small price to pay for the grouping advantages we would get.
| Automatically creating one changeset for every single change would be
| an option for a changeover time but not really desirable.

That sounds simple enough.
I was more thinking along the lines of josm,
where you download a part of the map, make
all your edits and then upload it.
But other editors may of cause upload every single
change as it is made. So your way may indeed be better
here.

What about implicitely creating a new changeset if none is
given, so that a transition is smooth?

Can anyone suggest to handle the case of user B
editing entities that are edited in the changeset
of A before AND after the edit of B?
So we have:

A: open changeset
A: move point X to the left
B: move point X a bit north
A: move point X to the a bit more left
A: close changeset

Do we need some kind of soft-locking with timeouts
for changesets like with the "conflicting edits"
in MediaWiki?

Marcus
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFHxr1Qf1hPnk3Z0cQRAvUJAJ42Wt6HFi2CotYV6SGlmY7Hfp9jugCeJPfe
8YhGs6W6F0rkKisaYJBRUuw=
=typu
-----END PGP SIGNATURE-----