[OSM-dev] What should the changeset api do?

Andy Allan gravitystorm at gmail.com
Mon Oct 13 11:07:02 BST 2008


On Sun, Oct 12, 2008 at 12:37 AM, Matt Amos <zerebubuth at gmail.com> wrote:
> On Sun, Oct 12, 2008 at 12:32 AM, Frederik Ramm <frederik at remote.org> wrote:
>> Matt Amos wrote:
>>> yeah, maybe its worth having a previous_version column in
>>> current_(nodes|ways|relations) for efficient lookup?
>>
>> Mh. *That* would always be version-1 (version is auto-incremented on every
>> update).
>
> would it? would it be worth not assuming this? we could allow other
> version ID allocations. for example, i think it would be useful to
> allocate version IDs globally, like subversion does, so that each
> version ID uniquely identifies all elements which were atomically
> committed together.

Much as I dislike quoting from wikis...

"Are there cases where new version number != old version number + 1?"
--Frederik Ramm 14:06, 5 May 2008 (UTC)

"Possibly no. But in the future with transactional updates we decided
we couldn't necessarily rule out the possibility that numbers might be
skipped, depending on where exactly the numbers were generated. So we
decided to play it safe: less assumptions is good." -- Kleptog

Cheers,
Andy




More information about the dev mailing list