Hi Reinoud,
Thanks!
The 0.7 spec makes the following statement about user IDs:
"One of the pieces of data that is always returned with a Person
object is the user's ID. The user ID must be alphanumeric (A-Za-z0-9)
and must uniquely identify the user in a container. This
standardization is intended to allow for prefixing IDs with a domain
name and separator to create globally unique IDs (e.g. "
orkut.com:
34KJDCSKJN2HHF0DW20394"). Note that there will likely be a size limit
placed on user IDs to help manage storing IDs in a database."
I think the language is phrased such that any given ID for a user
will need to always point to that user. For example, if user A gets
12345 as an ID for gadget X and 54321 as an ID for gadget Y, both IDs
(12345 and 54321) will always need to point to user A (there should
not be a case where user B gets either ID in gadget Z).
This still allows the possibility of returning different IDs for the
same user, but seems like it doesn't follow the intent of the
specification. I'm following up with the team about this and will
post any new information to this thread.
Cheers,
~Arne
On Jan 28, 11:32 am, "Reinoud Elhorst" <
goo...@claude.nl> wrote:
> First of all, congrats on getting out the 0.7 spec!
>
> I have a question on userIDs. The specification states that "The user ID
> [...] must uniquely identify the user in a container."
>
> From a containers point of view, does this mean that the userID has to be
> portable between
> applications? In other words, if gadget A and gadget B (A!=B) are on
> the same user profile, they MUST receive the same id for the owner?
>
> In our (data) API, we are used to encrypt the userIDs with
> some application-spec key, and would like to continue to do so for
> OpenSocial applications. We believe this increases privacy for our
> users.
>