How to store OpenID Identity in DB - any recommendation?

Dick Hardt dick at
Tue Aug 8 02:04:52 UTC 2006

Would this problem be resolved if the URL the user typed in is only  
used for discovery and a hint to the IdP as to the identifier the  
user would like to use, and the RP uses the identifier that the IdP  
returns? More likely that the IdP will be consistent in what is  

Perhaps this is something that can be addressed in 2.0.

-- Dick

On 7-Aug-06, at 6:26 AM, Lukas Rosenstock wrote:

> As far as I can remember the consent on this list was to *not* assume
> identities to be the same if they *can* be different, therefore it is
> necessary to store the complete canonical identity as given.
> HTTP/HTTPs remains a problem. The slash problem could be solved if the
> IdPs use redirection (30x) because the canonical URL that identifies a
> user is the one *after* the redirect.
> In my databases I use a table with local unique ids and another one  
> that
> matches OpenIDs to local ids, users can manually associate multiple  
> URLs
> with one account.
> Lukas
> Am 07.08.2006, 15:04 Uhr, schrieb Vladimir <vlach at>:
>> Hello,
>> I read some of the previous articles about HTTP/HTTPS identities,  
>> but I'd like to know, if you can recommend me how to store and  
>> identify  multiple identities as a one single user.
>> We all know, that user could enter:
>> [or other variants without http/https]
>> in most cases refers to one single identity.
>> What would you recommend to webmaster? How to store an OpenID  
>> identity in database at this moment based on current specs?
>> Would this be a good/bad idea?
>> (without ending trailing slash and http/ 
>> https?) Is this compatible with future OpenID development?
>> How do you manage this in your system?
>> Thanks for your suggestions,
>> Vlad

More information about the yadis mailing list