Dots are legal in a client ID (as per the spec), and there’s nothing inherently 
wrong with a client ID that’s a domain name. However, how can you be sure it’s 
google that gets the client ID “google.com <http://google.com/>”? And what if 
Google wants to have two clients?

In many implementations (including ours from MIT ITC), the client ID is random 
(we use a type 4 UUID) and we’ve got a separate field for human-readable names 
(client_name, defined in the dynamic registration specification).

 — Justin

> On Aug 29, 2016, at 12:51 AM, Viruthagiri Thirumavalavan 
> <[email protected]> wrote:
> 
> I'm working on a OAuth2 server project. Can I use domain name for generating 
> client_id ?
> 
> Ex: For instance if Google registering an app in my server, then the 
> client_id will be google.com <http://google.com/>
> Am i allowed to use dots in client_id? Is it wise to use domain name as 
> client_id ? what are the drawbacks?
> 
> Thanks
> -- 
> Regards,
> Giri
> _______________________________________________
> OAuth mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/oauth

_______________________________________________
OAuth mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/oauth

Reply via email to