I have to admit that I have never tested OPs or RPs for processing error 
messages according to the spec.

Another thing to add to the tests:)

I would be surprised if many of them properly implement direct errors.   
For direct errors it would be simple enough to add a error code to the 
parameters returned.
At the moment the only required elements are namespace and a human readable 
string.

For indirect errors the same issues apply.
I would rather define a new parameter for a numeric error code than mess with 
having RPs try and parse it out of a string.

The existing openID 2.0 spec accommodates this.

It is probably properly an extension to openID 2.0 that could be incorporated 
in V.next.

We need a WG to define the parameter for the error message eg openid.error_code 
 and a list of error conditions with say 4 digit numbers assigned to them.

Adoption will take a wile but it is probably worthwhile.

John B.

On 2010-01-06, at 12:57 AM, Johannes Ernst wrote:

> The general list had nothing to say so far. Is that something that should be 
> covered in an addendum / errata ... or other document?
> 
> 
> Begin forwarded message:
> 
>> From: Johannes Ernst <[email protected]>
>> Date: January 4, 2010 10:18:48 PST
>> To: [email protected]
>> Subject: [OpenID] openid error mode
>> 
>> Apparently a few people in the wild do implement openid.error. What do RPs 
>> do when they receive that? Display it to the user? Throw up their hands and 
>> say "something weird happened"?
>> 
>> There are plenty of problems with the way openid.error is (under-) defined, 
>> including:
>> - not internationalizable
>> - not computer-interpretable
>> - unclear whether the conveyed text message is formatted, or how.
>> 
>> Questions:
>> 1. Could we come up with an enumeratable list of error conditions, or are 
>> everybody's possible errors different? Error codes?
>> 2. Should we define that the error message be used for internal logging or 
>> to be shown to the user?
>> 3. If the latter, should we define that the locale of the error message 
>> should be the user's locale at the OP? What if the problem is that the user 
>> is not known at the OP? Should the RP pass on a locale?
>> 
>> Happy new year everybody,
>> 
>> 
>> 
>> Johannes.
>> 
>> _______________________________________________
>> general mailing list
>> [email protected]
>> http://lists.openid.net/mailman/listinfo/openid-general
> 
> _______________________________________________
> specs mailing list
> [email protected]
> http://lists.openid.net/mailman/listinfo/openid-specs

_______________________________________________
specs mailing list
[email protected]
http://lists.openid.net/mailman/listinfo/openid-specs

Reply via email to