[WRAP] Name space and prefix - OpenID Harmonization
Allen Tom
atom at yahoo-inc.com
Tue Mar 2 18:59:40 UTC 2010
As a datapoint OAuth 1.0a deliberately did not change the version string
from 1.0 to 1.0a, and it was determined that the version parameter did not
add any value. I can¹t quite recall the reasoning behind this off the top of
my head though.
Big +10000 to trying to harmonize OpenID and Oauth WRAP.
Allen
On 2/25/10 8:22 PM, "Nat Sakimura" <sakimura at gmail.com> wrote:
>
>
> On Fri, Feb 26, 2010 at 11:55 AM, Dick Hardt <dick.hardt at gmail.com> wrote:
>>
>> On 2010-02-25, at 4:11 PM, Nat Sakimura wrote:
>>
>>> Hi
>>>
>>> This may have come up earlier but ...
>>>
>>> I think Wrap should have a namespace / versioning syntax.
>>> Invariably, it will evolve, and will require version number etc. so, it
>>> seems better to me to have one from the beginning.
>>>
>>> e.g.,
>>>
>>>> wrap_ns=http://whatever/wrap/1.0
>>>> wrap_client_id ...
>>
>> Versioning was discussed. I don't recall the details, but it was decided it
>> did not add value.
>>
>
> I actually think it does.
> Perhaps not in the initial version, but in the future for sure.
> So, it is better to have it in the design from the beginning.
>
>>
>>>
>>> I would go further. Why is underscore '_' is used for the delimiter?
>>> If we make it dot '.', it will improve the future compatibility with OpenID.
>>
>> Or OpenID could change to using '_' :-)
>
> If you use '_' as the namespace delimiter, then '_' should be disallowed in
> the parameter name, which is not the case right now.
>
>>
>>
>>> So, we could do something like:
>>>
>>>> openid.ns=http://whatever/wrap/1.0
>>>> openid.client_id ...
>>>
>>> The same applies for OpenID. For an unknown reason, though OpenID has
>>> namespace so that we write:
>>>
>>>> openid.ns= http://specs.openid.net/auth/2.0
>>>
>>> the prefix "openid" is fixed. We should be able to change it like:
>>>
>>>> wrap.ns=http://specs.openid.net/auth/2.0
>>>
>>> Now, the third point.
>>>
>>> Could we not try to harmonize the variable names between the two specs?
>>>
>>> OpenID is in use widely, so it is kind of hard to change it,
>>
>> Interesting assumption. At IIW we discussed OpenID v Next that was NOT
>> backward compatible. It would seem that there is an oppportunity to make
>> changes to OpenID as well as OAuth WRAP.
>
> yes. The above also requires changes on the OpenID side, but I am seeing
> an opportunity to make the transition smoother.
>
>>
>>> so I would request Wrap community to come closer.
>>
>> WRAP followed OAuth, which has much broader adoption from what I know than
>> OpenID
>>
>
> Arguably yes, but at the same time, 'wrap_' is not 'oauth_' ;-)
>>>
>>> IMHO, we should try to harmonize/unite instead of fragmenting.
>>
>> Agreed, but perhaps the changes could happen in OpenID or a combination?
>
> Definitely in combination.
>
> It is good that OpenID Foundation finally can start creating WGs again.
>
>>
>> -- Dick
>>
>> _______________________________________________
>> specs mailing list
>> specs at lists.openid.net
>> http://lists.openid.net/mailman/listinfo/openid-specs
>>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openid.net/pipermail/openid-specs/attachments/20100302/3733bcee/attachment.htm>
More information about the specs
mailing list