On Mar 10, 2011, at 12:03 PM, Marc Petit-Huguenin wrote:
>
> A.48. Section 13.7
>
> The "RELOAD Data Model" registry contains a Code value which is on fact never
> used anywhere. I think it would be simpler to remove this (like it is done
> for
> the Access Control Policies) and redefine DataModel in section 6.2 like this:
>
> enum { single_value, array, dictionary } DataModel;
just removed it
>
> BTW, the select in 6.4.2.1 uses "model" instead of "DataModel."
fixed
>
>
> A.49. Section 10.1.1
>
> The Data Model, Access Control Policy and Kind names permitted in the schema
> does not and cannot match the definitions in the various IANA registries,
> because of the case. E.g. the Data Model registry contains "USER-MATCH" but
> the
> schema only accept "user-match". I suggest to replace the definition of
> "user-match" and other strings by the following:
>
> access-control-type |= xsd:token { pattern =
> "[uU][sS][eE][rR]-[mM][aA][tT][cC][hH]" }
I sort of hate the cause incentive stuff - it just seem to introduce bugs and
some people do it and some mess it up and just use what is in the examples.
Given this is all for machine not human consumptions, I moved it all to upper
cases. If this breaks anyones implementation, lets talk sort out what to do. I
only choose upper case over lower case as it seemed to match what was in the
IANA registries. That contradicts the relaxNG but I don't think many people had
used the relaxNG.
_______________________________________________
P2PSIP mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/p2psip
