Re: 2086upd comments

[Alexey Melnikov <[e-mail]>]

Answering to some of the comments.

Cyrus Daboo wrote:

> Section 1: says '/' is used as hierarchy separator, but many examples 
> also use '.'.

Fixed, even thought I can pretend that "INBOX.Drafts" is just a 
non-hierarchical name ;-).
...

> -- 
> What about existing extensions: NAMESPACE, UNSELECT, RLIST, IDLE? I 
> think 2086upd should say something about those since they do exit now.

I don't want to have any unnecessary dependencies. There is a separate 
draft (draft-melnikov-acl-rights-XX.txt) that lists rights for URLAUTH 
commands, I can describe NAMESPACE, UNSELECT and IDLE there.

RLIST - I think it is deprecated by LISTEXT, so I wouldn't bother.

> In formal syntax:
>
>   rights          = astring
>                      ;; only lowercase ASCII letters and digits
>                      ;; are allowed.
>     
>     why not use explicit character ranges for defining this?

I don't really care. Any opinions?

> -- 
> Is 2086 reference really only informative?

2086upd obsoletes RFC 2086. Can it have a normative reference to the 
document being obsoleted?

> -- 
> General issue with extensions: shouldn't the capability identifier be 
> clearly defined in the formal syntax? Most extensions just state what 
> the capability is in the text - I now think that is wrong and that 
> they should extend the 3501 capability syntax item. e.g.:
>
> for 2086upd:
>
>     capability      /= rights-capa
>
> for ANNOTATE:
>
>     capability        /= "ANNOTATE"

("/=" ==> "=/")

Sure, this is an easy addition.
  
Alexey