RE: early draft on 'best practices for registries' & Dec 8 call ACTION-531

For a finding to have any use, it should guide future actions in ways that wouldn't otherwise happen. Otherwise -- if recommendations are so weak that anyone could say "well, we're just one of the allowed exceptions" -- why bother writing it?

I think the process of gathering consensus on the document  includes determining which of the (currently strong) recommendations have use cases where they would not apply.  It would be great if you had any examples (real or imagined) where the language in the draft is too strong?

As it is, the document discusses several alternative ways of identifying extensibility points (using URIs, using namespaces, using vendor prefixes, using registered values), and focuses on the best practices for registries.


-----Original Message-----
From: Thomas Roessler [] 
Sent: Wednesday, December 07, 2011 1:59 AM
To: Larry Masinter
Cc: Thomas Roessler; Noah Mendelsohn;
Subject: Re: early draft on 'best practices for registries' & Dec 8 call ACTION-531


skimming this document, I notice that it's written in relatively strong language - all the way to an RFC 2119 MUST that's supposed to apply to all Working Groups.

For a "Best Practices" document, I'd expect a much looser style of language: "It is useful to do this ...", "Groups typically do that".  Or perhaps the good old imperative: "Best practice: Use IANA registries."

I'd suggest to either steer clear of strong normative language, or clarify that the status of the document is actually meant to be normative.

Thomas Roessler, W3C  <>  (@roessler)

On 2011-12-07, at 06:27 +0100, Larry Masinter wrote:

> I wouldn't mind using call time for even an informal discussion of the attached...
> For my (way overdue) action-531.
> <draft-registries.txt>

Received on Wednesday, 7 December 2011 17:27:03 UTC