Re: [apps-discuss] "X-" revisited

[Mark Nottingham <[e-mail]>]

On 28/06/2011, at 7:29 AM, Dave CROCKER wrote:

> It can't be only for documents within the IETF.
> 
> The entire purpose of the X- construct, when we first introduced it in RFC822, was to support /private/ activities.
> 
> It defined a safe haven for that private work, within a naming structure defined by the equivalent of the IETF.  But 'private' means outside the IETF.
> 
> What motivates the current document is that 'safe haven' turns out not to be nearly as important as making it easier to take the private efforts that have become popular and make then standardized, with the minimum transition pain.


I don't think that it goes far enough, as written. What's needed here is something we can point people at when they ask why they shouldn't use X-. As it is, it says:

> Therefore, this document recommends against the creation of new names with the special "X-" prefix in IETF protocols.


We understand what that means, but to the outside world -- i.e., someone who's NOT an "IETF insider" -- that statement isn't nearly clear enough.

I'd like to see the document rewritten with that target audience in mind -- i.e, instead of talking about applications protocols passively, speak actively about why using X- is a bad design choice for *your* protocol extension. 

Peter, I'll offer to help again, if you're amenable.

Cheers,


--
Mark Nottingham   http://www.mnot.net/