Re: draft-housley-two-maturity-levels

[Keith Moore <[e-mail]>]

On Jul 28, 2011, at 10:06 AM, Peter Saint-Andre wrote:

> On 7/28/11 10:03 AM, Eric Burger wrote:
>> And the real question is, are we moving forward? I think that we are
>> not moving as far as we originally wanted. However, I offer we are
>> moving a baby step forward, and as such is worthwhile doing.
> 
> We are more closely aligning our documentation with our organizational
> running code. All other things being equal, that's a good thing.

Hmm.  I've long believed that :

- trying to document existing practice
- trying to document desirable practice

are both worthwhile endeavors, as long as you don't try to do both at the same time.  When you try to do both at the same time, there is a conflict.

If someone wants to write a document that says we generally follow RFC 2026, except that:

- drafts hardly ever advance to Draft Standard and even more rarely to Full Standard, unless there is significant use of the protocol and there are bugs that need to be fixed (in which case the ability to advance can sometimes serve as an incentive of sorts)
- we have never been serious about periodic review of standards and we don't have enough time/energy to do that
- we've never really nailed down what Historic meant, and when it was appropriate to use it

etc.

that would be a fine thing.

And real changes to the process, say to bring in formal cross review earlier, to clarify the nature of community consensus and the need for it, etc. might also be a fine thing.  Unfortunately, such discussions are always contentious and difficult, because they affect the whole community, but they also attract a lot of interests from individuals with particularly unique axes to grind.  So we keep trying to fix the substantive problems with incremental changes.  I forget who it was who said yesterday that we can't really do that, but I certainly agree with him.

Meanwhile, it's not clear to me that simply changing from one document that we don't strictly follow, to another document that we won't follow much better, is helpful.  And I don't think IETF's problems with standards quality or process can be addressed merely by changes to the number of maturity levels.  That strikes me as a bit like rearranging deck chairs...it might make people feel better but is of little consequence.

In other words, I'm not convinced that this change will do much harm, but I'm also not convinced that it will help much.  And yet we keep flogging this idea...

Keith