Re: registries and designated experts

[Dave Crocker <[e-mail]>]

oh boy...

On 6/15/2012 3:57 AM, Graham Klyne wrote:
> On 12/06/2012 15:56, Dave Crocker wrote:
>> It's almost inevitable that many designated experts will, in fact,
>> act as gatekeepers.
...
> The effect may sometimes be similar to being a gatekeeper but,
> speaking for myself, that's not how I see my role.

Herein lies the core problem with these types of discussions:  Most 
participants use themselves as the metric.

Designing human processes requires consideration of other folk, not 
ourselves.  Other folk display widely varying behaviors.  They have less 
information and education on the topic than current participants.  They 
have moods.  They have distractions. They have biases.

In other words, real-world, human processes are "noisier" than we tend 
to design for.

When looking for factors to consider, a posting with "here's how I 
behave" can be extremely helpful.  When looking for what can go wrong, 
it actually tends to undermine the exercise, IMO.

Note that this is not about diligence.  It is about typical human vagaries.


> My experience is that no amount of review completely bullet-proofs a
>  spec against misinterpretation. So we do the best we can.

The second sentence is such a natural choice, given the first.  What it 
misses is that "the best" often means excessive hassle.  That is, very 
poor cost/benefit tradeoff analysis.

Quality control processes can and do help.  Carried to excess, they make 
the activity more effort than it is worth.

In the face of being unable to ensure perfect bullet-proofing, we often 
think that we need to put in more effort.  In fact what is often 
appropriate is /less/ formal bullet-proofing and trusting market forces 
to do the rest, since they will anyhow.

d/


-- 
  Dave Crocker
  Brandenburg InternetWorking
  bbiw.net