Re: Rule of Least Power

I wrote:
>> Len, I think your analysis completely ignores the effect of time
>> and independent evolvability of components in a multi-organizational
>> system like the Web.

On Mar 8, 2006, at 7:49 AM, Bullard, Claude L ((Len)) wrote:
> I understand that aspect perfectly, Roy.

Sure, but you left it out of the analysis.  We are talking about a
design principle (I don't know why the TAG changed it to a "Rule"
since we don't have rules) that declarative communication improves
visibility of interactions, which in turn enables greater evolvability
as the participants in the communication change over time.  Thus,
an analysis that talks about the run-time operation of a short-lived
or closed system (like Airbus) is not going to encounter any benefits
from the design principle.  In order to get the benefits you have to
look at the system as it evolves over the scope of years, not hours,
and independent of any "flag-day" decisions wherein one organization
decides it is time to upgrade.

> I think you ignore that subjective systems
> can be globally stable but create local instabilities.

I am afraid I don't know what you mean by subjective systems.  The Web
is a social system and one of its design goals is to remain globally
stable while encouraging local innovations.

> As a context, the web is global.  As a rule, RLP is
> second-order friendly (a subjective or observer-based
> cybernetic system).  So far so good, but applying that
> rule locally can create instabilities.

It is a design principle, not the answer to every question.  I think it
would be helpful if folks were directed to the Principle of Least  
Authority
that I described back in

   http://lists.w3.org/Archives/Public/www-tag/2005Dec/0032.html
   http://szabo.best.vwh.net/interpretingpower.html

because that also is a design principle that places long-term safety
(knowing that change is inevitable) over short-term efficiency.

> The trick is to tune it, Roy.  See PID controls.
>
> The business intelligence guys are way out front
> on this one.

Umm, no, I think that completely misses the point.

....Roy

Received on Wednesday, 8 March 2006 17:25:43 UTC