[Karl Dubost:QOTD] On Features, Bugs, Systems, and Social Adaptation

by M. David Peterson

Update: Question,

Think about the way you design. Do you do the controls first, or do you work out the data, the relationships, the rules, then build controls to measure and set those?


Len Bullard provides his answer below.

Update: Two nice additions to provide some food for your lunchtime (or dinner if you're in the UK, *VERY* early bfast if in Tokyo ;-) mental fodder,

First from Len Bullard,

Controls are emergent. People don't get that. As a result, many systems are preconceived notions based on superstitious relationships resulting in wasted motions instead of frugal applications of measured and just-in-time force to encapsulated data.


Second from Piers Hollott,

It is interesting, people seem eager to adopt the principle of "survival of the fittest" (bestest?) while ignoring that in any evolutionary system, "fittest" and "survival" are both very much rooted in context and subject to change.


[Original Post]
Disagreement, ambiguity, variability, lie, etc are not bugs, but a feature of the system. Each system which becomes too rigid looses its flexibility and often die. The social agreement makes it happen in a *context* keeping the possibility of an error, a mistake and by it, giving the possibility to fork, evolve, etc.

It's a question of balance.

1+1 = 2 most of the time, but not necessary in poetry.


Karl Dubost from a recent post to the W3C SemWeb Mailing List

8 Comments

len
2007-06-14 05:58:55
Or law.


"But that's good; you want to pick a single standard for a given purpose, to the exclusion of others, as long as all suppliers can implement the standard without unrelated restrictions." Groklaw


How is it Spock put it to Kirk: "Indications 2 dimensional thinking."


Controls are emergent. People don't get that. As a result, many systems are preconceived notions based on superstitious relationships resulting in wasted motions instead of frugal applications of measured and just-in-time force to encapsulated data.

piers
2007-06-14 08:51:19
It is interesting, people seem eager to adopt the principle of "survival of the fittest" (bestest?) while ignoring that in any evolutionary system, "fittest" and "survival" are both very much rooted in context and subject to change.


@len: "Spock put it to Kirk" - and now M:D will be flooded with homoerotic fanfic from googling slashers!

piers
2007-06-14 08:52:49
ooops. caffeine-fingers.
M. David Peterson
2007-06-14 10:34:45
@len,


Controls are emergent. People don't get that. As a result, many systems are preconceived notions based on superstitious relationships resulting in wasted motions instead of frugal applications of measured and just-in-time force to encapsulated data.


Nicely stated! +, as usual, it's forcing me to think: *ALWAYS* a good thing, and *ALWAYS* appreciated. Thanks! :D

M. David Peterson
2007-06-14 10:38:02
@piers,


It is interesting, people seem eager to adopt the principle of "survival of the fittest" (bestest?) while ignoring that in any evolutionary system, "fittest" and "survival" are both very much rooted in context and subject to change.


Well put! I think we should add a disclaimer to all text books on the subject matter that states,


The definitions of "survival" and "fittest" are subject to adaptation without prior notice. Please adjust accordingly. ;-)

M. David Peterson
2007-06-14 10:41:39
@piers,


ooops. caffeine-fingers.


Yeah, we need to add a "disable button after first click" feature to the comment "Post" button. For whatever reason it seems that as of late the time it takes for a comment to get processed is quite a bit longer, making it easy to think that the submission didn't take and hitting the submit button again.

len
2007-06-15 09:29:11
Think about the way you design. Do you do the controls first, or do you work out the data, the relationships, the rules, then build controls to measure and set those?


Think of the venerable feedback diagram (Input to Process To Output To Input). First you measure, then you change. Pick the control according to need but determine the need first. Then adapt it. But the measure twice cut once method is good for more than carpentry.


For example, until you know the target framerate for real-time rendering, adding designs for objects is a waste of design time. So pick the measures first, then create the measuring systems and routing that to the review systems, and so on. Not doing that is how bad standards get made.


Controls are emergent.

M. David Peterson
2007-06-16 04:43:48
@len,


So pick the measures first, then create the measuring systems and routing that to the review systems, and so on. Not doing that is how bad standards get made.


AMEN, Brother! :D