RoR: Forum

Ed:

One suggestion for you to ponder.

[Context1]
{
[Event A] --> [intellect * desires] --> [choice B] --> [action B]
}

[Context2]
{
[Event A] --> [intellect * desires] --> [choice C] --> [action C]
}

[Context3]
{...
}

...

Those chains can occur inside of different contexts, which can effect the actions.

(By adding Context, you've basically rediscovered BDD: Behaviour Driven Development, a kind of modeling in which every characteristic of a system is descibed as:

In [Context] When [Event] Then [Action.]

The efficacy of BDD, in theory, is that requirements can be expressed in business language, so as not to wig out management. But it maps directly to a related paradigm, TDD (Test Driven Development) in which applications are developed by writing things called UnitTests first, and then developing until all red/green testing is green.

So, BDD is for modeling, when sharing the design space with non-technical shareholders who don't care about TDD or nuts and bolts; but BDD maps almost directly to TDD, so is, in theory, a coherent scheme for extracting requirements from an organic thing like most committee efforts behind a development project. But... still wishful thinking, still needs buy-in.

But, forget all that; in your modeling of when/why the same actor makes different choices in response to the same events, a rational way of modeling that is via different context.

A context can also have a history...past events that have occurred in that context.

Events occur in Contexts...and you know what I am going to say next; plural.

regards,
Fred

(Edited by Fred Bartlett on 6/06, 12:13pm)

Post 21

Thursday, June 6, 2013 - 12:20pm

In [Church] when [desire to fart arrives] then [squirm uncomfortably but suffer in silence]

In [Bed:Marriage year0] when [desire to fart arrives] then [suppress until after sex]

In [Bed:Marriage year30] when [desire to fart arrives] then [go full boat DutchOven]

See? Same event, different choices.

Same context, different history, different choices.

Don't ask me what 'DutchOven' is; if you don't know, you want to keep it that way.

regards,
Fred

Post 22

Thursday, June 6, 2013 - 12:31pm

How will a given actor react to an event in a given context?

It might depend on what has happened to that actor in the past in that same context.

When we are young, our context history is shallower; we are more open, trusting. Fresh. Naive.

When we are older, our context history is deeper. We are more closed/guarded. Stale. Cynical. Jaded. Burned fingers.

Not uniformly, and not in every context. But that variability can be ...modeled.

Same actors, same events, same contexts,... different histories, possibly different choices/actions.

regards,
Fred

Post 23

Friday, June 7, 2013 - 8:56pm

Fred,

Ed:

One suggestion for you to ponder.

[Mock code]

Good stuff. I was thinking in code, too, but it looked like this:

using System;
using Heavy.Brain.Power;
import Bad.Ass.Attitude;
import Modicum.Of.Restraint;

namespace VolitionRules
{
class IsInSession
{
...

:-)

Good point about the nested events (events inside of differing contexts) leading to different outcomes even from the same general equation -- with a subtle difference in original inputs, and possibly even the dynamic of "emergence" due to an interplay of simultaneous sub-contexts.

Ed

(Edited by Ed Thompson on 6/07, 8:56pm)

Post 24

Friday, June 7, 2013 - 9:14pm

Fred,

Not uniformly, and not in every context. But that variability can be ...modeled.

Same actors, same events, same contexts,... different histories, possibly different choices/actions.

I learned this recently the hard way. For instance, I have recently acquired software which can be utilized to model up to 4900 actor-agents for 50+ generations in a spatial, evolutionary, agent-based economic game. Much to my chagrin, when I tweak the pay-off variables (in order to represent abstract things like "capitalism", "communism", etc.) I don't get just one outcome, but a disturbing range of outcomes.

In spatial games, your neighbors matter. Your neighbors (8 neighbors in a 9-square grid) are the ones with whom you interact, and they can sometimes make or break you. There is only something like a 99.9% chance that choosing a perfect strategy will lead to a perfect outcome -- and this 99.9% chance is taken 4900 times independently. A tenth of a percent isn't really that far off the mark of perfection, but with 4900 independent chances to err, you can still end up with some modest outcome variation.

I guess I'm just initially frustrated that the empirical validation of abstract philosophical truths can be messy, that's all. Things are so much more beautiful when viewed from my armchair. I will continue with the empirics, however -- though I do not plan to ever let go of the soul-nourishing, inspirational beauty of correct ideas during this process.

Ed

(Edited by Ed Thompson on 6/08, 1:17pm)

Post 25

Saturday, June 8, 2013 - 10:14am