Programmator creates softcoded routines.
Programmator itself is a hardcoded
routine.
Input:
{specified Concept ID}
Output:
program for specified Concept.
This is a sample
implementation of Programmator, which creates softcoded routine for specified Concept (SpecifiedConceptId):
1) Get the
most important cause-effect relations which
have:
EffectConceptId = SpecifiedConceptId
Sort
results by coherence attribute values. Select
results with the highest coherence attribute.
Let’s name
the result as: “RelatedRelations”.
2) Get
cause concepts for the specified concept.
Hint: ConceptId (of the cause concept) = CauseConceptId
(of “RelatedRelations”)
3) Select
only the most important concepts.
Hint: sort
by concept strength or by Desirability
attribute.
Let’s name
the result as: “TheMostImportantCauseConcepts”.
5) Create {softcoded routine}
as a sequence of TheMostImportantCauseConcepts.
4) Put {softcoded routine}
to softcoded routine entity.
PCnous wants to achieve
“collect information about an ICQ pal”.
Programmator finds out
that are the most important cause concepts for
“collect information about an ICQ pal”.
They are:
So Programmator creates softcoded routine:
1) Execute softcoded routine of “Collect information by Google” Concept.
2) Execute softcoded routine of “Look at ICQ user details” Concept.
3) Execute softcoded routine of “Ask a person about himself” Concept.
Programmator can work successfully only with “trained” concepts
and cause-effect relations. Learning process can be accomplished by:
Softcoded
routines should be evaluated by experiment. Such
evaluation helps to:
Advanced
programming process might include:
More flexible process of programming (because of using softcoded routines
which can be reprogrammed).
Select information from the memory