Mapping computer state changes as a space.

Click For Summary

Discussion Overview

The discussion explores the concept of mapping computable programs into a spatial representation, focusing on the state changes of programs and the implications of complexity in measuring these states. Participants consider theoretical frameworks, practical examples, and challenges related to defining and quantifying program states.

Discussion Character

  • Exploratory
  • Technical explanation
  • Conceptual clarification
  • Debate/contested

Main Points Raised

  • One participant proposes mapping the space of computable programs by analyzing simpler problems to identify distributions and potential clumping of variables.
  • Another participant questions the measurability of program states, highlighting the challenges posed by complexity metrics like McCabe's complexity and the implications of variable content on understanding program states.
  • Concerns are raised about the vastness of possible program state spaces, particularly when considering precision limits and undefined behaviors in floating-point arithmetic.
  • A participant suggests that for small functions, it may be possible to map their state space effectively by focusing on start and end points while ignoring intermediary paths.
  • An example of a state machine controlling a traffic signal is provided to illustrate a simple application of state mapping in a practical context.

Areas of Agreement / Disagreement

Participants express differing views on the feasibility of measuring program states and the implications of complexity, indicating that multiple competing perspectives exist without a clear consensus.

Contextual Notes

Participants note limitations related to the abstraction process, the vastness of state spaces, and the challenges of precision in floating-point arithmetic, which remain unresolved in the discussion.

lostminty
Messages
80
Reaction score
0
I want to see if we can map the space that is computable programs.

You wouldn't consider complex problems, once you mapped the simplest problems into a space of points you can get a distribution. You may see that these clump. Seeing as several of the variables are proportional to the history of the variables and one is time from then on. You could develop 2D field lines. This could be then turned into.
 
Technology news on Phys.org
How to measure program state is what? I'm missing something here.

If you use some arbitrary tool, like McCabe's complexity, then larger programs -- ones that have a large number of parameters and several functions, for example, the complexity values that pretty much suggest the program is not completely understandable nor completely testable. And I assume not measurable, at least not easily.

I get that you are restricting complexity, but I do not see what you think is 'measurable'. Content of variables? Using this approach gets us in trouble fast. Here is why: If you have a simple addition of two double precision variables then the upper limit of possible program state "space" is the cartesian product of all values from DBL_MIN -> DBL_MAX by increments of DBL_EPSILON for both variables. --- This is an enormous number. Really, really big. This also runs afoul of the problems with limits of precision DBL_DIGITS - e.g. adding 10^50 + (1/10^50) results in 10^50 - and the problems of undefined behavior and overflow - getting a NAN or INF result.

So. I'm confused.
 
Last edited:
I excluded the abstraction process. For very small functions you can map it's space. The process described is effectively path independent. If you know the start and end points and can parameterise it as a function of the localised change you can then ignore the space in between and consider them sequences that have a determined path.
 
One of the simplest examples of a state machine would be a rom only program for a microprocessor that controls a dumb traffic signal, which consists of a series of delay loops for each of the 6 states in this example {all signals red, north south green, north south yellow, all signals red, east west green, east west yellow}. Registers would be used for the loop count for the delay in each state.
 

Similar threads

  • · Replies 18 ·
Replies
18
Views
2K
Replies
29
Views
6K
Replies
7
Views
4K
Replies
9
Views
2K
  • · Replies 7 ·
Replies
7
Views
2K
  • Sticky
  • · Replies 13 ·
Replies
13
Views
8K
  • · Replies 59 ·
2
Replies
59
Views
6K
Replies
4
Views
1K
Replies
4
Views
4K
  • · Replies 0 ·
Replies
0
Views
2K