Difference between revisions of "Natural 2D visualization of 1D bit data"

From apm
Jump to: navigation, search
m (Apm moved page Natural visualization to Natural 2D visualization of 1D bit data: more descriptive title)
(→‎Related: added 2021 "Can Tensor Programming Be Liberated from the Fortran Data Paradigm?")
Line 27: Line 27:
 
== Related ==
 
== Related ==
  
 +
* 2021 [https://github.com/conal/talk-2021-can-tensor-programming-be-liberated Can Tensor Programming Be Liberated from the Fortran Data Paradigm?] (github shownotes)
 +
----
 
Conal Elliott (draft 2018) – "Data-Parallel Programming without Arrays"
 
Conal Elliott (draft 2018) – "Data-Parallel Programming without Arrays"
 
* http://conal.net/papers/data-parallel-programming-without-arrays/
 
* http://conal.net/papers/data-parallel-programming-without-arrays/

Revision as of 11:17, 11 July 2023

Natural 2D visualization of 1D bit data.

Question: How can a one dimensional stream of data be displayed as 2D visual graphic in the most natural way
when there is no information at all provided about correct interpretation?

Specifically:
How could the progression of evaluation of an untyped lambda expression be
most naturally mapped to a black and white pixels 2D visualization?

The right approach cannot involve any 2D array type interpretation since this introduces
an arbitrary line-end wraparound length. Rather one wants to use an in-an-obvious-natural-way serializable representation of 2D bit data that
does not contain an arbitrary line-end wraparound length.
Tries/MemoTries might do the trick.

In a sequence of evaluations of a lambda calculus expressions much of the structure often stays static and unchanged.
So one might want to visualize the differences instead.

Related

Related

Conal Elliott (draft 2018) – "Data-Parallel Programming without Arrays"

Trie datastructure

Retrieved from "http://apm.bplaced.net/w/index.php?title=Natural_2D_visualization_of_1D_bit_data&oldid=15225"