a comparison of hyperstructures: zzstructures, mspaces, and polyarchies by: mcguffin & schraefel...

A Comparison of Hyperstructures: Zzstructures, mSpaces, and Polyarchies

By: McGuffin & Schraefel

Presented by: Travis Gadberry

Abstract

Background

Smaller chunks of info (not full pages)

Focus on the structures, not implementation

Desire for new ways of accessing data

Structures

Multitrees

Polyarchies

Zzstructures

mSpaces

Multitrees

Kind of DAG

Can contain multiple overlapping trees

Overlaps must share subtrees

Ex. Human genealogies

Multitrees

Polyarchies

Can contain multiple overlapping trees

Overlaps may contain subtrees

Overlaps may happen at arbitrary nodes

Coloring edges distinguishes different trees

Polyarchies

Zzstructures

Kind of directed multigraph Subject to a single restriction R:

Each node in a zzstructure may have at most one incoming edge of each color, and at most one outgoing edge of each color

Edges of each color form paths/cycles that do not intersect within the same color

Zzstructures

Zzstructures

mSpaces

Difficult to visualize

Ability to organize data points in multivariate space

Allows for dimensional sorting, changing structure of the tree.

mSpaces

mSpace polyarchy for a 3D 2x2x2 multivariate space. 3! (6) overlapping bi. treesEach row displays 1 slice less (3D, 2D, 1D, 0D)

mSpaces

Taxonomy

Analysis

Zzstructures ?= edge-colored directed multigraphs ?

(ecdm)

Zzstructures == ecdm + R

R can be simulated by node cloning

Advantages?

Comparisons

Zzstructure’s Space non-Euclidean

Not easy to flatten and visualize May be changed independently of content

More freedom Can be much more confusing

Dimensions are like containers Nodes have a relative position in some dimensions

Comparisons

mSpace’s Space Euclidean

Like a high-dimensional grid Slices can be taken and visualized

Space determined by attributes on content More structured Changing location of nodes doesn’t affect space

Dimensions are variables Nodes have a value in every dimension

Comparisons

Nature of overlap between trees Polyarchy – arbitrary Zzstructure – arbitrary Multitree – one subtree is shared mSpace – all subtrees at certain depth are shared

Conclusions & Future Work

New ways to create hypermedia systems

This paper has shown the differences

Visualization applications are needed

Other hybrid or extended structures