axiom-developer
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Axiom-developer] documentation and the crystal


From: root
Subject: [Axiom-developer] documentation and the crystal
Date: Sun, 28 Dec 2003 15:23:58 -0500

Bill,

more top level speculation about the crystal leads to an interesting
problem. right now we have the image of books and standard page
navigation allows us to put "next", "previous", "index", etc on
each page as navigation buttons.

at the other extreme is the web where "forward" and "back" are all
of the possible navigation choices.

with a crystal one assumes that each facet is somehow "near" to the
adjacent facets. since this is a virtual object the facets don't need
to be all equally shaped (e.g. 6 sided). some facets could have many
possible neighbors (where information views are rich) and some could
have relatively few neighbors. the notion of "next" facet could be
specified by an angle measure of some sort. "forward" and "back"
become paths on the crystal surface. hyperlinking can move from any
facet to any other. in particular, viewing a "corner" could show
several facets at once (e.g. code, it's location in the lattice,
and the documentation for that code) which are all joined at a
virtual corner.

more generally one could embed one crystal within another. so the
inner crystal facet shows the type hierarchy and the outer crystal
facet looks at code for a particular item in the hierarchy. this 
is the 3-D notion of 1-D pipes, that is, information from one source
is filtered thru a pipe and shown thru another source. (e.g. grep
the algebra for "Category" and filter out only the lines that define
the domain).

the crystal is a virtual object which is really nothing more than
a way to organize hundreds of functions to manipulate code. the
crystal could be specified as a file containing its structure
definition where each facet is just a series of pipe-connected 
functions and a list of adjacent facets with navigation tags.
it could be built on top of a standard browser.

if done right the crystal should not be axiom specific, only the
functions need to know the details of their data sources.

t




reply via email to

[Prev in Thread] Current Thread [Next in Thread]