Re: ascii fractals

pinbackstein,

I don't have a quick answer to your questions. But your article gave me an
idea on how to do "true" ascii fractals.

Write a little fractal algorithm that outputs into a little drawing
application, making sure that:

* you only use a set of drawing elements that resemble ascii characters, for
example you could limit yourself to the following subset of only four ascii
characters: / \ _ |
* you stay at the same element size throughout the algorithm
* run the algorithm for several hundred to several thousand steps, i.e. until
it "looks good" (You may need to add something into your algorithm to keep the
resulting pic within certain bounds, say, within a printed page)
* run the resulting picture which is a graphic, through an OCR that you
restrict to the above predetermined set of four ascii characters

and voila, you have a "true" ascii fractal, i.e. as true as a true ascii
fractal can be.

Rationale #1: If you try to output your little fractal algorithm directly into
ascii, then you will run into the same ascii rendering problems that you
described in your article below. Outputting into a little drawing application
avoids this ascii rendering problem.

Rationale #2: The ascii rendering problem is tackled in a separate step by the
OCR. This step will certainly introduce singularities into the ascii pic as
the OCR resolves ascii rendering problems that occur all over the pic. These
singularities will in all likelyhood cut both ways, i.e. they can make or
break the aesthetics. So you will have to diddle with the fractal algorithm.

Enjoy the surprises!

=======================================================
[]     +                                             []
[]   \^^/      Visit the Gallery of the 'steins!     []
[]   ( ">     http://www3.sympatico.ca/petecasso/    []
[]   (> )>   Birdstein, Queen of the backyard tree   []
[]    ^^                                             []
=======================================================

sgt pinback wrote:
> 
>   one type of 'fractal' (<-in quotes to avoid discussion of what a
> fractal really is) is a dragon curve (i think the name comes from a
> type of paper folding). they're made something like this: you start with two
> things (one i think is called an 'initiator' the other i forget the name
> of). like a line segment
>                _
>  _  or a box  |_|
> 
> and a pattern made of bits it can replace
> 
>  __/\__ (<- here i am considering two underlines to be one line segment)
> 
> the pattern goes in replacing the bits in the original:
> 
> . __/\__   or   __/\__
> .              |      |
> .             /        \
> .             \        /
> .              |__  __|
> .                 \/
> 
> and repeat (in non-ascii usually things get replaced by smaller and
> smaller version --here it can't...).
> 
> .                                          __/\__
> .and repeat:                               \    /
> .                                    __/\__/    \__/\__
> .  __ / \ __         or the         |                  |
> .                    box bocomes   /                    \
> .becomes:                          \                    /
> .         __/ \__                   |                  |
> .         \     /                __/                    \__
> .  __/\__ /     \ __/\__         \                        /
> .                                /                        \
> .(breaks added to help show      \                        /
> . what replaces what)            /_                      _\
> .                                  \                    /
> .                                   |                  |
> .                                  /                    \
> .                                  \                    /
> .                                   |__  __      __  __|
> .                                      \/  \    /  \/
> .                                          /_  _\
> .                                            \/
> 
> the aspect ratio distortion from using
> .  |     as the verticle version of __/\__
> .   \
> .   /
> .  |
> i think makes the box look pretty bad (as a fractal anyway). anybody got
> ideas how to fix it? this also messed up things when i tried to use
> things like __|\__
> 
> for the 'snowflakes' previously posted i
> used
> . __
> . \/   and __/\__
> 
> and (sort of)
> .  __
> .  \/   and __/\__   for those with inside bits, but didn't follow
> .             \/     the 'rules' exactly and only replaced some segments.
> 
> or using
> 
> \  and ____/\__    (two segments on the left)
> you get:           (slanted to avoid wrapping tho it get long)
> 
>    \
> 
> .  \
> .   \__
> .     /
> .     \
> 
> .  \
> .   \__
> .     /
> .     \
> .      \
> .       \__
> .         /
> .         \____/\__
> .                 /
> .                /
> .               _\
> .              /
> .              \
> .               \__
> .                 /
> .                 \
> etc.
> .                         __
> and a triange based one ( \/ and ____/\__ ) that i think looks kinda neat:
> 
>                          __/\
> .                        \   \__
> .                        /     /
> .       ____/\______/\__/      \____/\__
> .       \                              /
> .       /_                            /
> .         \                          _\
> .          \                        /
> .        __/                       /
> .        \                        /
> .        /                       _\
> .       /_  ____                /
> .         \/    \               \
> .               /_               \__
> .                 \                /
> .                  \        __  ___\
> .                   \      /  \/
> .                   /_    /
> .                     \  _\
> .                      \/