[rescue] octane question

[Joshua D Boyd]

On Thu, Jan 17, 2002 at 11:15:51PM -0500, s at avoidant.org wrote:

>>Now, another project I've been kicking around is a idea for a new type of
>>symbolic math system.  The initial revision would be handwriting recognition
>>of some sort (probably a modified graffiti) coupled to a system like Maxima

> May I suggest XMerlin? Much easier to modify that Graffiti.
> 
> http://www.hellkvist.org/software/index.php3#XMerlin

I've bookmarked it for close inspection.  My initial impression is that it does
more than I want.  When I said a modified graffiti, I meant graffiti in the 
sense of the idea of it.  I wouldn't actually be basing my work on graffiti's 
code.

I planned to start with something like libstroke (although libstroke might be
too simplistic for my needs).  It would need to be modified a hair though.
Basically how libstroke works is that it tells you what stroke was entered, 
and you match the strokes up with what they mean.  For instance, one stroke
is the letter a, another is the number 9, another could be the integral 
symbol.  For my idea to work, I need to know what the stroke was, but also
where it was entered.

Oh, looking through my notes, I find that I previously had decided that wayV
(http://www.stressbunny.com/wayv/) looked better because it has finer grained
feedback of what was entered.

What I really want to get away from with my system is the idea that palm pilots
have about how data is entered.  You enter it at the bottom of the screen in a
special area, and it is essentially like you are using a keyboard, you apps 
can't tell the difference.  What I want to do is preserve the spacial elements
to writing math.  I want to actually move my hand from left to write and be
able to properly enter superscripts and subscripts.  Getting the formula
entry system down is the first, and probably hardest step to my desired system.


-- 
Joshua D. Boyd