<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000000">
<br>
<br>
Sandy Fleming wrote:
<blockquote cite="mid1180168130.7593.30.camel@localhost.localdomain"
type="cite">
<pre wrap="">Jonathan,
It will be compatible in the sense that I'll provide features that make
it possible to store and read those other formats, and to convert
between my formats and those.
I've tried very hard to use the numbers in the Sutton Symbol Sequence
when storing the signs, but have given up on it. If you want to program
with complete freedom and to be able to store the orientations and so on
with precision, then you need to take an engineering approach, not the
SSS integer approach.
</pre>
</blockquote>
Hi Sandy,<br>
My program is using the numbers in the Sutton Symbol Sequence for
working with the signs. But it won't have all of the precision yours
does. I might use the database register number to actually store it
internally but that can easily be converted back to the SSS number. I
use the SSS without all of the dashes. But is should be easy to put
them back in when converting it to an XML format for sharing with
others.<br>
Do you feel the a format like the SWML-S would be adequate if it
had the sequence added to it? Do you think that colors should be
specified in the format too? Will you keep track of which symbol is
the right and and which is the left? Mine will. If my program is the
only one that tracks this then it might not be a good idea to put it
into the XML format. But if yours does then maybe we could include it.<br>
<blockquote cite="mid1180168130.7593.30.camel@localhost.localdomain"
type="cite">
<pre wrap="">What I do use, however, are the symbol _names_ in the SSS. These are
perfect for labelling symbols from within the program (whether you use
strings or variable names) and don't have the fill or rotation, which we
want to express in engineering terms rather than as a proprietary (and
chunky!) system of numbers.
So instead of:
"01-01-001-02-02-01"
I use:
"IndexOnCircle"
to label the symbol, plus variables holding the rotation, fill,
position, and symbol category.
Rotation is in degrees from vertically upwards (because I always start
off by thinking of handshapes symbols with their fingers vertically
upwards): normally anticlockwise, but for the left hand, clockwise.
Fill is a decimal number between 0 and 1 inclusive.
Position is expressed as the (x, y) distance from the origin. It's
important that the units used relate to the distances within the sign
and not something that can change such as the size at the screen. So I
use a handsbreadth as the unit of distance (the length of a side of the
fist symbol or, equivalently, the diameter of the circle handshape).
The category number is useful to store so that the program can use it to
know how to treat different symbols (head symbols behave differently
from hand symbols, which behave differently from movement symbols and so
on) although I have to adjust it for some things. For example, I treat
the features of the head (eye, mouth, nose, head circle etc) as symbols
in their own right, I don't try to use entire facial expressions as
separate symbols, because this causes a combinatorial explosion.
</pre>
<blockquote type="cite">
<pre wrap=""> The program I am working on is staying more conservative in using
strictly the IMWA. I think that it will use the next version of the
ISWA once it is ready. I was asking the above questions because I am
thinking about how my program could exchange information with other
SignWriting programs and thought that it would be insightful as to how
others plan on changing some of the technical details of computer
SignWriting for even better options and how this will impact sharing
documents between programs.
</pre>
</blockquote>
<pre wrap=""><!---->
Since my program stores more detailed information than other programs, I
have no problem writing something to store my data in other formats: the
other programs will lose some of the information, but that's OK since
they don't use it. Conversely, when reading a different format such as
SWML I can supply the missing information as defaults or calculations
within my program.
</pre>
</blockquote>
I feel the same for my program. Not everybody will use the information
in their formats. It will be good to have one format to interchange
the basic information.<br>
<blockquote cite="mid1180168130.7593.30.camel@localhost.localdomain"
type="cite">
<pre wrap="">Of course none of this allows me to store the sorting sequence within my
program, but that's no problem, I'll just make up a table to map between
the SSS numbers and SSS naming system. We'll need to try to share stuff
like this!
Sandy
</pre>
</blockquote>
Jonathan
<div class="moz-signature">
<div class="moz-signature">
<address><font face="Comic Sans MS"><big><br>
</big></font></address>
</div>
</div>
</body>
</html>