IMWA: Handshape Construction and Sequencing Rules

[Steven Aerts]

> I agree. But we should think for the future. Every symbol of
> SignWriting will have a representation in Unicode. And Unicode
> characters just are numbers. So, what I am trying to tell you? Even if
> you use names ultimately you will have to work with numbers again.
We don't think unicode is a good standard to write SignWriting.

The main (and only?) reason to use unicode is that you can make a font for it.  
The difference between fonts and SignWriting is that SignWriting can rotate 
and mirror it's symbols and fonts - which are purely character based - can't.  
So the fonts will have to contain a representation for all possible 
SignWriting symbol mirrors and rotations, which will make the font too big to 
handle for a normal computer.  We never expect that you can open a unicode 
SignWriting text in your browser by only downloading the SW-font.
If we can't use SW-unicode anywhere else expect in our own programs why use 
unicode if there is a better alternative: SWML?  That format is extensible in 
nature, has more (programming)-tools available, is more human-readable and is 
most likely smaller if compressed.

> I don't agree. A rotation of 45 degrees is not something you measure
> with "degree". It is just oblique. Why? I think it is the human
> perception of the steepness. Take a slope you are skiing down. Tell me
> how steep the slope is? You don't tell me it is 35 degrees. You tell me
> it is rather steep but manageable. SignWriting is not exact science.
>
> In a summary the rotation numbers aren't something you express exactly
> with degrees. It is vertical, oblique to the right, to the left,
> horizontal to the right, to the left, etc.
All right but we don't think the user has to know how the angle is represented 
when he types in his signs.  We do think however it is important to define an 
open and understandable format.  
We don't think the present format is understandable because if we give you a 
symbol with rotation 1 you will have to rotate it  45° in normal cases, 15° 
if it is a shoulder or -45° with some special variations of some special 
hands.  This is maybe manageable if you have only 256 symbols to work with 
but not in the future symbolsets.
We don't think it is open if you represent a rotation with cardinal numbers 
between 0 and 8: if for some reason in the future someone wants to rotate a 
hand 15 degrees he has to define a new standard.

We hope this becomes an interesting discussion,

        Bart & Steven