the benefits of ELAN

[Steve Slevinski]

Hi Mark,

Mark A. Mandel wrote:
> I can't believe that there are 33563 individual symbols in SW. I assume that 
> these are combinations of a much smaller number of elementary symbols, in 
> different spatial relationships -- a small enough set to include in UTF-8. It 
> must be possible
>   
The International SignWriting Alphabet 2008 beta has 33,563 symbols in 
the symbol set.  Each symbol has a unique name.

There are 30 SymbolGroups and 600 BaseSymbols.  Half of the BaseSymbols 
are hand shapes.  This symbol set is for writing any sign language, so a 
subset for a specific language will be much smaller.  The symbol set 
also includes advanced writing symbols for detailed writing and sorting.

Each hand shape is a specific BaseSymbols.  Each hand shape can have 8 
rotations, 8 mirrored rotations, and 6 orientations.  This is a total of 
96 symbols for each hand shape.

There are currently 512 spaces in Unicode assigned for SignWriting.  We 
could possibly compromise and encode by BaseSymbol rather than symbol, 
but I don't see the benefit.  We don't write by BaseSymbol, but by symbol.

SignWriting is spatial and not sequential, so even if SignWriting was 
encoded in Unicode it could not be used in ELAN without a custom plugin.

Simplicity is the rational why we encode by symbol rather than 
BaseSymbol.  Writing is symbols in space.  Each symbol is a character 
and has a 16 bit code.  If we used BaseSymbols instead, we'd need 8 bits 
for the BaseSymbol, 3 bits for the orientation, and 4 bits for the 
rotation (saving 1 bit).  Instead of a series of codes and coordinates, 
we'd need a series of codes with modifying codes and coordinates. 

The same can be said for sorting.  Each sign has a spelling for it's 
visual appearance and a sequence of symbols for sorting.  Currently, the 
sorting sequence is a simple sequence of codes.  If we encoded by 
BaseSymbol, the sorting sequence would be a series of codes and 
modifying codes.

The extra complication of encoding by BaseSymbol and using modifying 
codes makes using SignWriting more difficult without adding any benefit 
other than fitting better in the previous Unicode allotment.  And I 
repeat, even if SignWriting was available in Unicode it would not be 
usable in any application because we do not use SignWriting sequentially.

With the current character encoding model, we use 16 bits for the coded 
character set (x-iswa-208)  This is compatible with Unicode if they open 
a new plane for SignWriting.  I don't think this is unreasonable because 
the ISWA can be used with all sign languages.

Regards,
-Steve

_______________________________________________
SLLING-L mailing list
SLLING-L at majordomo.valenciacc.edu
http://majordomo.valenciacc.edu/mailman/listinfo/slling-l