On 8/14/2014 10:37 AM, Stephen E Slevinski Jr wrote:
> Regarding the mechanics of sorting, the first part of a sign's name is 
> an (invisible) ordered string of symbols that default to the standard 
> ISWA 2010 collation everywhere.  The signs will sort properly within 
> an Excel spreadsheet, in any database (SQLite, MySQL, ...), in any 
> programming language (JavaScript, Python, ...).  Anywhere ASCII is 
> supported, the signs will sort according to the standard ISWA 2010.
Completly agree with you here.
> Creating a customized collation order of the ISWA 2010 is possible 
> with a symbol key rewrite.
Wouldn't a symbol key rewrite mean changing the keys for the symbols?
I had thought of something similar but not changing the symbol key. We 
would create a new number the sort_weight.  We could combine a 
sort_base_code, a sort_fill_order, sort_rot_order to get this number.

In SignWriter Studio I would add four columns to my symbol table.

For hands, both hands have the same sort weight either right hand or 
left hand. See attached excel spreadsheat.

To sort no contact, then hands contacting hands, then  contacts of head, 
then contact of body can be really tricky because there currently isn't 
any way programatically to determine whether a hand is contacting 
another hand.  There are some symbols to show that there is contact the 
head or the body, these should work.

For heads modifying sort_base_code to be all the same, sort_fill_order 
to 1, sort_rot_order to number they should be in sequence

See attached excel spreadsheat.

Then the ordered string fo sort_weight string will be created for the 
signs's SignSpelling Sequence.  The equivalent Unicode value for the 
sort_weight can be used in  DUCET table.

Which puddles would be the best to try this sorting on so we can check 
to see if it is sorting the way we want?  It would have to already have 
the SignSpelling Sequence filled out for most of the signs.

I still don't see a way to sort on symbols that don't exist in the 
ISWA2010 like the one used in LIBRAS for dinosaur until we add them to 
the symbols set because they don't exist yet.  This would be for a 
future version of the symbol set.

Comments welcome

> For the ASCII, a small text file contains the rules.
> Key S10f becomes S10e
> Key S10e becomes S10f
> For the Unicode, entries in the DUCET table are used to adjust 
> collation weights.
> U+FD82F becomes U+FD92E
> U+FD92E becomes U+FD92F
> For advanced collation needs, you can rewrite entire keys. Charles 
> could rotate the Head Rims clockwise rather than counter-clockwise if 
> he needed.
> Head Rims
> The clockwise sort can be enabled with the following rules.
> Key S30007 becomes S30001
> Key S30006 becomes S30002
> Key S30005 becomes S30003
> Key S30003 becomes S30005
> Key S30002 becomes S30006
> Key S30001 becomes S30007
> Whatever Detailed Location symbols Charles requires, he can customize 
> the sorting to an order of his choosing.
> For Andre, regarding rotation and sorting, that is possible with a 
> specially written SignSpelling Sequence.  Any rotation used in the 
> 2-dimensional sign box could be ignored for the SignSpelling Sequence.
> For the sign, in the sign box we would have this 2-dimensional 
> arrangement:
> For the sorting, the symbols in the SignSpelling Sequence do not need 
> to be the same symbols as in the sign box.  Here is an order that 
> would sort the index hand above the head without rotation.
> First is, then
> Historically, sorting has always been an issue for every script. In 
> 11th century, the Song Chinese developed the movable type printing 
> press.  Great printing houses each developed their own blocks and 
> unique characters.  Many words were common between all houses and some 
> words were unique for each house.  A huge volume of blocks were 
> accumulated over time.  Sorting these blocks became an issue.  For 
> complex blocks, little slips of paper were attached to the block to 
> properly explain the sorting.
> SignWriting is very similar, except for uniquely carved 2-dimensional 
> blocks, we have a name with 2-dimensional order. These names can be 
> created, copied, and searched with ease.  For sorting, instead of a 
> slip of paper explaining the sort, we have an invisible prefix that 
> orders according to a theory.
> I believe Valerie's design of the SignSpelling Sequence is the most 
> productive theory.  I believe it is the easiest to use for readers, 
> writers, and programmers.  Outside of Valerie's theory, a wide variety 
> of possibilities can be supported in the standard model.
> Regards,
> -Steve
> ________________________________________________
> Valerie Sutton SignWriting List moderator [log in to unmask]
> Post Messages to the SignWriting List: [log in to unmask]
> SignWriting List Archives & Home Page 
> Join, Leave or Change How You Receive SW List Messages 



Valerie Sutton
SignWriting List moderator
[log in to unmask]

Post Messages to the SignWriting List:
[log in to unmask]

SignWriting List Archives & Home Page

Join, Leave or Change How You Receive SW List Messages