Print

Print


Hello Honza,

     Congratulations with your project, so nice to see a sign language 
to sign language dictionary and it also supporting several spoken languages.

P.S. for those of you who would like to see the 
https://beta.dictio.info/ site google translated to another language go 
to 
https://translate.google.com/translate?hl=&sl=auto&tl=en&u=https%3A%2F%2Fbeta.dictio.info%2F

I know what you mean that some symbols look the same in different 
positions, that's why the fist was drawn as Y.  I tried to make a symbol 
chooser that has 4 rotations instead of the common three. But sometimes 
I think it would be clearer if we were looking at the baby finger side 
of the hand or the thumb side of the hand.  I had though at one time to 
add a thumb nail to the thumb side of the fist symbols that don't have a 
thumb shown.

As the baby finger side of the right hand, is the same symbol as the 
thumb side of the left hand.

   But I never got to that project and probably never will.

And the first really need the arm symbol in the case of dictionaries 
when people are trying to read the symbol for the first time. as most of 
them are identical.



Sometime between 2011 and 2015 I worked on a few attempts to sort my 
dictionaries by symbols.  Which led me to have to make an algorithm to 
create syllables to tell which symbols came first in a sign and also to 
attempt to identify right and left hands in the sign.  As you and Adam 
have pointed out it's not very straight forward when the algorithm has 
to look at the arrows that are moving those hand symbols to know if the 
symbol is a right or a left.

This lead to my Spelling suggestion feature in SignWriterStudio.


It mostly gets signs ordered the way I programmed it. It's similar to 
the Sutton-Syllable-Sequence (ok maybe that's not what it's called but I 
know it's SSS) but with a few changes.

Sometimes when there are two sets of hands and it's not clear for it 
which arrow is moving which hand, then it asks.

And in this case I only select the first part of the sign so as to get 
too crazy.  Here the dominant hand is considered to the right, though 
the algorithm could also do the left as dominant if we set one of the 
settings.

I think if I would've made another construct to group some of the 
symbols into groups to analyze group by group it would be even better.

My plan was to run the algorithm on all the signs in my dictionary and 
not add the spelling if it needed to ask for more information.  Then go 
in manually and have it suggest the other ones and correct them if it 
got them wrong as it sometimes does.

I don't currently have it set up to run through all the signs in a 
dictionary.

You are welcome to use or inspire yourself from my code.  It is open 
source at https://github.com/JonathanDDuncan/Signwriterstudio

The sign spelling algorithm is here.

https://github.com/JonathanDDuncan/SignWriterStudio/blob/4865a876dbbf4d4456bf45292dc0822d475740ee/SWEditor/SignSpelling.vb


I was also working on a way to sort the hand symbols by shape similarity 
different from the ISWA with SortWeight, an extra column in the database 
for a different number for sorting purposes. Which when run on the 
sorting sequence gave a string that would sort the dictionary entries in 
that order.  You might want to stop at the previous step.  But I am 
including it here in case you wanted to use it.

https://github.com/JonathanDDuncan/SignWriterStudio/blob/fa37f48e64ea140085146615f85a6885a9705ef6/SWClasses/Dictionary.vb

If you have funds for spelling and sorting algorithm and you want my 
help, let me know.


Jonathan


On 10/29/2020 2:23 PM, Honza wrote:
> Hello Adam,
>
> Thank you very much for your answer and explanation. I don't know why, 
> but I have always thought all symbols are divided in two halves as 
> follows: upper part right hand, lower part left hand.
>
> In the past, we used some symbols from the lower part, since this 
> symbols are not included in the first part, but we thought it is not 
> correct since it is the symbol designated for left hand. Ok, that is 
> my misunderstanding for all that years...
>
> Ok, now people will ask me, why this "forth" symbols are not sorted on 
> behalf the three standard ones.. I see there is a reason, but from the 
> perspective of rotations, it would be good to have it in one row (but 
> some twice, I know).
>
> Adam and Val, we use SignWriting symbols for sign search in our 
> dictionary and we use the distinction of left and right hand symbols 
> (upper and lower part), so we are able to search among one-hand signs 
> and two-hand signs separately. Now if there is in principle no 
> difference between left and right hand symbols, there are only 
> movement arrows left for sorting one and two-hand signs, but there is 
> nothing left when the writing contains only handshapes and contact 
> (star) symbols. Do you have an idea, how to sort them?
>
> Honza
>
> On Thu, 29 Oct 2020 at 13:47, Adam Frost <[log in to unmask] 
> <mailto:[log in to unmask]>> wrote:
>
>     Hello Honza,
>
>     If I understand your question correctly, you are asking about this
>     symbol being for the dominant hand:
>
>     Below is the sign that you have written:
>
>     It looks well written and matches the video that was provided.
>     (Even when noticing that the signer was left handed and the
>     writing is for right handed dominant signing.)
>
>     It seems the confusion comes from the fact that the handshape
>     symbol is found in SignPuddle on the lower half of the grid when
>     showing all of the rotations. People generally assume that the
>     rotations are sorted to be “right hand” on the upper half and
>     “left hand” on the lower half.
>
>     While this does generally turn out to be true, this isn’t how the
>     rotations are sorted. The symbols are only sorted by rotating
>     counter-clockwise then flipping and rotating clockwise. There is
>     no consideration for “right hand” or “left hand” as the symbols do
>     not hold such inherit handiness in their meaning.
>
>     If I were to take your initial sign and use a different palm
>     facing - facing towards the back - you will notice that the
>     handshape symbol is now on the upper half of the rotation level’s
>     grid:
>
>     So, there is no information about handiness in the handshape
>     symbol in of itself. To further illustrate my point, consider the
>     writing below:
>     Notice that I use the same handshape symbol. We know that it is
>     the left hand, not because of the handshape symbol, but because of
>     the arrowhead. The handshape symbol only tells us the handshape,
>     the fingers are pointing down, and that the palm is facing towards
>     the left.
>
>     So there is nothing wrong with the writing you had. It was written
>     correctly. It is, however, a perfect example of how the rotating
>     order does *not* actually imply right hand or left hand for
>     handshape symbols. ;-)
>
>     If you wanted to take into consideration of the rotation of palm
>     facing of a hand beyond the standard 3 palm facings, it could be
>     done as followed:
>     Notice that the last palm facing of this series is generally not
>     displayed on the grid because of two reasons: it can be found on
>     the other half of the rotations by using the second symbol, and it
>     is generally not used for most sign languages. This choice was
>     made to reduce space and redundant symbols.
>
>     So if I were to using the same ordering of palm facing rotation
>     with the fingers point down, I would get this:
>     The last symbol is the one that you are writing with. If the
>     reduction was not made, then you would be able to find this symbol
>     on both halves of the grid because the fourth column would just be
>     the same as the second with the upper and lower halves switched.
>
>     I hope this explanation helps understand why it is best not to
>     think of the rotation grid as "upper half" = "right hand" and
>     "lower half" = "left hand”. Especially since the side palm facing
>     can double up as both as shown above.
>
>
>     Adam
>
>>     On Oct 29, 2020, at 12:36 AM, Honza <[log in to unmask]
>>     <mailto:[log in to unmask]>> wrote:
>>
>>     Hi Val,
>>
>>     I hope, you are still going well ;)
>>     We are facing problem with handshapes orientation with fingers
>>     pointing to the ground.
>>     See this Czech Sign Language sign
>>     https://beta.dictio.info/czj/search/text/10627
>>     <https://beta.dictio.info/czj/search/text/10627> (first video)
>>     (note that the signer is left handed, but SW is always switched
>>     on right hand for ther purpose of the dictionary.)
>>     Sign is already written down using SW, but with incorrect
>>     handshape (just to show better the orientation - without finger
>>     it is not very good visible in this case).
>>     There is a symbol showing this orientation, but it is non
>>     dominant hand and that is not correct.
>>
>>     So it seems to me, we are missing symbols with this orientation
>>     (fingers pointing to the ground).
>>
>>     We use SignWriting as well for searching, so it is necessary tu
>>     use correct hand.
>>
>>     Thanks for your ideas.
>>     Honza, Czechia
>>
>>     ________________________________________________
>>
>>     SIGNWRITING LIST INFORMATION
>>
>>     Valerie Sutton SignWriting List moderator [log in to unmask]
>>     <mailto:[log in to unmask]>
>>
>>     Post Messages to the SignWriting List:
>>     [log in to unmask]
>>     <mailto:[log in to unmask]>
>>
>>     SignWriting List Archives & Home Page
>>     http://www.signwriting.org/forums/swlist
>>     <http://www.signwriting.org/forums/swlist>
>>
>>     Join, Leave or Change How You Receive SW List Messages
>>     http://listserv.valenciacollege.edu/cgi-bin/wa?SUBED1=SW-L&A=1
>>     <http://listserv.valenciacollege.edu/cgi-bin/wa?SUBED1=SW-L&A=1>
>>
>
>     ________________________________________________
>
>     SIGNWRITING LIST INFORMATION
>
>     Valerie Sutton SignWriting List moderator [log in to unmask]
>     <mailto:[log in to unmask]>
>
>     Post Messages to the SignWriting List:
>     [log in to unmask]
>     <mailto:[log in to unmask]>
>
>     SignWriting List Archives & Home Page
>     http://www.signwriting.org/forums/swlist
>     <http://www.signwriting.org/forums/swlist>
>
>     Join, Leave or Change How You Receive SW List Messages
>     http://listserv.valenciacollege.edu/cgi-bin/wa?SUBED1=SW-L&A=1
>     <http://listserv.valenciacollege.edu/cgi-bin/wa?SUBED1=SW-L&A=1>
>
> ________________________________________________
>
> SIGNWRITING LIST INFORMATION
>
> Valerie Sutton SignWriting List moderator [log in to unmask] 
> <mailto:[log in to unmask]>
>
> Post Messages to the SignWriting List: 
> [log in to unmask] 
> <mailto:[log in to unmask]>
>
> SignWriting List Archives & Home Page 
> http://www.signwriting.org/forums/swlist 
> <http://www.signwriting.org/forums/swlist>
>
> Join, Leave or Change How You Receive SW List Messages 
> http://listserv.valenciacollege.edu/cgi-bin/wa?SUBED1=SW-L&A=1 
> <http://listserv.valenciacollege.edu/cgi-bin/wa?SUBED1=SW-L&A=1>
>

________________________________________________


SIGNWRITING LIST INFORMATION

Valerie Sutton
SignWriting List moderator
[log in to unmask]

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

SignWriting List Archives & Home Page
http://www.signwriting.org/forums/swlist

Join, Leave or Change How You Receive SW List Messages
http://listserv.valenciacollege.edu/cgi-bin/wa?SUBED1=SW-L&A=1