Rotations

[Daniel Noelpp]

Hello Valerie, hello Bart and Steven, hello SignWriting list!

Bart Braem wrote:

> One other thing we would like to ask: please make all symbols turn in
> the same direction, so all counter-clockwise for example. Currently
> some
> symbols do turn the other way round and that's pretty irritating when
> programming: if you want to display a mirrored or rotated symbol you
> always need to check the category. When you just specify one rotation
> programmers can always be confident that when the format reads
> "turn 45 degrees" it will always be the same direction.

I am not sure that I understand you. Are you telling, that there were
some "weird" symbols rotating the other way round? I didn't find any
symbols in any category rotating differently than the other symbols in
the same category except for symmetry. My brother did a lot of detail
work when we configured the symbolset together. So I have to ask him
whether he discovered something strange that I might have overlooked.

Perhaps we are talking about different things:

1. About the ordering of rotations whithin the Sign Symbol Sequence, or
2. the rotation code within the files or
3. the exact behaviour of the rotation key?

I don't know. I am going to try explain my view about point 3 and what
I have discovered so far. Let's start with an example: the Index Finger
of both hands pointing forward and then both hands turned 45 degrees
outward:

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: text/enriched
Size: 1437 bytes
Desc: not available
URL: <http://listserv.linguistlist.org/pipermail/sw-l/attachments/20040605/efd3d8b5/attachment.bin>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: rotation.PNG
Type: image/png
Size: 678 bytes
Desc: not available
URL: <http://listserv.linguistlist.org/pipermail/sw-l/attachments/20040605/efd3d8b5/attachment.png>
-------------- next part --------------


If you do the Index Finger together with both hands, both symbols are a 
mirror of each other. If you turn the left hand about 45 degrees to the 
left then you have to turn the right hand the same amount to the right 
so they still are a mirror of each other. That's why the mirror was 
turned in the other direction and this is very natural.

The problem I see with this approach: It is not immediately evident or 
very visually clear that something is "righthanded" or "lefthanded". It 
will confuse and even bewilder users when they rotate things. And 
voil?: Neither SignWriter DOS nor Java rotate depending on mirror. They 
rotate always counter-clock-wise (and with the Shift-key pressed 
clock-wise).

In GUI designing there is the principle of the least surprise. An 
action should contain as few surprises as possible and always work the 
same as expected. Making rotation dependent on mirror would bring a lot 
of surprise and suspense.

I have to confess that I still need some thinking about this subject, 
especially about point 1 and 2. I agree with Steven and Bart that 
rotations should always have the same direction, but not because the 
programmers get confused, but because the users could get confused.

In the SymbolBank the mirrors of the hand symbols rotate the other way 
round. Perhaps it is a good idea to leave it this way? So we have the 
best of the two different possibilities? The GUI (the rotation key) 
always rotates the same, but the data knows about the mirror and 
rotates differently for the mirrors?

Daniel
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: text/enriched
Size: 1551 bytes
Desc: not available
URL: <http://listserv.linguistlist.org/pipermail/sw-l/attachments/20040605/efd3d8b5/attachment-0001.bin>