[charmod-norm] Canonically equivalent but may have different appearance.

ntounsi has just created a new issue for https://github.com/w3c/charmod-norm:

== Canonically equivalent but may have different appearance. ==
(May be  a font rendering problem, but worth to notice)

Same example as previous issue #162 , but with Kasra instead of Fatha. Kasra (U+0650) normally appears bellow Shadda (U+0651) and sometime bellow a letter. So regardless of the order  
![b-k-s](https://user-images.githubusercontent.com/15004708/38435158-c50a7cf4-39c8-11e8-8bd1-b2125e10898d.jpeg)  

or   

![b-s-k](https://user-images.githubusercontent.com/15004708/38435177-cc6f6e5a-39c8-11e8-8b97-2ebebfdfc8f5.jpeg)

you may have:

(A) 
![bks2](https://user-images.githubusercontent.com/15004708/38435247-075b71da-39c9-11e8-91f6-b91fc0bfa2b1.jpeg) (with Amiri font for example )

or

(B) 
![bks](https://user-images.githubusercontent.com/15004708/38435265-13c514d0-39c9-11e8-8fa3-bd000788ef28.jpeg) (with other fonts, more simple IMO)

But, for some (broken IMO) fonts, according to the order of combining marks, if Kasra is _before_ Shadda, the appearance _is_ like (A) above, and if Shadda is _before_ Kasra, it _is_ like (B).
However, the principle of canonical equivalence, at ยง2.2.1, second paragraph, states that "_**When correctly displayed**_ [ sequences of Unicode characters that represent the same abstract character] _should always have the same visual appearance and behavior_". 

The last example appears to contradict this principle... The appearance differs according to the sequence order of the pair of diacritics. Unless one notices what is, carefully, said "When correctly displayed". 
Does the example here, (A) for Beh+Kasra+Shadda and (B) for Beh+Shadda+Kasra, fit in the case of "correctly displayed" ? 



Please view or discuss this issue at https://github.com/w3c/charmod-norm/issues/163 using your GitHub account

Received on Friday, 6 April 2018 17:45:30 UTC