Overall Directionality

Hi all,

    I've started to work on implementation of the overall 
directionality. I've written a testcase (see attachment) that would 
certainly be useful to discuss the expected rendering and help to test 
implementations. For most presentation elements I think the way they 
have to be drawn in RTL is clearly described so I just give a few remarks:

- For some elements such that mpadded or mfrac, I think it could be 
useful to give a reference to section 3.1.5.1 to recall how the terms 
leading/trailing and left/right are used.

- For mmultiscripts, the sentence "It supports both postscripts (to the 
right of the base in visual notation) and prescripts (to the left of the 
base in visual notation)" does not take into account the RTL case and is 
redundant with what is said at the beginning of the parent section: 
"Note that ordinary scripts follow the base (on the right in LTR 
context, but on the left in RTL context); prescripts precede the base 
(on the left (right) in LTR (RTL) context).".

- For menclose, the directionality is mentioned twice:

    "The case of notation="radical" is equivalent to the msqrt schema 
(and thus affected by the directionality)."

    "(Note that "updiagonalstrike" ("downdiagonalstrike") is taken to be 
a line from the lower-left to upper-right (upper-left to lower-right, 
resp.), independent of directionality.)"
   
    As I understand, only radical depends on the directionality. Hence 
rather than these two comments in parentheses I suggest to do as in 
other sections i.e. use a separate sentence where all the discussion 
about the directionality of menclose is moved. I think something 
containing "Among the notations from the recommended list, only radical 
is affected by the directionality" could indicate clearly what is 
expected for menclose.
   
- For mfenced, directionality is well described via the equivalent 
markup, so I just give an off-topic remark: I think that the two 
equivalences given in section 3.3.8.1 are, strictly speaking, not the 
same as the general equivalence of 3.3.8.2. The parenthesis and comma 
are marked as fence="true" and separator="true" in the Operator 
Dictionary but this one is just a suggestion and an implementation may 
not treat them as fences/separator.

Regards,

Frédéric Wang