[MathML3-last-call] RTL and mlabeledtr

[Neil Soiffer]

I didn't see anything that said what to do with the label for mlabeledtr.  In fact, it doesn't say what to do with the main table elements although it is not hard to guess given what it says in mtr, but it probably should be explicit.  For the label, mtable's "side" attr controls the placement, so if something should be done for RTL, that seems like that's where a statement should be made.

Neil Soiffer
Senior Scientist
Design Science, Inc.
www.dessci.com
~ Makers of MathType, MathFlow, MathPlayer, MathDaisy, WebEQ, Equation Editor ~

[Bruce Miller]

On 10/22/2009 10:50 AM, Neil Soiffer wrote:
> I didn't see anything that said what to do with the label for
> mlabeledtr.  In fact, it doesn't say what to do with the main table
> elements although it is not hard to guess given what it says in mtr, but
> it probably should be explicit.  For the label, mtable's "side" attr
> controls the placement, so if something should be done for RTL, that
> seems like that's where a statement should be made.

The label itself is positioned left or right according
to the mtable's side attribute.  I think we been
consistent with the interpretation that "left"
always means "left", even in an RTL context, so that
we shouldn't need to reinforce that.

The only exceptions are things like lspace, rspace where
we could (and had to) fudge the interpretation to
"leading space"; and of course things we've inherited
from elsewhere (eg. "LEFT PARENTHESIS").

For the remaining children, the 1st paragraph of mlabeledtr
says:
   "The rest of the children represent the contents
   of the row and are identical to those used for mtr;..."

[Bruce Miller]

Neil Soiffer wrote:
> I didn't see anything that said what to do with the label for
> mlabeledtr.  In fact, it doesn't say what to do with the main table
> elements although it is not hard to guess given what it says in mtr, but
> it probably should be explicit.  For the label, mtable's "side" attr
> controls the placement, so if something should be done for RTL, that
> seems like that's where a statement should be made.

I've clarified that the children of mlabeledtr (after the first)
are treated identically to the children of mtr.

I've also split out a paragraph in the bidi section to make clear
from the outset what leading and trailing mean, and that otherwise,
left means left, etc.

In trying to be explicit without being repetitive, it became clear
that the menclose cases updiagonalstrike and downdiagonalstrike could
suggest a dependence on directionality.  Discussions within the group
decided that they should not have such dependence and so I have added
clarifying text there, as well.

Checking in spec/xml/presentation-markup.xml;
/w3ccvs/WWW/Math/Group/spec/xml/presentation-markup.xml,v  <--  presentation-markup.xml
new revision: 1.295; previous revision: 1.294
done


> Neil Soiffer
> Senior Scientist
> Design Science, Inc.
> www.dessci.com <http://www.dessci.com/>
> ~ Makers of MathType, MathFlow, MathPlayer, MathDaisy, WebEQ, Equation
> Editor ~