From: David Carlisle <davidc@nag.co.uk>

Date: Wed, 22 Aug 2012 00:48:49 +0100

Message-ID: <50341E61.2070105@nag.co.uk>

To: "'www-math@w3.org'" <www-math@w3.org>

Received on Tuesday, 21 August 2012 23:49:17 GMT

Date: Wed, 22 Aug 2012 00:48:49 +0100

Message-ID: <50341E61.2070105@nag.co.uk>

To: "'www-math@w3.org'" <www-math@w3.org>

Recently Alexander Krutik asked an interesting question on the mathjax list about the proper display of mmultiscripts. the thread is archived at: https://groups.google.com/d/topic/mathjax-users/SAmlSkWQ3TY/discussion It turns out that MathML (1 2 and 3) are less than explicit about how the scripts should be aligned, it says that they should be "vertically-aligned " this is distinct from msubsup where the positioning can take account of the italic correction of a base letter. (Compare P_1^2 and {P{}}_1^2 in TeX which produce the attached ps.png results). However the specification does not currently state whether the sub/sup pairs should be left or right aligned. The test suite only has examples with single letters so does not produce a very good guide for horizontal alignment. In order to check the current state of implementations we tested the following file with results being attached. <!DOCTYPE html> <html> <head> <!-- <script type="text/javascript" src="http://cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML"> </script> --> <title>mmultiscripts test</title> </head> <body> <math xmlns="http://www.w3.org/1998/Math/MathML"> <mmultiscripts> <mi>A</mi> <mn>1</mn><mn>111</mn> <mn>2</mn><mn>222</mn> <mprescripts/> <mn>3</mn><mn>333</mn> <mn>4</mn><mn>444</mn> </mmultiscripts> </math> </body> </html> test case of various systems rendering the above, (without the script element except for mathjax) pmml2tex: left aligned (tex.png attached) firefox 17 (nightly) center aligned (ff.png attached) mathjax (in Chrome) left aligned (mj.png attached) mathplayer 2 center aligned (mp2.png attached) After some discussion the current feeling is that we should have an errata to the spec (equivalently an update to the editor's draft) that 1. Points out that the spec currently does not specify horizontal alignment, so systems are free to align as they desire 2. Mentions that an attribute to control alignment may be added in future versions of MathML. 3. Mentions that sub/sup prescripts of different lengths are more common in chemistry and in that case it is best to right align prescripts and left align postscripts; because of the difficulty of recognizing that use case, it may be a good default alignment. An example of the Chemistry use case might be seen in the mhchem latex package documentation where the example \ce{^{227}_{90}Th+} is given on page 6 of http://mirrors.ctan.org/macros/latex/contrib/mhchem/mhchem.pdf Thus might want the possibility of not always left-aligning prescripts. It might be noted that the similar \sideset macro from the amsmath latex package for mathematics left aligns prescripts. However (not entirely coincidentally) the AMS documentation doesn't explicitly mention horizontal alignment, and the examples typically just involve single digits. Comments as usual welcome on this list, David for the Math WG.

(image/png attachment: tex.png)

(image/png attachment: ff.png)

(image/png attachment: mj.png)

(image/png attachment: mp2.png)

(image/png attachment: ps.png)

*
This archive was generated by hypermail 2.2.0+W3C-0.50
: Tuesday, 21 August 2012 23:49:17 GMT
*