Re: Prime

> What is the correct representation of "X'" in MathML: (1) or (2)?
> 
> (1)     <math>
>           <msup>
>             <mi>X</mi>
>             <mo>&#x2032;</mo>
>           </msup>
>         </math>
> 
> (2)     <math>
>           <mrow>
>             <mi>X</mi>
>             <mo>&#x2032;</mo>
>           </mrow>
>         </math>


(1) is the correct way to markup up x'.  The Math WG has taken up this
question several times, and always agreed that the superscript
notation better captures the structure, and it the preferred markup.

With X' there isn't much difference, but with (f+g)' and so on, method
(2) is as bad as superscripting a parenthesis in (f+g)^2.

> However, the glyph corresponding to &#x2032; in standard Unicode fonts
> (for instance, the X11's Misc-Fixed fonts or Microsoft's Arial) is
> analogous in height to &#x27;, the ASCII apostrophe. 

Yes, this is true, and the Unicode description of the character states
explicitly that this is the raised, shrunken, accent version of the
character.  Of course this is at odds with TeX and (as indicated
above) MathML usage, which is a real drag.

> Moreover, only (2) is rendered as intended by Mozilla.

Since (1) is preferred, this is a bug in Mozilla.  

However, since the Unicode description predates, and is at odds with,
what MathML wants to do to encode structure, its an understandablebug.  

What MathPlayer does is detect both situations and "correct" the size
and positioning of the prime character when it occurs in a script. 

Unfortunately, what that means is that it is possible to use the "wrong"
markup in a cross-platform way, while it is not possible to the
"right" markup.  That is really irritating, since it inevitably means
that people will start generating "wrong" markup, and we will be stuck
with it forever.  

I don't really know what do do about it.  Any ideas would be welcome.

--Robert

------------------------------------------------------------------
Dr. Robert Miner                                RobertM@dessci.com
MathML 2.0 Specification Co-editor                    651-223-2883
Design Science, Inc.   "How Science Communicates"   www.dessci.com
------------------------------------------------------------------

Received on Tuesday, 19 November 2002 13:18:27 UTC