′ should result in $'$, not in $\prime$.

What do you mean? 'prime' is an overloaded character. Do you have a testcase/screenshot of what you see?

I got the following explanation from the reporter: > The MathML equivalent of LaTeX's $f'$ should be > <math><mi>f</mi><mo>&prime;</mo></math>, but > at the moment it's > <math><msup><mi>f</mi><mo>&prime;</mo></msup></math> It makes the bug report clear now. However, it isn't what the spec says. In fact, the issue came up in the n.p.m.mathml newsgroup and the WG confirmed that what Mozilla does is OK, though it was suggested that the other behavior could be an enhancement. So this bug is a RFE. Thread: http://groups.google.com/groups?hl=en&th=c907286f698426d8&rnum=1 Post from Robert Miner on the question: > The question of marking up primes has come up several times with the > Math WG, and everytime we have decided that in MathML, a prime should > be attached as a superscript, eg > > <msup> <mi>f</mi> <mo>&prime;</mo> </msup> > > The strongest argument is that this makes explicit to what the prime > applies. Updating the title to reflect the reality: OLD title was: "MahML entity &prime; is wrongly interpreted" NEW title is: "[RFE] render MathML &prime; as a mathematical prime in contexts other than <msup>" This could become a non-issue if generators become good enough.

What about when the prime is not an operator? e.g. is <msup> <mi>k</mi> <mo>&prime;</mo> </msup> ...preferred over <mi>k&prime;</mi> ...when trying to tell the difference between the variable /k/ and the related but distinct variable /k'/ ?

That's is all quirky... "<mi>k&prime;</mi>" is always understooned means the identifier k' (i.e., concat of a glyph for k and a glyph for ' -- nothing special expected). It is a very bad (MathML) markup for that purpose. Anyway, I have little interest in seeing this bug fixed (I am tempted to mark WONTFIX :-) As roc jotted in the other bug 121748, any quirk that is implemented is likely to last forever, whereas generators can improve -- in which case, this bug will become a non-issue.

s/That's is all quirky... "<mi>k&prime;</mi>" is always understooned means/ /That's all quirky... "<mi>k&prime;</mi>" is always understood to mean/

Cool, so what should I use instead? <msup> <mi>k</mi> <mo>&prime;</mo> </msup> ...? <msup> <mi>k</mi> <mi>&prime;</mi> </msup> ...? I'm kinda confused as to which is appropriate... In this context, the &prime; is not an operator (which would imply differentiation), is it an identifier?

> <msup> <mi>k</mi> <mo>&prime;</mo> </msup> yes this is the recommended markup. (if you were to 'hint' at differentiation, you could do: <msup> <mo>k</mo> <mo>&prime;</mo> </msup>) > <msup> <mi>k</mi> <mi>&prime;</mi> </msup> In general, there isn't much that is done for '<mi>' (except the special-casing of non-stylable characters such as set R, or the italicized rendering in the case where the textual content consists of a single character). It is <mo> that is heavily overloaded (it is looked-up in the Operator Dictionary, and not only its content is significant, its position plays a role too -- whether it is prefix, infix, or suffix). So as a rule of thumb, '<mi>' is for things that are meant to be (italicized) identifiers while '<mo>' is for other things.

Alrighty. Thanks for the help!

My Firefox 1.0 now shows it as I wanted it to be. Is this a "meta" bug or has the policy changed?

I doubt if it's still working as Torsten wanted in Firefox 1.5.0.9. There is inconsistency, both in Linux and WinXP, in the presentation handling of &prime; and &Prime; (double prime). The latter works sanely as a postfix operator, while the former does not. With &prime;, however, there is an acceptable CSS workaround using {vertical-align: super; font-size: 0.6em} for the <mo> container that mostly -- but not always works. See the XHTML page http://www.albany.edu/~hammond/gellmu/primeaccents.xhtml and the pictures http://www.albany.edu/~hammond/gellmu/primeaccentsLinux.png and http://www.albany.edu/~hammond/gellmu/primeaccentsWinXP.png. I say the CSS snippet is an acceptable workaround because for most purposes &prime is used by mathematical authors as a superscripted postfix math accent (see Lamport's sample2e.tex). For GELLMU I am reluctant to step on an author's usage by superscripting it in the XML behind his back; touching it with CSS is much less of an infringement. On the other side, if I tell the author he must superscript it, he will walk away. I think the MathML spec at section 3.2.3.1 needs revision. For Mozilla meanwhile please make &prime; behave like &Prime;. Also I'm not happy with the response above to Ian Hixie; I think, in fact, <mi>k&prime;</mi> is a sensible math token. Otherwise one is messing with cdata rather arbitrarily. (The spec provides for this inside <mo> but mostly not inside <mi>.) For case distinctions attribute settings on the <mo> container should be the best way to go -- with postfix superscripting the default -- when the content is one of prime, Prime, tprime, qprime. Thanks.

In case the thrust of my comment a week ago is insufficiently clear, I have written a slightly more elaborate version of the first http reference in that comment. It may be found at: http://www.albany.edu/~hammond/gellmu/primeaccents2.xhtml

The MathJax fonts render the prime at the right size, so that may explain why some people see this bug "fixed". For the STIX fonts, font-feature-settings can be used to select an alternative glyph. Compare <math style="font-family: STIX;"> <msup><mi>x</mi><mo>&#x2032;</mo></msup> <msup><mi>x</mi><mo style="-moz-font-feature-settings: 'ss03'">&#x2032;</mo></msup> </math> (however, this will be font specific)