Répondre à : MathMl et commandes maths wims

Accueil Forums Gestion d’un serveur WIMS Administration d’un serveur WIMS MathMl et commandes maths wims Répondre à : MathMl et commandes maths wims

#2286
eric
Participant
Up
0
Down
::

Thank you Joke,
but I think this fix for \cal in mathml doesn’t solve completely the problem.
Unless we want to abandon totally the image mode, the same formula written should give the same result in both mathml and image mode, provided they respect some (maybe restricted, ideally not) wims syntax.

I don’t know what mathml syntax is, but latex syntax is the following :
\cal is a font change, hence valid for all characters after it inside its group.
\mathcal is a function applied to one argument, which can be one character or a group of characters embraced inside brackets.

This is not what directexec gives on Leiden server for mathml :
\cal A gives correctly a calligraphic A in both modes
\cal AB gives a calligraphic AB in image, but in mathml the B is back to roman
\cal {AB} gives a calligraphic AB in image, but not understood in mathml
\mathcal A gives strange result in directexec (the letter A appears 3 times, possibly a bug of directexec). Probably the image mode is correct, not mathml.
\mathcal AB gives a calligraphic A and roman B in image mode, and only roman for mathml (+ same « bug » as above).

The other point cannot be tested (at least I don’t know) inside directexec : it is the fact that some « error » in mathml seems to affect later formulas. If you try
\statement{matrice \([1,0;3,-4]), base \(\mathcal B) , matrice \([1,0;3,-4])}
in createxo under mathml mode, the second matrix is presented as an image, while the first is written in mathml, probably because the \(\mathcal B) makes mathml sick.

So in my opinion we should first decide which syntax is allowed for calligraphic letters in wims (cal and/or mathcal, brackets or not, acts as font or funtion, …) and then try to make the parser render the formula in both modes as described by the proposed syntax. Since I am used to latex, I would of course prefer taking latex syntax as reference and try to adapt the wims parser for mathml, but you may have good reasons to object to this choice.

Once the choice is done and the technical difficulties overcome, some pedagogical modules may have to be changed (I encountered the problem in the exercise module=U2/geometry/oefgeomeucl.fr&+exo=rotationanti&+cmd=new, which is now ok on your Leiden server).

Eric