deal with multirelation@latexml

[kohlhase]

If LaTeXML sees something like $A=B=C$, then it creates

<om:OMOBJ>
  <om:OMA>
    <om:OMS cd="latexml" name="multirelation"/>
    A
    <om:OMS cd="relation1" name="eq"/>
    B
    <om:OMS cd="relation1" name="eq"/>
  </om:OMA>
</om:OMOBJ>

this is reasonable, but the OMDoc we generate does not support latexml as a theory (importing it where necessary). so in particular, it is not taken into account in the rendering among others.

[kohlhase]

I am sure that there are similare cases like these that we cannot do at the moment.

M

[deyan]

We could:

• Edit our branch of LaTeXML to produce something else
• Edit the XSLT to substitute these occurrences for some reasonable alternative.
• Handle this in JOMDoc

These "multirelation"s are firstly introduced by the LaTeXML grammar as XMath and then converted to what we see by the OpenMath? postprocessor. So there is no direct fix to this in the bindings, that I am aware of, at least not a general one.

[kohlhase]

you are right in pointing out that we can fix this in the XSLT. So here is what I am going to do; make a theory/OMDocCD multirelation, that works exactly like the one latexml supplies and rewrite it in the XSLT. I also have to see that I import that theory whenever necessary. This should be simple to do, and should also solve all similar cases.

I had forgotten that possibility (again; I fear that you will have to keep reminding me).

[kohlhase]

This will only really work when we have solved #1477

[kohlhase]

and of course we will also need #1472, since I do not see that I can build an sTeX notation definition for multirelation.

[kohlhase]

I have taken care of this, seems to work.