Hi Max, On Tue, May 18, 2021 at 08:02:12AM +0200, Massimiliano Gubinelli wrote: > > The question is whether TeXmacs would really be in its role here. > > This could be considered to be the job of plug-ins. > > It could. But then the question become if we want to provide a unified > mechanism to parse mathematical formulas, e.g. from the TeXmacs format to an > abstract form which could then used to create expression in some target > language (scheme, mathemagix, Python, etc...) for evaluation.
We already have such a mechanism. It is used by each plug-in that supports mathematical input. > Thanks, I will give a look to the parsing support. I looked already in the > C++ code but I do not understand how to extract a syntax tree from the > parsing execution, it seems that it returns only whether the formula can be > parsed. We would need to add productions to the parser and our grammars. That is what I planned to do one day, when I would have time. > > Yet another issue concerns the semantics of symbols/formulas. > > In different areas, the semantics may differ, > > so not everybody will necessarily agree with our decisions. > > > > I think the common need for plugins is to take a TeXmacs document which > represent a mathematical expression and extract some AST e.g. a scheme tree, > then a plugin can process this tree to target the final language. No, this is not what users of plug-ins expect. They expect to write programs using the syntax of the plug-in. They do not want TeXmacs to mess up the syntax. For instance, Mathematica uses [] instead of (), and Mathematica users want to use the first notation. Best wishes, --Joris _______________________________________________ Texmacs-dev mailing list Texmacs-dev@gnu.org https://lists.gnu.org/mailman/listinfo/texmacs-dev