Adrien, there is an ongoing project to define a successor to MusicXML, and Daniel is one of the key members of the team.
The problem with any standard (like MusicXML) is that to get any traction it has to cater for the least common denominator of all the interested parties. So it tends to end up fossilizing what is already more of less obsolete. (I know first hand - I’ve been on such projects, not in the music field).
MusicXML is too closely tied to the way music looks, on the printed page not to what the notation means. But if you break that connection to violently in a new standard, the Sibelius / Finale / Musescore generation of notation software wouldn’t be able to do anything with the standard anyway. (If you are a Sibelius user, think what Sibelius could do, importing music described in terms of flows, layouts, and players, for example - not much I think, without a complete rewrite. Not to mention laying out the page of a score with frames.