As MicroXML[1] is pretty much a work in progress but has reached at least some form that is useable and very well translates to JSON, we should keep an eye on it's evolution and evaluate for version 5.0 of OData, if introducing the format MicroXML as first class format (parallel to JSON and probably remove ATOM or remap the latter to MicroXML[2,3]) might be beneficial for formulating the ideas of OData and implementing/using it the like.

References:
[1]: "MicroXML", https://dvcs.w3.org/hg/microxml/raw-file/tip/spec/microxml.html
[2]: "Introducing MicroXML, Part 1: Explore the basic principles of MicroXML", http://www.ibm.com/developerworks/library/x-microxml1/
[3]: "Introducing MicroXML, Part 2: Process MicroXML with microxml-js", http://www.ibm.com/developerworks/library/x-microxml2/