It seems to support a lot more than traditional MIDI, especially in terms of control and synchronisation.
Does anyone have any idea how well MIDI 2.0 and OSC will compete with or complement each other?
By contrast, MIDI (both MIDI 1 and 2) are flexible application level protocols. For example, among other things, MIDI 1.0 describe schemas for musical notes, parameters, transport control, and time synchronization. Built-in application schemas allow devices that fit the application model to communicate in a relatively seamless way. I believe that MIDI 2.0 provides a more extensive schema, that includes (for example) device discovery and capability queries, and removes some limitations of the old schema. I'm not familiar enough with the details of the final MIDI 2.0 spec to say much more than that.
As I recall, some of the features of MIDI 2.0 (e.g. capability queries, discussed elsewhere on this page) were proposed for an "OSC 2.0", however the fine people at CNMAT who produced the OSC 1.0 spec didn't have the resources to sponsor 2.0 development, and no one else stepped up. In contrast, the MMA (MIDI Manufacturers Association), who sponsored the 2.0 spec, have all of the major music corporations as members (e.g Roland, Korg, Yamaha). That said, as I understand, the MIDI 2.0 process was open to anyone, and I know of at least one independent developer who was involved.
Will they compete? I suspect that the situation will continue much unchanged: commercial hardware will support MIDI (1 and/or 2), and as is currently the case, few commercial music devices will support OSC. OSC will likely continue to be the protocol of choice for custom projects using custom hardware, software and application schemas. Perhaps with time, as the tools improve and we get API support for MIDI 2.0 in operating systems and embedded libraries, it might become easy enough to develop MIDI 2.0 software to choose between OSC and MIDI 2.0.
Edit: clarity.