given the perspective of your OER project, I understand some of the confusion my considerations might cause. Therefore I try to discribe, how I came to my approach:
1) First there was MIT's OCW (OpenCourseWare) initiative. As an interested observer, I understand that MIT provides an editorial concept and support stuff for one goal: to make self produced course descriptions and learning materials available for free. In principle, these materials are produced by individual faculty for use in their own residential classroom teaching. It is learning material that has to be produced anyway. Publishing just adds value, since it raises visibility of the teacher and the institution, and the material might be of use for others (in not very specified ways). Just making things available, which are already there, with the least possible additional burden for the individual teacher.
MIT's OCW just gives away materials, but not credits, interaction, teaching, ect. This second part, the educational activities and services, are still separat.
2) In principle, I want to follow the idea to keep OpenCourseWare seperated from educational services. OCW should be freely accessible, while educational services should be restricted to enrolled students.
My approach only differs with respect to the production process: As far as I understand, MIT has a large group of support staff that collects materials from faculty to edit and publish it, after the course is over. Therefore, I understand MIT's OCW as a post-teaching publication. Materials have to be transfered from any learning environment to the publishing environment of MIT's OCW.
In my opinion, this approach of developing and sharing learning materials is only viable for MIT or other institutions that can afford large support units. But many European institutions do not have these facilities and therefore have to rely on the work and the commitment of individual faculty members. My hypothesis is that they only will buy into OCW, if they can produce and publish materials on their own, while they are teaching the course. (After the term is over, they would not invest additional time for post-teaching-publication. The only chance I see is to ask faculty to publish, while they are teaching their course.)
I also want to make materials freely available and to restrict acces to educational services. But I want to link both spheres technically/logistically, so that the individual teacher can control and shift between both spheres without much additional help, making it easier to contribute to OCW without much additional work. The technical solution can/should be Moodle, by creating a public (freely accessible) and a private (restricted, for enrolled students only) area for the same course. (Thanks for your recommondations, I think, this is the way it could work.)
3) "The OER project is following the approach of building resources that will then be used in different course settings - so they'll be highly modular for ease of customisation."
As far as I understand your project, it significantly differs from MIT's approach to OpenCourseWare. I think, you are more concerned with the planning and the designing for possible re-uses of materials than MIT does. One of the charming things about MIT's OCW approach is that it needs less coordination with possible users, and that it strongly supports the authorship of the individual lecturer. (I see an analogy to publishing a journal article: If I want to publish, I have to meet editorial standards and pass a peer review, but I do not have to pre-determine, which part of my text will have to be used by whom in what specific way.)
If your project primarely produces for a national context, I assume that is possible to influence (or: know about) some of the framework conditions for re-use, like national curricula, etc. I am really interested to know, if you have any mechanisms to influence re-use of materials, e.g. by setting up joint developer groups, creating obligations/committment for use of your materials, etc. There exist similar examples of national efforts for joint content production (e.g. Finnish Virtual University), which might be interesting for you.
Summing up, I think, Open Educational Ressources are an important issue for eLearning in general, and for the Moodle community specifically. MIT's OCW approach is one approach, the New Zealand Open Educational Ressources project is another. Personally, I think it is very interesting to compare the similarities and differences of these approaches, since they represent different ways to achieve the same goal: to share educational ressources.