I just finished a paper on 'interpretation' of ORM models (e.g. to/from KL or 'other' formal theories). If it's accepted (for this years conference), it'll be one place to start (beyond Brian's points, which I believe are spot-on and right to the point).
One heads up on 'interpretation' is this. 'Interpretation' of any formal theory/structured information is up to you (within logical limits). e.g. If you could look at the binary data of a 'Microsoft Word' document (i.e. computer based 'file' with a '.doc' extention) and consistently interpret verses of the Koran from that binary stream of data....then not only would that be remarkable, but it would be a valid 'interpretation' of a '.doc' file (binary stream of data). The fact that that is probably very unlikely to happen, is probably a good thing for Microsoft, the Koran and our belief system in general (i.e. it confirms what we believe to be true, in that '.doc' files are best interpreted as 'documents' by software such as 'MS Word' and can be printed out on paper and read). It is only incidental that some of those documents may be verses of the Koran.
Does that answer your question? Specifically, what Brian said is true...nORMa interprets the ORM meta schema as 'objects' and displays them as an ORM diagram (interestingly....those 'objects' have 'methods'...but they are predefined, are not in the ORM model, and are restricted to ORM diagramming in nORMa). In the 'relational view' of the ORM diagram (in nORMa), you start to see what looks like 'objects'.
Attributes of Objects
To interpret an ORM meta-schema as 'Objects with Attributes' (like ODL)...you'd need to go through the 'relational transfomation' steps (as well documented in Terry's books and/or doctoral thesis), or as supplied in the open source of the nORMa tool. Sounds like you have read Terry's thesis...but you can find it at www.orm.net just the same. If you are lucky...all the source code is supplied for you within nORMa.
Methods on Objects
As you most likely know, Object Role Modeling (ORM) does not support 'methods' or 'functions' (as at ORMv2.0) on Entity Types/Nested Fact Types.
Tony Morgan is researching the depiction of 'functions/processes' in ORM diagrams...and an abstract of his paper can be found in the Library here: http://www.ormfoundation.org/files/folders/orm_2007/entry81.aspx ("Business Process Modelling and ORM").
Or, there are ways to interpret Fact Type Readings as 'function headers'..by way of isomorphic interpretation of ORM models/Fact Type Readings....but that may be beyond your immediate needs.
If you only want the attributes of your objects...then certainly, the relational transformation steps documented by Terry Halpin (and works by Nijssen and Leung) will be good enough.
I hope this is helpful information.
P.S. I have read, on other sites/forums, that there is contention as to the fact that the ORM methodology has little to do with modelling 'Objects' (in the OO sense of the word, with the inclusion of 'methods'). On the whole, I think that is a fair criticism of the 'name', but not the methodology. At any rate, if you have chosen ORM, then I don't think that is going to worry you.