Just wanted to advertise that Harold has posted a draft specification of how an RF2 to OWL transformation should work, described in English with SQL rather than using the Spackman PERL script as the gold standard. See Representation of SNOMED in OWL.v0.9
A number of issues came to light during this work and I'd like to invite you to consider these and modify the issues page as you see fit: Transformation Issues
It seemed more efficient to keep them all on one page and add text directly, but if people prefer (or if the text becomes uncomfortably large) I could split each issue to its own page and suggest using the comment functionality rather than editing the page. As you wish.