Well, we're not going to force folks to provide Oracle ports of new packages unless they want to (nor PG ports of new Oracle packages etc) so the sequence isn't necessarily going to be right to do this.
But ... you should really e-mail Steve Woodcock directly, he should be back from holiday now.
Dan's got some hack he used to pull apart PL/SQL packages - that's where all those nice PG functions with default values given as comments come from, and is why so many packages still have the Oracle original embedded as comments. Steve's original thought was to parse the .sql files when doing package instantiation, but we decided that was too fragile.
But ... ripping up the PL/pgSQL functions to generate the inserts into metadata tables would be OK. They could be hand-fixed if necessary.
Of course, at some point we have to ask "which is more work, dealing with positional parameters or doing the work to support a nicer paradigm?"
