On Fri, Mar 12, 2010 at 08:33:03AM -0800, Geoffrey Broadwell wrote: > On Fri, 2010-03-12 at 12:04 +0000, Allison Randal wrote: > > Any final discussion on docs/pdds/draft/pdd31_hll.pod before I launch it > > out of DRAFT? Does its current form accurately reflect the current > > implementation work on HLL interop? > > Tene and I discussed this at some length in private; our conclusion was > that the API represented in this document is incomplete for some > necessary tasks, had some specification fuzziness, and generally needed > one or two working implementations to sanity check it.
+1 to this. HLL interop is one of those areas where I think we need to go through a process of (1) estimate what might work, (2) let some HLL implementations use the estimate as a guide (as opposed to specification) to figure out what really does work, and then (3) revise/rebuild the PDD to document what actually happens. So, in that sense, the current pdd31_hll.pod really represents #1 above, and as Geoffrey says, get some implementations to sanity check it and flesh out some of the details. Whether this means the current pdd31 can move out of draft is probably up to the architect. My suggestion would be go ahead and move it out of draft, but add a note somewhere that the PDD is expected to be revised and improved in response to HLL implementation experiences (unless this is assumed for all PDDs already, which hasn't been my impression thus far). Pm _______________________________________________ http://lists.parrot.org/mailman/listinfo/parrot-dev
