We had a meeting of the Mercury group on Monday, 19 November 2001. Attendees: Zoltan Somogyi (zs) Tyson Dowd (trd) Ralph Becket (rafe) Mark Brown (dougl) David Overton (dmo) Fergus Henderson (fjh) Progress reports Zoltan Somogyi (zs): - Sent in the stack-slot optimization paper on time. - Benchmark results show a 1 - 1.7% speedup and a code size reduction of a few tenths of a per cent. Static analysis shows stack frame size reductions in 20% of cases. - Has been marking exam scripts. Tyson Dowd (trd): - Finishing implementation work on the foreign type and related syntax. - Has written a detailed dissertation outline. Ralph Becket (rafe): - Has continued working on the loop invariant hoisting optimization. This work is nearly ready for testing. Mark Brown (dougl): - Finished off support for .NET data types in the retry aspect of the debugger; only managed pointers are not unhandled and Mark has a design to accommodate them. David Overton (dmo): - Has done some more work on the mode inference paper. - Started work on a detailed dissertation outline. Fergus Henderson (fjh): - Gave Ralph some help on the loop invariant optimization. - Has been working on accurage GC for the high-level C back end. Has done some preliminary benchmarking to get an idea of the cost of the extra bookkeeping involved. Other Business Five students will definitely be working with us over the summer. Their exam's finish on 27 November and the plan is to give them the Mercury introductory courses on the 29th and 30th. One student will be working on the accurage garbage collector. Various students have expressed an interest in working on the Java back end. Simon Mei (not one of the five definites) who was a summer student last year has explained that he may be able to work with us part time. Minutes taken by Ralph Becket (rafe@cs.mu.oz.au)