We had a meeting of the Mercury group on Fri March 30, 2001,
from 4.15pm to 5.15pm.

Attendees:
	Zoltan Somogyi (zs)
	Fergus Henderson (trd)
	Adrian Pellas-Rice (apell)
	Simon Taylor (stayl)
	Tyson Dowd (trd)
	Mark Brown (dougl)
	David Jeffery (dgj)
	Kevin Glynn (keving)

Report from Microsoft .NET DevLab

	Zoltan, Tyson and Fergus reported on their recent visit for Redmond
	for the .NET DevLab.

	At the DevLab they and the other language developers gave feedback on
	the suitability the .NET platform as a back end for other 
	'non-standard' languages.

	They also spent quite some time negotiating the various software
	licences that we are affected by.

	Agreements were signed that should allow us to get the latest
	developer's version of Visual Studio (presumably with a 60-day expiry)
	quite soon. During the visit they were able to get a rudimentary 
	version of syntax highlighting working for Mercury through Visual 
	Studio.

Progress reports

	We went around the table, with each person in turn discussing what
	they had been working on recently and/or what they were planning
	to work on.

	Tyson 'Jetsetter' Dowd (trd):
		- Will be leaving for Europe for six weeks on Saturday week.
		  He will be spending 4 weeks at MSR in Cambridge and 2 weeks
		  at Mission Critical in Belgium. At Mission Critical he
		  plans to work on support for types defined in foreign
		  languages.

		  Aside from organising his travel plans he has been working
		  on allowing C# to work through Mercury's foreign language
		  interface. It is mostly working, although it does not yet
		  automatically select between C# and Managed C++ versions
		  of the same procedure.

	Mark Brown (dougl)
		- Has spent most of his time reading the .NET documentation,
		  including the ECMA specs and the Microsoft SDK 
		  documentation.

		  He has also reviewed Simon Wei's quickcheck implementation,
		  and will tell him to commit soon. Other people should also
		  read the www based quickcheck tutorial.
	
	David Jeffery (dgj)
		- Has been working on optimising the CPLEX based solver.
		
		  The HAL implementation now uses the main branch Mercury
		  compiler, and this is working fine. The merge of the
		  HAL and main branches has caused two Mercury bugs which
		  affect everyone and need to be fixed RSN.

		  He has also been working on getting more of the HAL library
		  (some of which only works under the SICStus backend) to
		  work with the Mercury back end. This has shown up a bug
		  (or maybe a number of bugs) in the way Mercury handles the
		  'any' inst. Fergus, David and Maria have been communicating
		  about this. The fix for this is not simple. Maria should be
		  at Melbourne Uni. next week, so Fergus, David, Maria and
		  David Overton can sit down and work out what to do then.

	Zoltan Somogyi (zs)
		- Working on fixing bugs in deep profiling. The latest bug
		  fixes are in the process of bootchecking.

		  He is setting up a way of automatically testing the output
		  generated by the deep profiler, to make sure that at the
		  very least we continue to be able to generate the HTML pages
		  that display the profile.

		  He will start on performance testing the system soon. He also
		  has a scheme for deep profiling code that uses exceptions,
		  but it will only work for the low level C backend at this
		  stage.

	Fergus Henderson (fjh)
		- Has re-arranged the material on the Mercury www site.

		  Apart from that, he has spent most of his time recently
		  working on various licensing and copyright issues that
		  have cropped up with respect to the Visual Studio integration
		  and the new GCC back end. There is lot of red tape to
		  work through!

		  There was also some discussion of whether the GCC back end
		  should be included in some versions of our binary
		  distributions; it will make the distributions larger, but 
		  will make installation easier. There was also discussion of
		  building some very cut down versions of our binary
		  distributions; a user on comp.lang.functional commented that
		  our distributions are prohibitively large.

	Adrian Pellas-Rice (apell)
		- Has been debugging his sequence quantification changes. He
		  doesn't have a lot of time to work on this any more, so is
		  handing his changes over to Peter Schachte to work on.
		  
		  Fergus suggested that he post his current diff to
		  mercury-reviews at the same time.

	Simon Taylor (stayl)
		- Has been working on a system to keep track of compilation
		  dependencies in a fairly fine-grained way. 

Minutes taken by David Jeffery (dgj@cs.mu.oz.au).