Notes for meeting at Hewlett Packard in Palo Alto on Thursday
	7/14/05 in the Sigma Room in Building 4.  Peter Markstein hosted
	us.  Jeff Kidder, Prof Kahan, David Hough, Alex Liu, Dick Delp,
	Debjit Das Sarma, Mike Cowlishaw, Ivan Godard, Jim Thomas, Joe
	Darcy, Alex Fit-Florea, Jon Okada, John Harrison, Peter Tang,
	David James, Liang Kai Wang, & Dan Zuras attended.  Leonard
	Tsai, Mark Erle, Steve Carlough, Chuck Stevens, & Eric Schwartz
	were on the phone.

	We began with expression.htm (1-2).  Expression evaluation,
	widento modes.

	Next: widento (3-2).  Widento modes.  While Jason's example does
	not actually cause the problem he fears, we acknowledged that,
	at least IN PRINCIPLE, there might exist similar codes that
	would behave mysteriously.  As the conversation drifted to what
	Cobol does, Prof Kahan pointed out that the people in Leon,
	France MIGHT be able to do float^integer correctly rounded with
	only a modest cost.  Cool.

	Next: optimization.htm (1-3).  Optimization modes.  Both Jim
	& Ivan objected to reassociation through parentheses but for
	slightly different reasons.  Jim, because the language has a
	specified order for associative expressions.  Ivan, because
	he felt that the overloading of ordering & binding with the
	characters '()' meant that the compiler has insufficient
	information to know what is actually desired.  For all of what
	Chuck outlined, he had to admit that there is only one
	implementation that only partially implements 'standard
	arithmetic'.  Perhaps that will improved in the near future.

	We took a break around 2:45.

	Peter began his presentation of BID at 3:00.  His initial
	investigation suggested there was no need for Intel to devote
	hardware to decimal arithmetic.  Therefore, he argues that most
	implementations will be in software for some time to come.  He
	concludes that DPD "might not be the most friendly encoding" for
	software implementations.  Thus, his BID (Binary Integer
	Decimal).

	Mike began his rebuttal at 3:45.

	I started around 4:30.

	I finished around 4:45.  At which point, a very lively
	discussion ensued.

	We adjourned at 5:35.