Notes for meeting at HP in Palo Alto on Friday 2/17/05 in the
	Sigma Conference Room in Building 1.  Peter Markstein hosted
	us.  Dick Delp, David Hough, Jon Okada, Jim Thomas, Ivan Godard,
	Prof Kahan, & Dan Zuras attended.  Leonard Tsai, Alex Liu, Eric
	Schwartz, Jeff Kidder, Mike Cowlishaw were on the phone.


	We started with underflow.  It is short.  It is almost word for
	word what was voted on last month.  And, of course, since this
	was agreed to without objection last month, a lengthy discussion
	ensued.  The vote was 8 for, 2 against, & 2 abstentions.  So the
	proposal was accepted.

	Mandatory conversions.  After a fairly short discussion (which
	was nonetheless longer than the topic deserved) it was sent back
	to committee.  It looks like we cannot separate this issue from
	the rest of the expression evaluation issues.

	ISSCC: http://www.blachford.info/computer/Cells/Cell1.html
			The CELL Microprocessor

	We took the 3:00 break at 2:30.

	After the break, we took up the topic of the naming convention
	for datatypes that Prof Kahan wanted to discuss.  He wishes to
	replace binary32 with "4-byte binary' & decimal64 with "8-byte
	decimal" for the basic formats only.  Or, more specifically,
	"4-byte binary(127x24)" & "8-byte decimal(384x16)".  And, for
	non-basic formats (extended or optional) "binary(>16000x>63)"
	or "decimal(>13.3)" where "13.3" means a 16 digit fixed point
	with 3 digits behind the decimal point.  A lengthy &, I thought,
	unusually relevant discussion ensued.  We decided to leave it
	as is & Ivan offered to come up with a notation that meets with
	everyone's needs.

	We discussed ARITH-17 briefly & finished up somewhat after 5.