Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!wuarchive!sdd.hp.com!elroy.jpl.nasa.gov!decwrl!pa.dec.com!decuac!shlump.nac.dec.com!jareth.enet.dec.com!edp
From: edp@jareth.enet.dec.com (Eric Postpischil (Always mount a scratch monkey.))
Newsgroups: comp.sys.handhelds
Subject: Re: A Bug or not ?
Message-ID: <22201@shlump.nac.dec.com>
Date: 19 Apr 91 17:24:10 GMT
References: <00946F29.17B42000@MAPLE.CIRCA.UFL.EDU> <1991Apr12.111705.4699@ifi.uio.no> <1620@fs1.ee.ubc.ca>
Sender: newsdaemon@shlump.nac.dec.com
Reply-To: edp@jareth.enet.dec.com (Eric Postpischil (Always mount a scratch monkey.))
Organization: Digital Equipment Corporation
Lines: 29

In article <1620@fs1.ee.ubc.ca>, jmorriso@ee.ubc.ca (John Paul Morrison)
writes:

>However there _is_ a temperature bug, even though I believe it is
>performing as documented:
>
>Try raising a temperature to a convenient integer power:
>
>ie 5_C^4. You get 625!!! That is _not_ right! HP does not even follow
>its 
>own rules.

I do not see how you can say both "it is performing as documented" and
"HP does not even follow its own rules".  Page 201 of the Owner's Manual
says pure temperature units are converted to absolute temperatures
before adding or subtracting and the sum or difference is then converted
to the level 1 unit.  It doesn't say any such thing is done for
exponentiation.

The "rule" is that temperature units on the 48 can mean either relative
or absolute units, and the 48 tries to figure out which from context. 
For adding and subtracting pure temperature units, it assumes the
meaning is absolute.  For other operations or other forms of the units,
it assumes the meaning is relative.


				-- edp (Eric Postpischil)
				"Always mount a scratch monkey."
				edp@jareth.enet.dec.com