On 5/1/07, skaller <[EMAIL PROTECTED]> wrote: > In principle I think 'complex' and 'complex with infinity' > are different types, but then complex+inf only has ONE > infinity, quite different from 'number too large' which > might retain a phase angle.
Infinity+phase angle is different than "number too large"+phase... infinity+phase angle is needed so identities such as angle(K*x) === angle(x) to hold when the scalar K "equals" infinity. (I use "equals" here in the IEEE-754 sense.) But certainly I agree that a complex type where infinities are disallowed (and 1/0 raises an exception) would be useful. > Felix doesn't box products :) Well, you need at least one bit to store the polar/Cartesian flag, which means an extra word in the representation, which means a structure of odd size... but in polar form we only need a subset of the range available in an IEEE double (only positive magnitudes, -pi:pi angle) so maybe we can do some trickery there to shove everything into two doubles... - Chris ------------------------------------------------------------------------- This SF.net email is sponsored by DB2 Express Download DB2 Express C - the FREE version of DB2 express and take control of your XML. No limits. Just data. Click to get it now. http://sourceforge.net/powerbar/db2/ _______________________________________________ Felix-language mailing list Felix-language@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/felix-language
