ksr1492@cec1.wustl.edu (Kevin Scott Ruland) (03/14/90)
Someone posted hp's bug list earlier. On it was a bug that incorrectl translates symbolic complex numbers in polar form to rectangular form. This bug was the only one present in the four current rom revisions. How does hp suggest "getting around" this bug? And, when do they expect to fix it? This bug seems at least as serious as the 8x8 inversion bug and is preventing me from buying a 48sx today. Kevin Ruland
ccplumb@lion.waterloo.edu (Colin Plumb) (03/14/90)
In article <1990Mar13.174255.2517@cec1.wustl.edu> ksr1492@cec1.wustl.edu (Kevin Scott Ruland) writes: > Someone posted hp's bug list earlier. On it was a bug that incorrectly >translates symbolic complex numbers in polar form to rectangular form. >This bug was the only one present in the four current rom revisions. >How does hp suggest "getting around" this bug? And, when do they expect >to fix it? The bug arises when entering a complex number in polar form, and then only when you're using symbolic (lacking a value) variables. Where you want r cis theta, it actually returns theta cis r. The solution is to transpose the arguments to the function. I wouldn't call it a fatal bug. > This bug seems at least as serious as the 8x8 inversion bug and is >preventing me from buying a 48sx today. The 8x8 inversion bug isn't very serious, either... if you need to invert larger matrices, divide by an identity matrix. You might lose a bit of accuracy, but you can define your own invert function which calls INV on smaller matrices. -- -Colin
dan@Apple.COM (Dan Allen) (03/15/90)
In article <1990Mar13.174255.2517@cec1.wustl.edu> ksr1492@cec1.wustl.edu (Kevin Scott Ruland) writes: > > Someone posted hp's bug list earlier. On it was a bug that incorrectl >translates symbolic complex numbers in polar form to rectangular form. >This bug was the only one present in the four current rom revisions. >How does hp suggest "getting around" this bug? And, when do they expect >to fix it? > This bug seems at least as serious as the 8x8 inversion bug and is >preventing me from buying a 48sx today. It is really not a big deal at all. It is only when entering a symbolic complex number in polar mode that the bug occurs: the numbers are incorrectly converted to rectangular coordinates during command line processing. All you have to do is enter the complex arguments in reverse order if in polar mode and then everything is fine. Everyone should remember that internally all numbers are kept in rectangular coordinates: only the display shows things in polar mode. It is during this input conversion process that the bug occurs. Once in the machine, all is fine. Dan Allen Apple Computer