home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: sci.math.symbolic
- Path: sparky!uunet!stanford.edu!EE.Stanford.EDU!siegman
- From: siegman@EE.Stanford.EDU (Anthony E. Siegman)
- Subject: Re: Mma and linear algebra
- Message-ID: <1992Jul22.025710.17724@EE.Stanford.EDU>
- Organization: Stanford University
- References: <SMITHW.92Jul21184319@sofya.hamblin.math.byu.edu>
- Date: Wed, 22 Jul 92 02:57:10 GMT
- Lines: 28
-
- From Dr. William V. Smith):
-
- >Someone asked about Mma and linear algebra, in particular, finding the
- >eigenvalues and eigenvectors of the matrix
- >
- >{{0,1,0},{0,0,1},{1,1,0}}
- >
- >I tried "Eigensystem" on this on my home NeXT, and it did fail.
- >Mathematica has problems with complex numbers. I also just
- >tried it on a system at work (a MIPS) and it gave the algorithm
- >failed message.
-
- Try {{0.,1.,0.},{0.,0.,1.},{1,.1.,0.}}. On a Mac SE/30 with mma 2.0
- both the eigenvalues and eigenvectors come out almost instantaneously.
- Without the dots the eigenvalues come out in a few seconds, and I gave
- up on the eigenvectors after a minute or so.
-
- One thing I've learned from several helpful msgs from David Withoff at
- wri is that mma behaves _very_ differently in many situations
- depending upon whether you write integers (including especially zero)
- with or without a decimal point. Even something like
-
- m = 1; rootTwo = Sqrt[2.]; f[x_] := Exp[ - m rootTwo x^2]
-
- can behave differently for "m = 1" or "m = 1." despite the fact that
- you'd think the product "m rootTwo' would lose the pure number
- identification.
-
-