home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!portal!lll-winken!fnnews.fnal.gov!mp.cs.niu.edu!linac!convex!news.utdallas.edu!aicklen
- From: aicklen@utdallas.edu (Greg Aicklen)
- Newsgroups: sci.math.symbolic
- Subject: Inconsistent results across Mma platforms
- Message-ID: <BxoE87.DCz@utdallas.edu>
- Date: 13 Nov 92 22:24:07 GMT
- Sender: aicklen@utdallas.edu (Gregory H. Aicklen)
- Organization: Univ. of Texas at Dallas
- Lines: 23
- Nntp-Posting-Host: drips.utdallas.edu
-
- A short time ago I posted a message about the inconsistent symbolic results I
- was getting across platforms, even with the same Mathematica version.
-
- I contacted WRI, and sent examples to them, but I got no response. So I
- did a little poking around and I discoverd what the problem was. If you
- have had problems, or if you have recently upgraded, you might want to
- read this.
-
- I upgraded from Mma 2.0 for Windows (on PC) to Mma 2.1 for Windows. I used
- the standard installation program and followed the directions. But it
- appears that not all of the files were upgraded. In particular, all of the
- files in the "packages/startup" directory were left alone, and the
- "packages/preload" directory was not created (preload should have replaced
- the startup directory in the new version). So, I was preloading the OLD
- versions of such important features as ComplexExpand[], etc.
-
- By deleting everything and starting over, I got the right results for Mma 2.1.
-
- Seems WRI could spend a little more time making their installation programs
- robust.
-
- Greg Aicklen
- aicklen@utdallas.edu
-