From: SMTP%"RELAY-INFO-VAX@CRVAX.SRI.COM" 12-SEP-1993 16:48:05.01 To: EVERHART CC: Subj: AnalyRim vs. Lotus 1-2-3 on VMS Date: Fri, 10 Sep 1993 21:09:42 GMT From: "Glenn C. Everhart" To: INFOVAX@arisia.gce.com Cc: EVERHART@arisia.gce.com Message-Id: <930910210942.a2@arisia.gce.com> Subject: AnalyRim vs. Lotus 1-2-3 on VMS I got a demo copy of Lotus 1-2-3 to check out to see if anyone in my shop was interested. Put it up on my VS3100/76 to test. My test consisted of setting up a spreadsheet in which the range B2..H57 cells all contained the formula @avg(a1..c3) (for B2) or the equivalent (in other words, each cell is the average of itself and its 8 neighbors) and the edges were all set to fixed values. I set the thing to row-wise calculation and told it to do the calculation in the foreground when I demanded it, so I could get valid timing for it. Turned out that it recalculated in about 1.25 seconds, averaged over several tries (timed with a stopwatch), nothing else happening on the Vax. For laughs, I tried the same thing on AnalyRim. AnalyRim's display is much slower, but the calculation speed was better than Lotus: about 1.04 seconds for the same spreadsheet (with the formula set to AVG[A1:C1,A2:C2,A3:C3] in the central region cells). That surprised me. Stil, there it is: a freely available spreadsheet ran faster than the best selling one. If you use the VM command to turn off AnalyRim's display during recalcs, the slower display is less of an issue. Also, AnalyRim has hooks designed to let it call Gnuplot for its plotting. GnuPlot is DEFINITELY a non-shabby product!! Mind, in cases where "natural order" recalculation is possible, I suspect 1-2-3 will have a major advantage. However, AnalyRim has addressing modes that let it do some really oddball things that I suspect the commercial package would have trouble duplicating; pointer arithmetic is important in programming it, and well supported. Glenn Everhart Everhart@Arisia.gce.com