Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Posting-Version: version B 2.10.1 6/24/83; site sphinx.UChicago.UUCP Path: utzoo!watmath!clyde!burl!ulysses!mhuxr!mhuxn!ihnp4!gargoyle!sphinx!benn From: benn@sphinx.UChicago.UUCP (T Cox) Newsgroups: net.micro.mac,net.sci,net.wanted Subject: SUMMARY on What Scientists Want From Mac Software Message-ID: <1794@sphinx.UChicago.UUCP> Date: Tue, 11-Mar-86 22:48:17 EST Article-I.D.: sphinx.1794 Posted: Tue Mar 11 22:48:17 1986 Date-Received: Fri, 14-Mar-86 04:56:22 EST Distribution: net Organization: U of Chicago Computation Center Lines: 110 Xref: watmath net.micro.mac:5078 net.sci:578 net.wanted:8371 [] I originally promised to only post this on net.sci, but after several requests to the contrary, I'm cross posting to the groups where I asked for responses. Flame me if you will; it's too late now. :-) These are the replies I got to the question, "What do scientists want in a Macintosh [or any other] scientific stat/graphing package?" One hopes the developers out there are awake and listening. Thomas Cox ----- From: ihnp4!gargoyle!sphinx!fdot (Tom Lippincott) How about error handling? I've never seen *any* program where you could, for example, add 1.23+/-.02 to 4.56+/-.04 and get 5.79+/-.06, let alone perform more complicated math, graph them with error bars automatically, etc. Also, if statistics are incorporated, make it possible to flag errors as "systematic" or "spastic", as the systematic errors are not affected by averaging. What about including units in the math, and flagging errors of mismatched quantities (e.g., 5 meters + 3 joules) and working out the units of mathematical results (5kg * 10 m/s = 50 kg*m/s)? Built-in conversion tables? Settings for relativistic units (time=space, energy=mass, &c.)? ----- From: ihnp4!mhuxt!evans I would like to see 3d graphs with fit surfaces ----- From: ihnp4!mhuxl!smh >From the keys of Steve Henning, AT&T Bell Labs, Reading, PA mhuxl!smh On my wish list would be statistics/data handling capabilities and the ability to do qqplots, probability plots or whatever you want to call them. The statistics capability should handle missing values where appropriate. If you are familiar with S, it has most of this in it, but rather than having a spreadsheet, it has matrix math. If one could could marry S to the mac along with the tools that we know and love, it would be fantastic. ----- From: dubois@uwmacc.UUCP (Paul DuBois) Standard deviation bars! ----- From: ihnp4!tektronix!reed!bart (Bart Massey) The ideal package for the Mac would have about 80 windows -- a layout (like pagemaker) window, a text entry (like macwrite) window, a graphing window, a table window (like vax tbl), a equation typesetting window (like vax eqn), a graphics editing window (like macdraw), would be good starters. Basically, one would use the various subwindows to do various parts of a document, and then combine them all on the layout window. Clearly one could not emulate all the functions of each referenced program in its respective window, but one should be able to do a reasonable subset. The other alternative is to just do DAs that generate picts for graphing, tables, equations, and graphics, and then use a word-processor and the clipboard to tie them together. This might be gobs simpler... ----- From: ihnp4!bellcore!decvax!genrad!panda!talcott!cmcl2!rna!cubsvax!peters (Peter S. Shenkin) Build in all the features of APL! ----- From: Philip Leverton (Postgrad, Dept. of Physics, University of Melbourne) I already use,quite heavily, the S statistics and graphics software from AT&T. I rather like the multi-functional interface that allows you to do what you want by building up from simple tools but still has advanced routines available. - Rather like the so-called Unix philosophy. I doubt whether a "spreadsheet" interface has enough power to display and manipulate arbitrarily long arrays of numbers and strings - a function that S handles well. Its "implied vectorisation" - giving a vector to a scalar function applies the function to every element of the vector - is a very important attribute. A macro facility in which complex procedures could be constructed from the basic commands would be very useful too - how one could marry this with an essentially mouse oriented system would need some careful consideration. As you can see, my favourite package is S :-) ----- Well, that's it. Send your additions, comments, alpha test packages (:-) to: Thomas Cox ...ihnp4!gargoyle!sphinx!benn benn%sphinx@uchicago.bitnet We men speak of love and desire as though they were two seperate things, but they are not. They are two facets of the same thing. --