Re: Re: Mathematica skill level snippet(s)

*To*: mathgroup at smc.vnet.net*Subject*: [mg104942] Re: [mg104852] Re: Mathematica skill level snippet(s)*From*: "Vincent N. Virgilio" <virgilio at ieee.org>*Date*: Sat, 14 Nov 2009 01:55:57 -0500 (EST)*References*: <200911917250.789293@jfultz2winlap> <hdbhn2$jol$1@smc.vnet.net>

Bobby, I'd be crazy only if I took myself seriously. I hope there's not a conflated design decision lurking behind Mathematica's ubiquitous Automatic. There's a bright line between automatic algorithmics/aesthetics and Automatic options. Sure, the two overlap, so the line has some width. PlotStyle in Plot probably doesn't always have the same Automatic value. Still, this is about engineering good software, not perfecting a mathematical edifice [* soapbox], so WRI can approximate. They can and should specify usual, common, or frequent values for Automatic options and similar elements. There are examples which demonstrate that they already understand this; viz, the documentation of MeshFunctions clearly lists default values across host functions. Vince Virgilio [* soapbox] The mantra "everything is an expression" in Mathematica is about as helpful as "everything is an atom" is in the physical universe, or "everything is an utterance" is in linguistics. It's a beautiful observation that more can enjoy than benefit from in economic time. WRI should not overuse it. Certainly not to present Mathematica as so plastic that it defies documentation. Perhaps it's time to start making some "hard decisions" and evolve the Mathematica language from compiler-level abstract syntax tree mechanisms into something more human, with more convenient system-supported idioms; now there are at most a few. More constraints, perhaps, but likely also more efficiencies. As it stands, Mathematica is a strange mix of ultra high-level semantic terminals (NDSolve) and their low-level grammatical glue. WRI probably needed this scaffolding to get so far, but eventually scaffolding is taken down. This might be a timely suggestion, since Stephen Wolfram's personal to-do list for Mathematica will probably complete in the next couple of major releases ( http://blog.wolfram.com/2009/11/06/a-remarkable-year-ahead-for-mathematica/). Plenty of time for linguistic evolution and ramp-up, no? On Thu, Nov 12, 2009 at 1:52 PM, DrMajorBob <btreat1 at austin.rr.com> wrote: > Documentation that documents? Are you CRAZY, Vince? > > Why would we want to know what Automatic means, for goodness sake! > > Bobby > > On Thu, 12 Nov 2009 05:00:13 -0600, Vince Virgilio <blueschi at gmail.com> > wrote: > > On Nov 11, 4:36 am, fd <fdi... at gmail.com> wrote: >> [snip] >> >>> What I'm not quite happy with is when some options are given as >>> strings (for e.g. the ColorFunction option in DensityPlot[Sin[x y], >>> {x, 0, 3}, {y, 0, 3}, ColorFunction -> "BlueGreenYellow"] ), this >>> confuses me a bit as I can't know all the options available apart from >>> looking in the documentation . . . >>> >> [snip] >> >> "documentation" is precisely where all available options should be >> listed [*]. Fun as it might be, Mathematica should not assume or >> tacitly rely on users spelunking to find approved options (not >> obsolete, deprecated, or experimental). >> >> Vince Virgilio >> >> [*] This applies as well to default option values. For example, >> PlotStyle's spec as Automatic in Plot[] would benefit from some >> elaboration. >> >> > > -- > DrMajorBob at yahoo.com >