Re: Re: Any Mathematica 6 book yet?
- To: mathgroup at smc.vnet.net
- Subject: [mg81658] Re: Re: Any Mathematica 6 book yet?
- From: "sdw" <warwick at jps.net>
- Date: Mon, 1 Oct 2007 04:39:37 -0400 (EDT)
- References: <200709261035.GAA06256@smc.vnet.net> <email@example.com> <firstname.lastname@example.org>
having dipped into Mathematica for many years, I have to agree with
Nasser on this.
what I often find is that Mathematica has the basic functions to
"get you started" but never the richness of functionality in any one
area to make it usable out of the box.
recently, I needed to do some simple continuous constraint linear
programming. yes, Mathematica has the function. but ask for basic analysis
feedback such as shadow prices, breakpoints, standard linear programming
solver responses and you're dead in the water.
try to do statistical regression analysis. yes, you can get some
fundamental model statistics, now, try to get forecast statistics - dead
try to build a bar chart showing simple data, like a pretty line chart
- no problem. try to build a bar chart with grouped categories with
multiple series of data, with each series colored differently - dead again.
(remember, Mathematica is supposed to be a data visualization package too! )
for the theorists out there, you have a great and powerful language for
developing extraordinary constructs. no intent to deny
what needs to happen is for Mathematica to conduct usability analysis
in different disciplines and attempt to incorporate in the base product
sufficient functionality to support 90% of the typical uses, not 20%
as it seems today. we just cannot justify paying $500 for extra packages
to reasonably cover each new requirement as it emerges, or spend
hours developing the tools individually.
An interesting survey would be to ask the developers how many hours it took
to develop each of the "demonstrations" in the demonstrations project.
Another pointer on future development. it would be interesting to see if it
were possible to emulate a feature that MS has implemented in their office
tools - the macro recorder. here, one has the ability to interact visually
and construct a graphic using drawing tools, with the results being
translated automatically into the underlying macro language, and then
allowing complex editing of the resulting code. Mathematica is currently
constrained to be a text-driven graphics engine. Think about ways to make
this more bi directional. even if a separate notebook is "generated" as part
of the process, and users end up copying the resulting code into their
applications. This could greatly improve the user efficiency.
"Nasser Abbasi" <nma at 12000.org> wrote in message
news:fdl98h$cvd$1 at smc.vnet.net...
> "Murray Eisenberg" <murray at math.umass.edu> wrote in message
> news:fdf236$20u$1 at smc.vnet.net...
>> I sometimes find it utterly amazing how deeply some mathematicians,
>> scientists, or engineers believe that Mathematica is deficient in being
>> handle this or that kind of problem (typically said about numerics, but
>> sometime about graphics, too).
> There are few things that would make Matematica 'easier' to use by those
> engineers who are so used to that 'other' tool (which we all know what it
> is). I am amazed that WRI still have not added such features after all
> these years. Some of these things are:
> 1. Ability to zoom in and out of any part of a plot using the mouse. (I
> this can now be programmed using the new Dynamic features, but this needs
> be a build-in feature, just like the ability not to rotate 3D plots using
> 2. an EASY to use debugger. Something as basic as seeing line numbers and
> setting a break point on a line number. The current debugger is
> useless for me, I can't ever figure how to use it. So I still debug my
> using Print. Using the 'other' system, never had to do this.
> 3. To Make Mathematica more attractive to 'engineers', add more control
> systems functions to basic Mathematica. May be combine the functions in
> control systems application and the functions in the image processing
> application into Mathemtica core. WRI might lose few bucks in sales from
> those applications, but will gain thousands more in sales of Mathematica
> itself (adding $5 to the price of Mathematica itself might end up
> things out).
> I can make a much much bigger list, but we can just start with the above
Prev by Date:
Re: Plot3D with NDSolve
Next by Date:
Re: Install problem, 6.0, Mac OS 10.4.10: No kernel connect?
Previous by thread:
Re: Problem with lists as matrices for OLS and statistical inference
Next by thread:
Re: Re: Re: Any Mathematica 6 book yet?