Which: Matlab, Mathematica, Mathcad

[Jim Thompson]

Clients are starting to shovel data at me in
Matlab/Mathematica/Mathcad formats.

Can some regular users guide me in choosing which to purchase?

Thanks!

...Jim Thompson

 

[Chuck Harris]

Clients are starting to shovel data at me in
Matlab/Mathematica/Mathcad formats.

Can some regular users guide me in choosing which to purchase?

Thanks!

...Jim Thompson

That's an easy one, Octave. It's free, it runs under Linux and 'doze,
and does a nice jobe of emulating Matlab. There are some slight differences,
but no show stoppers. The author is trying to eliminate as many of those
differences as possible.

Google Octave Windows download

You'll find a source.

-Chuck Harris

 

[Winfield Hill]

Jim Thompson wrote...

Clients are starting to shovel data at me in
Matlab/Mathematica/Mathcad formats.

Can some regular users guide me in choosing which to purchase?

It'll have to be the same format that's being shoveled at you.

 

[[email protected]]

1) Matlab

Excels at matrix-based numerical computation. Will crunch through
large data sets easily. Lots of specialized toolboxes (cost extra
unfortunately) for things like signal processing, financial operations,
neural networks, etc. They've integrated Maple's symbolic computation
abilities as a separate toolbox so you can (kinda) do symbolic
integration, etc. as well as numerical computation, but the latter is
really where it excels. Matlab has several clones (Scilab, Octave being
the ones that come to mind first...both are Open Source) but IMHO
Matlab is superior enough to actually be worth purchasing.

2) Mathematica

Excels at symbolic computation. Steep learning curve. Others can
fill in more details.

3) MathCad

Dreck.

Does symbolic computation like Mathematica, but not nearly as well
nor as extensively. Does numeric computation like Matlab, but not
nearly as well (just try importing a 50000-element vector from a text
file...I dare you) nor as extensively.

Mostly, it's a toy meant for educational use (which it is suitable
for...mostly, though it tends to make students lazy such that they will
fire up MathCad to solve a linear equation in 1 unknown...pathetic).

The user interface will drive you nuts.

 

[Tim Wescott]

1) Matlab

Excels at matrix-based numerical computation. Will crunch through
large data sets easily. Lots of specialized toolboxes (cost extra
unfortunately) for things like signal processing, financial operations,
neural networks, etc. They've integrated Maple's symbolic computation
abilities as a separate toolbox so you can (kinda) do symbolic
integration, etc. as well as numerical computation, but the latter is
really where it excels. Matlab has several clones (Scilab, Octave being
the ones that come to mind first...both are Open Source) but IMHO
Matlab is superior enough to actually be worth purchasing.

2) Mathematica

Excels at symbolic computation. Steep learning curve. Others can
fill in more details.

3) MathCad

Dreck.

Does symbolic computation like Mathematica, but not nearly as well
nor as extensively. Does numeric computation like Matlab, but not
nearly as well (just try importing a 50000-element vector from a text
file...I dare you) nor as extensively.

Mostly, it's a toy meant for educational use (which it is suitable
for...mostly, though it tends to make students lazy such that they will
fire up MathCad to solve a linear equation in 1 unknown...pathetic).

The user interface will drive you nuts.

When was the last time you used MathCad? I've never used Mathematica,
but I've been using MathCad for years, for computations far more serious
than solving 1st-order linear equations, and it works pretty well for me.

 

[Tim Wescott]

Jim Thompson wrote...



It'll have to be the same format that's being shoveled at you.

MathCad and MatLab use two markedly different formats, once you go
beyond comma-delimited files to store arrays they are totally
incompatible. I haven't used Mathematica so I can't comment on it.

Chuck's suggestion was spot-on: I started using Octave when I started
the business, with the intention of buying Matlab when I had to.

I'm still waiting.

 

[artie]

Jim Thompson wrote...

It'll have to be the same format that's being shoveled at you.

As others have mentioned, Mathematica does have a steep learning curve,
but excels at symbolic mathematics. It's also extensible into pretty
much any realm you want.

There are also free Mathematica notebook readers available from
Wolfram, so if all you need to do is read and print, that's the way to
go.

 

[Tim Wescott]

1) Matlab

Excels at matrix-based numerical computation. Will crunch through
large data sets easily. Lots of specialized toolboxes (cost extra
unfortunately) for things like signal processing, financial operations,
neural networks, etc. They've integrated Maple's symbolic computation
abilities as a separate toolbox so you can (kinda) do symbolic
integration, etc. as well as numerical computation, but the latter is
really where it excels. Matlab has several clones (Scilab, Octave being
the ones that come to mind first...both are Open Source) but IMHO
Matlab is superior enough to actually be worth purchasing.

2) Mathematica

Excels at symbolic computation. Steep learning curve. Others can
fill in more details.

3) MathCad

Dreck.

Does symbolic computation like Mathematica, but not nearly as well
nor as extensively. Does numeric computation like Matlab, but not
nearly as well (just try importing a 50000-element vector from a text
file...I dare you) nor as extensively.

Mostly, it's a toy meant for educational use (which it is suitable
for...mostly, though it tends to make students lazy such that they will
fire up MathCad to solve a linear equation in 1 unknown...pathetic).

The user interface will drive you nuts.

MathCad (and I would hope Mathematica) is what a spreadsheet should be.
You define functions & variables in a freeform manner, interspersed
with comments where applicable and graphs where needed to illustrate
your points, and the tool takes care of all the computations. I use
MatLab (well, Octave or SciLab) for heavy number crunching, but for real
math I use MathCad.

 

[Bob Monsen]

That's an easy one, Octave. It's free, it runs under Linux and 'doze,
and does a nice jobe of emulating Matlab. There are some slight
differences,
but no show stoppers. The author is trying to eliminate as many of those
differences as possible.

Google Octave Windows download

You'll find a source.

-Chuck Harris

I'll second Octave, but, again, the input language isn't exactly the
same as Matlab. If you have hundreds of files to parse, this may be an
issue.

Also, I found it a bit tricky to get gnuplot working with it under
windows for some reason. Gmuplot is used to do any plotting of results.
My desire to use it to study digital filters was one factor that
prompted a recent move from XP back to Linux.

 

[Joel Kolstad]

1) Matlab
Matlab has several clones (Scilab, Octave being
the ones that come to mind first...both are Open Source) but IMHO
Matlab is superior enough to actually be worth purchasing.

This probably doesn't work for Jim, but educatoinal use versions of Matlab are
cheap -- something like $99... although you don't get any of the various
toolkits with it.

I think Matlab is great, but the full version really is spendy -- many
thousands of dollars!

3) MathCad
Dreck.

No it's not! MathCAD is certainly limited relative to Matlab (or anything
else ), but it's often the fastest way to just "mess around" with some
data and graphs. As soon as you need to write some reasonably programmatic
functions, though, it becomes somewhat painful and Matlab is often the better
choice.

Mostly, it's a toy meant for educational use

Meant for educational use... mmmm, yeah, I suppose. A toy? No. It's used
everyday by folks doing Real Engineering Work, including the likes of Hans
Camenzind (inventor of the 555).

MathCAD is quite cheap relative to the rest...

(which it is suitable
for...mostly, though it tends to make students lazy such that they will
fire up MathCad to solve a linear equation in 1 unknown...pathetic).

It is pathetic, but I'd have to say it's often due to the students never being
required to learn how to do it by themselves. Technology is so complicated
these days that you want to ask students to do things like solve very
difficult differential equations, build fancy convolutional encoders, perform
FFTs, etc., and there's no time (and often no reason) to teach them the
mechanics of doing so themselves, so instead you teach them how to solve the
problems using readily available tools.

The disservice occurs when students are "given" (meaning, "required to
learn!") so few "core" tools that they end up being helpless if they don't
have a perfectly round hole for their round pegs. This happens at the lower
(high school) levels where the emphasis is all on passing some standardized
tests rather than actually learning what's going on.

The user interface will drive you nuts.

I think it's OK...

---Joel Kolstad

 

[Evgenii Rudnyi]

Clients are starting to shovel data at me in
Matlab/Mathematica/Mathcad formats.

Can some regular users guide me in choosing which to purchase?

Thanks!

Well, tastes differ. Below there are some my observations about
Mathematica.

I have been working with Mathematica for the last 5 years and I like
it.

1) Fast symbolic calculation and since Mathematica 5.0 fast numerics as
well. Now numerical linear algebra is done with the same speed as in
Matlab.

2) A very consistent functional programming language. I should say that
functional programming is very important feature for scientific
computing.

3) A nice working environment. A project is a single file, so called a
notebook. You can even find books written within Mathematica.

4) webMathematica gives you a smooth transition to the Web.

We have a collection of functions in Mathematica related to MEMS
simulation, you may have a look

http://www.imtek.uni-freiburg.de/simulation/mathematica/IMSweb/

If to speak about free tools, I think Python is not bad. There are many
add-ons to Python for scientific computing now.

Best wishes,

Evgenii Rudnyi

 

[J. B. Wood]

No it's not! MathCAD is certainly limited relative to Matlab (or anything
else ), but it's often the fastest way to just "mess around" with some
data and graphs. As soon as you need to write some reasonably programmatic
functions, though, it becomes somewhat painful and Matlab is often the better
choice.

Hello, and hey! I still use my ancient MathCad 6.0 Pro (it runs fine
under Win XP) for electronics engineering work. It easily handles matrix
algebra with complex numbers as matrix elements. I've also included
MathCad pages containing text and formulas as-is into technical reports.
As stated above you can easily "mess around" with MathCad much in the same
way as you would a handheld scientific calculator or a spreadsheet. You
can put MathCad away for several months or longer and pick right back up
on it without significant relearning. Again, I'm speaking about my
antique version. My .02 worth. Sincerely,

John Wood (Code 5550) e-mail: [email protected]
Naval Research Laboratory
4555 Overlook Avenue, SW
Washington, DC 20375-5337

 

[Chuck Harris]

We have a collection of functions in Mathematica related to MEMS
simulation, you may have a look

http://www.imtek.uni-freiburg.de/simulation/mathematica/IMSweb/

If to speak about free tools, I think Python is not bad. There are many
add-ons to Python for scientific computing now.

Best wishes,

Evgenii Rudnyi

Yes, but Python is a general purpose programming language. I really
doubt that we will be turning Jim Thompson into a Programming Wonk
in this lifetime.

-Chuck

 

[Rob Young]

That's an easy one, Octave. It's free, it runs under Linux and 'doze,
and does a nice jobe of emulating Matlab. There are some slight
differences,
but no show stoppers. The author is trying to eliminate as many of those
differences as possible.

Google Octave Windows download

You'll find a source.

-Chuck Harris

Also "Scilab". There is a newsgroup for Scilab, comp.soft-sys.math.scilab

 

[Chuck Harris]

Also "Scilab". There is a newsgroup for Scilab, comp.soft-sys.math.scilab

Scilab is an excellent program, but I don't like the terms. They only offer
the free license to non commercial entities. That leaves consultants out
in the cold. They want something like $500 for the license for commercial
use. Might just as well use Matlab and Mathematica. At least
with them, there is a paid staff to provide maintenance.

I'll continue to use Octave. It is a very powerful, and well done program.
And, the price is right.

-Chuck Harris

 

[John Hudak]

Hi Jim:
FWIW, I have used Mathematica, Matlab, and Octave. I like Mathmatica
for symbolic manipulation/solutions of equations. It has (in my
opinion) a steeper learning curve than Matlab/Octave. I have not used
Mathcad. Being an engineer who does ckt design as well as control
systems engineering, as well as signal and image processing, I like
Matlab. The toolboxes contain all the functions that can be applied to
heavy duty applications. The programming paradigm in basic Matlab
carries over directly to the toolboxes. If interfacing to the realworld
for realtime analysis (signal processing and control), Matlab does a
very good job.
From my usage, Octave is very similar to Matlab in the analysis
portion, but I never tried to use it in a real-time mode. I must admit,
it is great for the price.

There is another tool that I really like, and depending on your needs,
may be a better fit. Its IDL. IDL is: a mathematical analysis library
(similar to matlab in both programming paradigm and functionality),
programming language, and the ability to embed/integrate your analysis
programs into your application for very good time response.
IDL has been around for a long time, has good support, and runs on MANY
operating systems (WinX, Linux, UNIX, VMS) and you can develop
GUIs/applications that are truly portable (X, Win, web). If a function
you need is not in it, or, you want to craft a display, or interface to
a DAS, you can write your own functions to do it...there is also a user
area that ppl contribute their functions. It does cost a bit (don't
recall, its been a while since I bought it but, it can be cost
competitive with Matlab..
http://www.rsinc.com/
Good luck
John

 

[Joel Kolstad]

Hey, can anyone out there who's familiar with Mathematica compare it to Maple
in terms of symbolic manipulation?

 

[Charlie Edmondson]

1) Matlab

Excels at matrix-based numerical computation. Will crunch through
large data sets easily. Lots of specialized toolboxes (cost extra
unfortunately) for things like signal processing, financial operations,
neural networks, etc. They've integrated Maple's symbolic computation
abilities as a separate toolbox so you can (kinda) do symbolic
integration, etc. as well as numerical computation, but the latter is
really where it excels. Matlab has several clones (Scilab, Octave being
the ones that come to mind first...both are Open Source) but IMHO
Matlab is superior enough to actually be worth purchasing.

2) Mathematica

Excels at symbolic computation. Steep learning curve. Others can
fill in more details.

3) MathCad

Dreck.

Does symbolic computation like Mathematica, but not nearly as well
nor as extensively. Does numeric computation like Matlab, but not
nearly as well (just try importing a 50000-element vector from a text
file...I dare you) nor as extensively.

Mostly, it's a toy meant for educational use (which it is suitable
for...mostly, though it tends to make students lazy such that they will
fire up MathCad to solve a linear equation in 1 unknown...pathetic).

The user interface will drive you nuts.

Hi Jim,
Mathcad got me through grad school, so I have to recommend it, although
that was (mumble) years ago, and the program has changed quite a bit
since then.

Pluses: WYSIWYG interface, so nice for equations and such, decent
import/export capabilities.

Minuses: Learning the user interface (Oh, you have to enter the limits
for the integration... Here, not there, OK, so, why do I have that
little box there?...)

Charlie

 

[John Hudak]

Hi Joel:
I used Maple probably 5-7 years ago...at that time, Mathematica was much
better...but a lot has changed.
John

 

[Chris Carlen]

Hey, can anyone out there who's familiar with Mathematica compare it to Maple
in terms of symbolic manipulation?

Mathematica is the most advanced symbolics package available.

There was a time when I was learning to analyze circuits, by writing the
time domain differential equations. I wrote out the whole system
equations for a full wave bridge rectifier including transformer
inductance, diode models, ESR of the filter caps, and the filter C.
Mathematica could solve this system (numerically), but Maple couldn't
even deal with the equations. Ultimately I discovered that SPICE was
much more effective at solving these sorts of things due to being
optimized for this form of system. But it was very educational to solve
the equations in a general purpose math package.

The ability to combine symbolics with numerics makes it much more useful
to me than Matlab, since I am often seeking to avoid pages of tedious
algebra in deriving some relation. But then I wish to immediately apply
my new function to data or generate data, which it can do as well.


Good day!

--
_______________________________________________________________________
Christopher R. Carlen
Principal Laser/Optical Technologist
Sandia National Laboratories CA USA
[email protected]
NOTE, delete texts: "RemoveThis" and "BOGUS" from email address to reply.