Original Message
Subject: Re: Beginner's choice: R or MatLab ?
Date: Thu, 2 Apr 2009 04:42:09 -0700 (PDT)
From: Dennis E. Slice dsl...@morphometrics.org
To: morphmet@morphometrics.org
References: 49d49be7.2090...@morphometrics.org
Matlab is certainly a good choice for folks with a little money. For
instance, I dropped in on a presentation by the Avizo/Amira folks
yesterday (http://www.tgs.com/products/avizo.asp), which is a very nice,
but pricey, package, and they provide direct links to/from Matlab. Also,
I recall that when working with engineers, they really like Matlab for
its signal- and image-processing features.
If you primarily want to get some thing done, then the choice comes
down to which does what you need to do. My interest is more towards
helping as many people as possible get their things done, so I focus
on cross-platform, (semi)open-source solutions. I don't believe it is
the case with Matlab, but in the past I have too often surrendered
significant personal or research funds to a company only to have them
disappear after I have committed myself to their product - e.g., the
old Morpheus.
Some years ago, it was unclear to me whether or not the investment in
learning R was worthwhile. That is not an issue today. Yes, it is worth
it. I advise my students now that if they can't easily do what they want
to in Java or R, then they probably shouldn't be doing it without
professional programming help.
Ciao, ds
morphmet wrote:
Original Message
Subject: Re: Beginner's choice: R or MatLab ?
Date: Wed, 1 Apr 2009 14:26:49 -0700 (PDT)
From: G. Alex Janevski galex...@umich.edu
To: morphmet@morphometrics.org
References: 49d3652c.1070...@morphometrics.org
I second most of what Matt said. I would also argue that R has a fairly
steep learning curve for a programming or statistics novice. I don't
know how that compares to what would be faced using MatLab. However,
once you get past the learning curve, the benefits of R are
extraordinary. As others have commented, R is free and multi-platform.
This means anyone can use it at any time, including poor students and
researchers. Also, it's a fairly small piece of software, given what it
does. It doesn't require add-ons, though packages are available to do
almost everything imaginable. In fact, if I have one major complaint,
it's that I've often done things by hand before realizing someone else
had already created a package with a function that did it better and
faster than I could (thoroughly searching the archives is a must). R is
structured to encourage users to contribute their work in the form of
packages. This is really useful for a research community. If someone on
this list develops a great morphometrics package and puts it on the R
site, it's readily available for everyone else to use, and, arguably
more importantly, to modify. Also, the graphics capabilities are
outstanding. I don't know whether that also applies to MatLab, but it's
a nice part of R.
morphmet wrote:
Original Message
Subject: Re: Beginner's choice: R or MatLab ?
Date: Tue, 31 Mar 2009 11:50:46 -0700 (PDT)
From: Matthew Burton-Kelly matthew.burtonke...@gmail.com
To: morphmet@morphometrics.org
References: 49d1f872.6090...@morphometrics.org
Considering I've never used Matlab, I have little to contribute to
that particular question. Using R, however, has highlighted some
particular weaknesses in my understanding of mathematics and
statistics, so I think it has helped me get ahold of theory a little
better because you have to think everything out first.
I like R because it is open-source and because it has a programming-
like interface, but having never taken a modern programming course,
some of the data types completely baffle me, so I end up doing things
the long way round.
Matt
--
Dennis E. Slice
Associate Professor
Dept. of Scientific Computing
Florida State University
Dirac Science Library
Tallahassee, FL 32306-4120
-
Guest Professor
Department of Anthropology
University of Vienna
--
Replies will be sent to the list.
For more information visit http://www.morphometrics.org
