MATLAB vs PV-WAVE
Henrik Nordborg
In article <3abslc$m...@access4.digex.net> dwb...@access4.digex.net
(Dennis W. Brown) writes:
From: dwb...@access4.digex.net (Dennis W. Brown)
Newsgroups: comp.soft-sys.matlab
Date: 15 Nov 1994 22:01:00 -0500
Organization: Express Access Online Communications, Greenbelt, MD USA
For various reasons, I'm looking at replacing MATLAB with PV-WAVE. I
haven't had a chance to load and play with the demo version of PV-WAVE
yet, however, from the literature, it appears the two are similar. It
also looks like PV-WAVE has some very distinct advantages (ex: it's
supports multiple data types (vice the RAM hogging doubles of MATLAB)).
Anyone have any opinions comparing the two?
Dennis
dwb...@access.digex.net
This is interesting, since I am considering doing just the opposite. I think
Matlab is definitely the superior, and I have been working with both of
them. Matlab has a much better programming language, a much
better user interface and can do at least as good graphics.
Regards,
--
--------------------------------------------
Henrik Nordborg
Institute of Theoretical Physics
Swiss Federal Institute of Technology (ETH)
CH-8093 Zurich
phone: +41-1-633 25 73
fax: +41-1-633 11 15
hen...@itp.phys.ethz.ch
--------------------------------------------
John Cristion
>
> For various reasons, I'm looking at replacing MATLAB with PV-WAVE. I
> haven't had a chance to load and play with the demo version of PV-WAVE
> yet, however, from the literature, it appears the two are similar. It
> also looks like PV-WAVE has some very distinct advantages (ex: it's
> supports multiple data types (vice the RAM hogging doubles of MATLAB)).
> Anyone have any opinions comparing the two?
>
> Dennis
> dwb...@access.digex.net
>
Ok, here's my $.02.
I'm an engineer and I like Matlab better than PV-Wave. I think that for
simulations, quick looks, etc. it is easier for "ME" to use. But, the
programers that I work with like PV-Wave better, because it is more flexible
for their uses (i.e., better at memory management, faster, easier to integrate
with C, etc.).
So, I think that it depends on what you are trying to do. I'll stick with
Matlab for now and allow the programers here to stick with PV-Wave. An
optimal, albeit expensive, solution.
John
____________________________________________________________________
John A. Cristion
cris...@aplcomm.jhuapl.edu
The Johns Hopkins University/Applied Physics Laboratory
Johns Hopkins Rd., Laurel, MD 20723
____________________________________________________________________
Ralph Finch
of code for that. At home I use Matlab.
I would say Matlab is better at numerical things; IDL for data
analysis. IDL's big advantage is support of different data types.
However Matlab's numerical capabilities strike me as significantly
superior. I prefer somewhat Matlab's syntax for arrays.
Matlab support is better, IMHO. I haven't done a lot of graphics
programming in Matlab, but the Handle Graphics concept seems way
superior to the standard graphics concept in IDL. Of course the
*implementation*, for a particular application, might be better
supported in one or the other, but in the long run I have to say
handle graphics will win.
Overall, Matlab seems the better designed product to me. If Matlab
would support different data types; real chars; structures; then it
would clearly be superior in a large majority of cases, IMO.
--
Ralph Finch 916-653-8268:voice 916-653-6077:fax
rfi...@dop.water.ca.gov / finger for PGP public key
Any opinions expressed are my own; they do not represent my employer
David B. Enfield
> I'm an engineer and I like Matlab better than PV-Wave. I think that for
> simulations, quick looks, etc. it is easier for "ME" to use. But, the
> programers that I work with like PV-Wave better, because it is more flexible
> for their uses (i.e., better at memory management, faster, easier to integrate
> with C, etc.).
There was a thread on Matlab vs. IDL some months ago and I remember
very similar comments. I know that the things that bother me about
Matlab are the things that IDL is supposedly good at, such as
I/O capabilities for HDF, netCDF, and more cooperative multitasking
& better memory management. But overall, I still wouldn't give up
on Matlab to get those advantages. I just hope that somebody at the
Mathworks is paying attention, because we DO want these things and
there IS competition...(and we ARE paying a lot for our Matlab license).
Dave Enfield
NOAA/AOML
Miami, FL
Dennis W. Brown
Alain Jacques
No experience with PV-WAVE but with a very similar product, IDL. From my point
of view, the advantages of IDL are its capability to handle multidim. arrays
and a much efficient handling of bitmap based data.
IDL demos for different platforms are available from ftp://gateway.rsinc.com or
ftp://ftp.germany.eu.net/shop/CreaSo. These are full featured but time limited
versions.
Alain
Joe Hicklin
> I just hope that somebody at the
> Mathworks is paying attention, because we DO want these things and
> there IS competition...(and we ARE paying a lot for our Matlab license).
>
> Dave Enfield
> NOAA/AOML
> Miami, FL
>
I would guess that more than 100 people here
read this newsgroup regularly, perhaps twice that many.
As an engineer here I find every description of a competitors
advantage over "my" product painful. You can bet that we watch
this news group carefully, and that it strongy influences our
development direction.
Karl J. Molnar
-From: dwb...@access4.digex.net (Dennis W. Brown)
-For various reasons, I'm looking at replacing MATLAB with PV-WAVE. I
-haven't had a chance to load and play with the demo version of PV-WAVE
-yet, however, from the literature, it appears the two are similar. It
-also looks like PV-WAVE has some very distinct advantages (ex: it's
-supports multiple data types (vice the RAM hogging doubles of MATLAB)).
-Anyone have any opinions comparing the two?
-Dennis
-dwb...@access.digex.net
>...
> I just hope that somebody at the
> Mathworks is paying attention, because we DO want these things and
> there IS competition...(and we ARE paying a lot for our Matlab license).
>
I've had the opportunity to use both IDL and MATLAB and the following
are the main good/bad points that I see of each:
IDL (version ?):
Good Points:
- The language is much more like a standard programming language
(whether this is good or not depends on your point of view!).
- The debugger works well and makes programming/debuging fairly easy.
- Simple to add C function calls.
- Speed advantage over MATLAB for code that has many branches/loops
(though still not great compared to a real programming language).
- Has multidimensional arrays and structures.
- Good plotting capabilities.
Bad Points:
- The standard MATH support stinks - supports only single precision in many routines.
- Very expensive for SPARC platform (others?).
MATLAB 4.2
Good Points:
- Good math routines (though don't ever use them blindly!).
- Very large amount of supporting code (ie. many toolboxes).
- Good plotting capabilities.
- Simple to add C function calls.
Bad Points:
- Poor performance when loops/branches in code.
- Debugger is poor at best.
- Has neither multidimensional arrays or structures.
- Very expensive for SPARC platform (others?).
Experiences/Comments:
I really like both of these tools for prototyping algorithms,
especially when linear algebra / matrix manipulation is involved.
For simulations that take a long time, however, you'll save more
time in the long run to code your programs in C/C++/FORTRAN or
something else. I've seen speedup times of 100 to 1 for IDL code
that has been turned into C code (IIR filter design code). Likewise
for MATLAB. Which is best? For me - the one that is less expensive
will win.
Karl Molnar
mol...@rtp.ericsson.se
Corbett Rowell
0
0
1.2
2.3
1.1
.
.
.
Each line is a density in a 3D grid, so the total number of lines is
y * y * y. Can Matlab visualize this dataset by creating a color
for each density? For instance, I have a globe with densities = 2 and
the surrounding space with densities = 0.
Thanks in advance.
Please reply to this email address: ph...@ee.ust.hk
-Corbett
David B. Enfield
Matlab's Good Points:
- Good math routines (though don't ever use them blindly!).
- Very large amount of supporting code (ie. many toolboxes).
- Good plotting capabilities.
- Simple to add C function calls.
----------------------------------------------------------------
I would add one very important plus to that list: - tech support
Joe Hicklin
> So, when will Matlab support other data types??
>
> Theo P.
Nice try Theo! I do understand your need to know what's coming down
the pipe. On the other hand, I'm pretty sure that the best way for
us to serve our customers is to announce products at the "right" time,
and this isn't the time to announce that feature.
Joe Hicklin
Theodore G. Pantazopoulos
Beau Paisley
>would support different data types; real chars; structures; then it
>would clearly be superior in a large majority of cases, IMO.
>--
You might look at O-Matrix if you want to have different data types.
It will support character, logical, integer, real, double and double-complex
type.
Ralph Finch
>
> So, when will Matlab support other data types??
Joe> On the other hand, I'm pretty sure that the best way for
Joe> us to serve our customers is to announce products at the "right" time,
Joe> and this isn't the time to announce that feature.
Please reconsider. It is important that we know in general what
companies are working on. It would be very helpful if Mathworks said,
'Yes we probably will have data types within a year' or 'No we have no
plans to use other data types.'
How would the Mathworks like it if, for instance, they couldn't
receive Win95 from Microsoft except when it hit CompUSA's shelves like
everybody else? I am very sure you have had Win95 beta's for at least
a year, not to mention frequent contacts with Microsoft concerning the
direction of Win95.