On Mon, May 09, 2011 at 11:33:32AM -0700, Colby Boyer wrote:
One of big reasons I think that people struggle with GNURadio is that
is jams so many different fields of expertise into one package.
1. Digital Comms people (aka the Maths people) cannot program
themselves out of a wet paper bag, for the most part. This is what I
have seen in industry and academia.
2. Software people get lost in all the digital comm and signal
processing lingo. While they can program, they really don't understand
what each block actually does.
3. Hardware people also get lost in the digital comm stuff, and also
some of the software. However, they tend to be less confused than the
'maths' people on the programming aspect
This sounds like it's from the 80ies. Things have changed a bit; education
has become more 'fluid' and it's not a big deal to know a bit of
everything. Of course, you'll never know all of everything, but if you
engage in a new project, you always have to learn new stuff as well.
I believe that our university is no different from many others in a
sense that if you graduate with a major in communications engineering,
you're supposed to know at least two of your these; and let me also
point out that you can omit a *lot* of hardware knowledge if you want to
use GNU Radio.
Let me rephrase: People struggle with GNU Radio because it jams so many
*difficult* fields into one package. SDR is not, and never will be,
easy. Writing a receiver for standard XYZ (which can operate in
real-time) is always a daunting task, and there is no way to make it
simple.
The difference between Matlab and GNU Radio is that the guys from
Mathworks make it *look* easy, whereas are favourite software package
scares away the people who are not used to scarce documentation and
autotools.
My apologies, Colby, for interpreting your email in a way you probably
didn't mean it. But what you wrote sounds like it's the user's fault GNU
Radio isn't used more often. And well, it is of course, but frankly, we
haven't made the biggest effort to make the decision between
Matlab/Simulink and GNU Radio flip to GR's benefit.
To a non GPL-philic, non-nerd, why choose GNU Radio? There is no reason:
- Matlab is generally free of charge for universities
- Matlab is used by the industry
- Matlab is better documented and has a wider user base
- Simulink has more blocks already incorporated
- Matlab/Simulink has a much wider applicability outside realtime DSP
Here at CEL, the majority of student's projects are done using
Matlab/Simulink. In those cases where the students chose GNU Radio, all
these things were true:
- The student had some background in open source dabbling (e.g. using
Linux)
- The student was not scared away when I explained that the learning
curve is intimidating
- The motivation to do the thesis was beyond simply wanting to earn
credits
I guess if we really want more people to use GNU Radio, it's up to us,
the active community, to get them on board.
MB
PS: I'm not really good at short posts :)