[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Discuss-gnuradio] dive into gnu-radio
From: |
Marcus D. Leech |
Subject: |
Re: [Discuss-gnuradio] dive into gnu-radio |
Date: |
Wed, 16 Mar 2016 22:23:22 -0400 |
User-agent: |
Mozilla/5.0 (X11; Linux x86_64; rv:31.0) Gecko/20100101 Thunderbird/31.7.0 |
On 03/16/2016 08:34 PM, Timothée COCAULT wrote:
I think there is cruel lack of explanation (not documentation) for the
GNU Radio blocks.
Reminds me of a Dylan Thomas piece "A Book with everything about wasps,
except why."
Gnu Radio is "embedded" in a number of different complex disciplines,
like it or not. There's no getting around it. One might as well
declare that the 2nd law of thermodynamics is awfully inconvenient,
so let's just wish it away.
There is a difference, I think, between the usability of a tool, and the
ease-of-understanding of the discipline for which the tool was
created. Some of the documentation complaints are, quite rightly,
about usability of the tool. Some are, I would posit, about the
steep learning curves of the relevant disciplines. The Gnu Radio
development team is well-placed to make great progress in the
former, and less, I would assert, in the latter. It may be "nice"
if every block included the equivalent of a chapter or two
treatise on the subject matter at hand, as if from a college-level
textbook, but I don't see that happening, unless someone
(a technical writer) puts in the time to do it. It would a poor
substitute for curling up with the appropriate reading material
oneself in front of a fire, with a cup of Joe, and ones favorite dog
at ones feet.
Let me use one of my semi-famous "proximate analogies". When you
acquire an eCad system for the first time, to layout circuits and
circuit boards, ones frustration is usually about the tool. Becoming
frustrated that it requires that you understand simple concepts
like current and voltage, what an op-amp is, and how to use one, and
what the truth-tables are for a 74LS74 (and again, why you would
use one), is not likely to happen. One doesn't expect ones tooling
to substitute for the necessary background knowledge. I cannot imagine
anyone sitting down in front of a VLSI design tool for the first
time, and loudly exclaiming that the "documentation sucks" because it turns
out that being successful with the tool requires that you understand
a bit (or a lot, really) about VLSI design as a technical discipline.
But I see folks arrive, every day, at Gnu Radio, expecting that no
relevant background should be required, that if they aren't immediately
successful with the tool, that it must be the tools fault. Granted,
some of Gnu Radio's documentation could use improvement, and in
fact my decades of experience in technology would suggest that no
documentation is EVER "good enough". Because we're all different,
and we all learn in different ways. But I think it's exceedingly
important to understand the difference between "tool usage" documentation,
and "a substitute for a 4 year EECS degree". This isn't elitism or
snobbery, it's just that expecting the tool itself to be a substitute for
the necessary background just isn't reasonable. The Gnu Radio
project doesn't have a large team of technical writers, textbook writers,
course-ware writers, etc, etc just waiting to contribute. The
emerging Gnu Radio *ecosystem*, however, may be a good place to turn.
Yearly Gnu Radio conferences, folks like Jonathan Corgan who run
intro-to-SDR courses, the excellent work of folks like Michael Ossmann.
Online forums like www.dsprelated.com, www.complextoreal.com.
The Gnu Radio project cannot possibly be your "one-stop shop" for all
things related to radio, DSP, real-time software design, etc, etc.
It's just not practical.
- Re: [Discuss-gnuradio] dive into gnu-radio, (continued)
Re: [Discuss-gnuradio] dive into gnu-radio, Tom Coleman, 2016/03/16