discuss-gnuradio
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Solved: [Discuss-gnuradio] gnuradio under Ubuntu under Parallels on


From: Marcus D. Leech
Subject: Re: Solved: [Discuss-gnuradio] gnuradio under Ubuntu under Parallels on MacPro dual Quad core
Date: Mon, 17 May 2010 17:42:57 -0400
User-agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.1.9) Gecko/20100330 Fedora/3.0.4-1.fc12 Thunderbird/3.0.4

On 05/17/2010 04:45 PM, HR Myler wrote:
>
> After seeing a post by Josh Blum stating that I should run HUD because
> he thought that since "the usrp2 gnuradio driver has a linux specific
> packet ring, and therefore wont work on a bsd environment." I was
> momentarily outraged. Does it not often seem that if you are not
> running what (the ubiquitous) "they" are running the answer is always
> the same? Go get on a different system is just the easy way out.
>
When developing software these days, it's important to start out life by
picking the environment
  most likely to be broadly useful *first*.  Eliminating windows for a
second, that leaves Linux.
  While it's true that it's possible to build many classes of
applications in such a way that they're
  utterly-portable from OS-to-OS, it's hard, and it's harder when the
application has to touch
  anything other than a highly-abstracted interface to the filesystem,
and few rigidly-standardized
  applications APIs.

It's not a big surprise to me that some of the bits 'n pieces of the Gnu
Radio network interface
  code use some Linux-specific networking "stuff".

You are using a VM that may, or may not, faithfully mimic the behaviour
of real hardware.  That's
  not the environment that Gnu Radio was developed on, so it should
hardly be surprising that
  the kinds of "weird cases" produced by a VM environment haven't been
tested-for, or accounted for.

The kinds of combinatorial explosions that happen in code when the code
*does* try to account for
  "all possible barriers to correct functionality" are truly hideous,
and lead to code that is
  profoundly hard to maintain--speaking from agonizing experience.  Over
and over again.
  For the past 31 years.

Consider a reductio ad absurdium argument.  What if there was hardware
out there that couldn't
  add certain numbers properly, and occasionally didn't do multiplies
correctly?  Software generally
  assumes that when you add 2+2, you get four.  It generally doesn't
need to verify that the
  underlying CPU is actually sane.  But one *could* argue that software
should perhaps be robust in
  the face of impossible odds, and do such checks, perhaps only when it
detects "weirdo environment
  number 57".   But it's easy to see how quickly such a piece of
software becomes elephantine and
  unmanageable.

Probably the *right* thing for MacOS, is to have native ports of the Gnu
Radio code, and UHD,
  and so on.  I think basic Gnu Radio functionality is already available
on MacOS, so some keener
  with BSD experience could perhaps lend a hand and get UHD running
natively on MacOS/*BSD.





-- 

Principal Investigator
Shirleys Bay Radio Astronomy Consortium
http://www.sbrac.org





reply via email to

[Prev in Thread] Current Thread [Next in Thread]