discuss-gnuradio
[Top][All Lists]
Advanced

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

Re: [Discuss-gnuradio] 3rd party software (was comedilib question)


From: George Nychis
Subject: Re: [Discuss-gnuradio] 3rd party software (was comedilib question)
Date: Tue, 09 Sep 2008 14:21:19 -0400
User-agent: Thunderbird 2.0.0.16 (X11/20080724)



Johnathan Corgan wrote:
On Mon, Aug 25, 2008 at 10:43 AM, Dan Halperin
<address@hidden> wrote:

On this note, but slightly tangential, there are a bunch of third party
gnuradio plugins that are not being kept up to date and even no longer work
with gnuradio without significant changes. (E.g., top block vs flow graph).
Do you think we can create space in SVN somewhere for the updated versions
of these codes? I'm thinking about gr-ucla and maybe the BBN 802.11 stuff as
well.

There are two issues with this.

The first is that the software you describe seems to not have a
maintainers, and thus the bit-rot has set in.  Having it in our SVN
doesn't change anything.

Secondly, with a few very rare exceptions, the host software in our
repository has all of its source code copyrights assigned to the Free
Software Foundation.  So, even if we had committed maintainers for
this 3rd party code, we'd need to get copyright assignments from the
original authors.

That said, I'd be happy to help you and/or others to get them up to
date.  Since they are all GPL, there is no issue with hosting them
somewhere else and patching them up to be current.  We could even have
a wiki page dedicated to these efforts.


I'd like to bring this back up. I think Dan is poking at something that has really bugged me for some time too. I think there is a ton of GNU Radio code floating out there that is potentially extremely useful to other people, that is never adopted in to GNU Radio because the barrier of entry of code into GNU Radio is extremely high.

Copyright assignments, developer branch, QA code, following conventions, awaiting approval, hesitation of maintainers to let in code they're not familiar with... let's face it, getting code in to GNU Radio if it's not a bug fix is a serious hassle that not everybody has time for. I completely understand the reason for all of the above, but it's the reason why we need something for 3rd party software.

There have been tons of postings to the list about projects people have done, that maybe only work with a certain snapshot of GNU Radio. I think this is OK. For example, someone might want to work with the gr-ucla code and they don't care about new enhancements to GNU Radio which make it unworkable. They check out a snapshot, and gr-ucla code.

So, you're probably thinking: OK so why do we need 3rd party support, they can just check out the gr-ucla code and then an old version of GR and be happy. Well... if they went to update that gr-ucla code they need to contact the UCLA people. Whereas if there was a third party repo, they could just makes changes for example, upgrading it to work with the current GR release.

Additionally, 3rd party support is a great way to gather code for someone to look at what else is available out there. For example, there was multi-path code someone posted a while back that they had to go through some hassle to setup their own server to host it. That server is likely down by now. If there was some additional repo they could commit it to, we'd have it right here right now to work from, and so could other people.

Also, I spoke with Kyle that he revamped some of the OFDM code and has been trying to get it in to GR but has had problems with either contact or acceptance of his changes. I'm sure the code could definitely be used by others... I think there should be somewhere that Kyle could upload his code and let others use it.

Maybe an additional SVN is what we need. Maybe a GIT repo. A wiki so that people can provide some information about their code. We need something. I don't care of its officially supported by GNU Radio or not, but I think it would certainly benefit the GNU Radio community.

- George




reply via email to

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