[Battlemesh] Thanks for the v5 ;-)

[Michel de Geofroy]

Just impose unique device and document tests then let each team pay for its
hardware and flash it and bring it... maybe Mike BB can provide more 0m2p
at reduced price for testing...

On Tue, Apr 3, 2012 at 6:03 PM, Gabriel Kerneis <kerneis at pps.jussieu.fr>wrote:

> On Tue, Apr 03, 2012 at 02:17:21PM +0800, Yeoh Chun-Yeow wrote:
> > I think that we need to ensure our binary images ready before the
> > event. Off course, we need identify the hardware platform earlier.
>
> Sadly, this is hardly surprising.  For instance, six months ago, on this
> mailing-list:
>
> On Thu, Sep 22, 2011 at 02:20:52PM +0200, Pau wrote:
> > 2011/9/22 Marek Lindner <lindner_marek at yahoo.de>
> > > On Thursday, September 22, 2011 06:52:22 Andrew Parnell wrote:
> > > > One thing I would like to suggest is to have a more formalised
> arrangement
> > > > for the test network that we will use for WBMv5.  Though I am a
> newcomer
> > > > to this, one thing that I couldn't help but notice was that many of
> us
> > > > were there for 5-6 days before we were actually deploying the
> Foneras and
> > > > building a test mesh.  It seems that this is not the most efficient
> use of
> > > > our time.  Perhaps we could find a way to require that each
> team/group who
> > > > wishes to submit code for testing must have this ready to go
> /before/ the
> > > > event begins.  This way, a firmware image can be prepared in
> advance, and
> > > > once people begin to arrive, we can set up the mesh immediately and
> have
> > > > much more time for testing.  We would probably also want to have
> another
> > > > deadline(s) sometime in the middle of the event perhaps, where
> updates can
> > > > be provided and a new firmware image(s) compiled.  Hopefully this
> would
> > > > allow us to use the short time we have to the maximum benefit, and we
> > > > could really get some good tests/statistics compiled as a result.
> > >
> > > It is not that we did not try in the past. We had deadlines / teams /
> etc
> > > but clearly lacked direction imposed by a "test leader" or "test
> group". The
> > > misery starts when you ask the question: What are we going to test ?
> You
> > > will find people chiming in that are fairly silent most of the time
> but feel
> > > they have to "defend" their protocol.  In short: As long as we don't
> have
> > > someone (preferrably a protocol neutral person) who takes matters into
> his
> > > hands I don't expect any improvement this year either.
> > >
> > I'm also a newcomer at WBM, the last one was my first one. Before I went
> > there, I thought that this was an event where the main objective was to
> test
> > mesh protocols and put them in battle. But after, my thought was that
> this
> > is an event to meet people and speak about some geek topics.
> > For me that is good, but maybe to push the original objective a little
> more
> > would be nice.
> > One of the main restrictions I found is the hardware. Fonera is a very
> > limited device and when you are running 4 or 5 protocols, it can do some
> > strange things, and the tests become unreliable. Maybe we should put some
> > efforts in to have another kind of hardware, we can find some sponsors
> who
> > can give us some hardware. Or maybe we can put a special tax for
> spending on
> > hardware (10€ per person = 600€ = 10 new devices).
>
> Proposing and discussing some test scenarios in advance is definitely an
> improvement, but it turned out to be useless without a reliable mesh setup
> to
> perform the actual measurements.
>
> --
> Gabriel
> _______________________________________________
> Battlemesh mailing list
> Battlemesh at ml.ninux.org
> http://ml.ninux.org/mailman/listinfo/battlemesh
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://ml.ninux.org/pipermail/battlemesh/attachments/20120403/6ec2d0a0/attachment.htm>