[Ns-developers] iNSpect Update

[Gustavo Carneiro]

2008/9/3 Ken Anderson <keanders at mines.edu>

> Gustavo,
>
> My name is Ken Anderson and I am working with Jeremy on iNSpect. I am
> writing you to let you know the changes we're making to iNSpect as a result
> of this conversation.
>
> The first change we made is to implement scaling in iNSpect again. This
> should help us move toward a program that can accurately guess the display
> parameters. Once autoscaling is complete, iNSpect will use default values
> until it finds them to be incorrect, and then try to correct itself during
> runtime.
>
> Another change I made was a minor bugfix where the usage text didn't
> display when it should. Now it does.
>
> Finally, I am now working on removing the configuration file requirement
> for iNSpect. Instead, the default values will be used whenever a
> configuration is not supplied. A custom configuration file can still be
> loaded by attaching -c <configuration file> argument. I expect this to be
> done early next week.
>
> We are very happy to hear all the feedback and concerns you may have about
> iNSpect. Feedback like we've seen here helps us focus on what's important
> regarding iNSpect.


Thanks.

Regardless of your progress, I started experimenting with a python based
visualizer [1] and have had some good progress.  I just can't help it, I
rather work with technologies I am comfortable with: Cairo and Python are it
for me.  I won't imply C++ and OpenGL are not good technologies, just not
something I can easily hack to make my quick-and-dirty visualization
customizations I expect to be doing in the future, mainly as debugging aid.

But I'll be sure to re-evaluate iNSpect periodically in the future, to
provide more feedback.  Additionally, perhaps some parts of my code can be
reused for iNSpect.  For instance, I have code to introspect the NS-3
scenario for available nodes, and uses graphviz to automatically layout
nodes without mobility model.  I also have some code that extracts node
positions every 100ms, and was working on tracing node packet transmissions
during each 100ms slice of simulation, to be fed to a GUI.  This code will
be refined and improved by me, as time permits.  Lately time has not
permitted much, unfortunately, due to the impending NS 3.2 release, but I
should get back to it soon.

[1] http://code.nsnam.org/gjc/ns-3-pyviz/

Take care.

-- 
Gustavo J. A. M. Carneiro
INESC Porto, Telecommunications and Multimedia Unit
"The universe is always one step beyond logic." -- Frank Herbert