[Ns-developers] wifi phy API rework

[craigdo@ee.washington.edu]

Hi Mathieu,

Sorry for the delay.  I keep getting interrupted.  This is the fourth time
I've started this email; hopefully I'll finish it this time :-)

> I believe that the API is complete and the implementation 
> reasonable so,
> I would like to move forward with this code. i.e., reviews would be
> welcome, especially from the helper API maintainer.

Comments:

1) There should be an Install method that takes a Ptr<Node> for Python
support purposes, i.e.:

  NetDeviceContainer Install (const WifiPhyHelper &phy, Ptr<Node> node)
const;

2) I think that the semantics of the trace functionality is unclear -- i.e.,
what is being traced.   I also think that the placement of some of the
tracing-related helpers in the phy helper is questionable -- i.e. not really
a phy thing.  I'm sure, though, that any expansion on these comments would
really just be another restatement of the problems I've already described in
bug 241 (http://www.nsnam.org/bugzilla/show_bug.cgi?id=241 -- Trace sources
not clearly documented, trace sinks not clearly documented).

So I think I'll take my concerns about tracing to an upcoming discussion
relating to that bug and leave you free to move forward without that baggage
here.

Other than that, your new code does look considerably easier to use.  Of
course, if you check this in, a lot of tutorial will be rendered obsolete.
Please, please update the tutorial to keep it in sync with the code.  Do a
"grep -i wifi: in doc/tutorial.   I know it's a pain, but it must be done;
especially since this is the first exposure people will have to our
codebase.  It's got to make sense ;-)

Other than that,

<stamp> Looks good to me </stamp>

Regards,

-- Craig