Re: NS DOCUMENTATION and ANSWERS to questions

[John Heidemann <johnh@ISI.EDU>]

On Wed, 24 Feb 1999 10:07:32 +0100, Thierry Ernst wrote: 
>John Heidemann wrote:
>> 
>> On Wed, 24 Feb 1999 02:45:58 GMT, Milind Kopikar wrote:
>> >Hi there.
>> >       How do I obtain TCP/IP trafic traces through NS?
>> >Regards,
>> >Milind kopikare
>> >IIT Bombay
>> >
>> 
>> Please consult the documentation and example programs.
>
>I do not want "to put oil on the fire" (I am not that aggressiv, 
>but I use to complain a lot - sorry), but most of the time 
>people are posting some interesting questions (especially
>for those in the mailing list who are beginners in NS)
>like this one, they only get a reply like this "please 
>consult the documentation".    I guess they already have;
>if they didn't, they should have read it before posting
>on the mailing list; but should they have read it, the
>NS documentation  is more or less incomplete and doesn't
>reply the questions.   Some methods exist in the code but are
>not documented in the NS doc; some of the methods described 
>in the doc do not exist in he code because those have been
>removed, etc.
>...

There are many areas of ns that need improvement:  the documentation,
the set of functionality, and the level of detail in answers posted to
the mailing list.  We're working on all of these, although it's a
balance that's often not perfect.  The documentation often lags
because we'd rather make the code available to highly motivated users
now than wait for documentation; the mailing list often lags because
we'd rather work on the code and documentation to help lots of people
than handle questions one at a time.  Such is the world of limited
resources.

To respond to part of your comment more directly, I'm less certain
that all posters here have always checked out the available
documentation.  I encourage people to check

    - the "Installation Problems and Bug Fixes"
    - Marc's tutorial
    -  "ns Notes and Documentation"

before posting here.  (All are linked to off of the main ns web page,
<http://www-mash.cs.berkeley.edu/ns/ns.html>).  Doing so will improve
the signal-to-noise ratio and usefulness of this resource for
everyone.

A second suggestion to people asking questions:

    Please try and make your questions as specific as possible,
    and to provide what information you think we'd need to help you.

The overhead of exchanging multiple rounds of e-mail means that
specific questions get much better (more specific and probably more
helpful) answers than vague ones.  If you get a segfault, can you
trace it to a specific function in C++ or line of Tcl code?  If you
have code that doesn't work, can you make it available on a web page
somewhere so we can look at it and not try and guess what you did?

For folks new to NS with general questions about how to get started, I
highly recommend Marc Greis' tutorial.  We also hold occational
tutorials (three so far), watch the mailing list and the main web page
for details.

   -John Heidemann