Re: [GSOC] ncat gui idea

[Fyodor <fyodor () insecure org>]

On Sun, Mar 27, 2011 at 01:45:33PM +0300, Shinnok wrote:
> I would like to probe my idea for this GSOC iteration
> for nmap against the nmap-dev list. What I am thinking
> about is a GUI for Ncat,

That sounds like a clever and potentially valuable idea to me.  We
would seriously consider such an application.

> written in native C++ code and Qt.

That part makes it a harder sell.  Nmap already ships with a GUI named
Zenmap which is written with Python and PyGTK.  Remember that
maintaining software is usually more work over the long term than the
initial work of writing it, and using a consistent set of software
infrastructure makes maintenance much easier.  It substantially eases
the cross-platform build process and also reduces the number of
software libraries and programming languages that maintainers must
know.

So I wouldn't rule out SoC funding for a Qt GUI for Ncat, but it would
be a very tough sell.  Of course you are welcome and encouraged to add
Ncat support to your NetcatGUI on your own.  The need to reduce
software dependencies only applies to packages we ship (e.g. with
Nmap).  But the main goal of SoC for us is to help students produce
code that can be integrated with Nmap and used by millions of people
after just a few months of development.

Remember that the hardest part of creating a great Ncat GUI is
designing an efficient, intuitive, well-organized, and powerful
interface.  Writing the code (whether in Qt or PyGTK) will likely take
a minority of the time.

> Another reason that i am sending this e-mail is to probe the nmap's
> team and community need for a new GUI for nmap.

Everyone prefers writing a new program from scratch, but (except when
there is a pressing need requiring a rewrite), we (Nmap project and
Nmap users) normally get the most value from spending the time
improving the existing code rather than trying to reinvent the wheel
from scratch.

Writing a new Nmap GUI would be fun and you're certainly welcome and
encouraged to do that on your own (including mailing test releases to
nmap-dev, etc.)  But we probably wouldn't sponsor such a thing for
SoC.  Again, we're mostly looking for code that stands a high chance
of being integrated into Nmap.  Now if you started an Nmap Qt GUI and
by next year it was competitive with Zenmap in terms of UI and
features, we'd consider a SoC project to enhance it.  But don't
underestimate how hard it would be to rewrite a GUI like Nmap.  We've
had numerous SoC students spend the summer just implementing specific
features for the existing Zenmap.

I hope this mail doesn't sound discouraging!  I'm just trying to guide
you toward the ideas we're most likely to accept.  I'd encourage you
to apply for Nmap SoC in any case.

Cheers,
Fyodor
_______________________________________________
Sent through the nmap-dev mailing list
http://cgi.insecure.org/mailman/listinfo/nmap-dev
Archived at http://seclists.org/nmap-dev/