[forum] About our effort at NoMachine
Sun, 23 Mar 2003 00:02:54 +0100
We spent last three years trying to make X so good as a network
computing platform to compete in performances and functionalities
with the leading proprietary technologies from Microsoft, Citrix and
Tarantella. NX was made available few weeks ago. You can find
all the modifications we did to the XFree86 code base as well all
the additional components (for example the X compression libraries)
at http://www.nomachine.com/dev_sources.php. Not only NX beats
those products in performances, but aims to make X-Window the
standard way to deploy applications to users of any OS.
This is not the place to advertise our software, but just let me
enumerate some facts:
- NX provides X, RDP and VNC desktops trasforming
foreign protocols to X protocol.
- NX is a complete X distribution for MS Windows
- NX compresses X protocol (and foreign protocols)
to a degree that makes possible to run complete
desktop sessions on a remote server across the
Internet, even through a modem link. A TestDrive
server is available to the public. You can run a KDE
or GNOME session from there, just to try.
- NX makes any Linux workstation a Terminal Server,
in the MS sense.
- NX core software is based on XFree86, OpenSSH,
RDesktop, VNC and other OpenSource components.
All X related stuff developed by NoMachine has
been released under GPL.
- NX higher level components are commercial software
but any company or good developer could implement
an OpenNX project in a few months, if not weeks.
We know we made a very good job and want to preserve
a competitive advantage, but it's our interest to
have competition. We want to push X and Linux as
a network computing platform. If X wins, we win.
NX embraces many aspects of X technology, from X protocol to
libraries and session management. Unfortunately documentation is
still poor and mostly oriented to end-users, not developers. We
are working hard these days to fix the (many?) bugs that are
Too many people are speaking about leaving network transparency
behind. Such people confuse poor implementation of X clients and
X toolkits with X flaws. I'm leading the NX project and implemented
the X compression software (maybe some of you remember mlview), so
I probably had time to explore the problem space widely enough.
I can tell you about the default terminal application included in
one of the two leading Linux desktop environments that's generating
so much X traffic to fill a 100Mbps LAN just to scroll a text
window, or can tell about a famous Linux distribution that included
a small applet to remember you to update your software that is
sending PutImage requests for gigabytes. I think that one of the
problems of X is that it's just too fast!
In our opinion, reason why people tend to forget or ignore X
networking is because of the gap between "the possibility" of network
transparency and the "availability" for the average user. NX was born
to fill this gap. We believe that once made simple to run X
applications everywhere in the world, poor clients will just disappear
due to lack of users.
During these years I've been partecipating to many discussions
about the future of X development. Noboby seemed to be interested
to aspects like X compression technology or X session management.
These aspects were considered marginal in respect to graphic cards'
support, font antialiasing and speed of rendering. I even remember
the lead developer of a very good window manager to say that our
effort was useless as X would never be able to offer remote
computing the way RDP and VNC are doing. Well, considering we are
in speech with a leading game console manufacturer and with a 3G
network provider for inclusion of X/NX in next generation's mobile
phones, I'm happy that he was wrong.
Gian Filippo Pinzari, email@example.com