Faulty Paradigms for Business Week

Posted Oct 25, 2005 9:35 UTC (Tue) by drag (subscriber, #31333) [Link]

Redhat is to Servers what Novell tries to be towards desktops.

There are a few reasons why Gnome is as good as it is today. Lots comes from Sun and their usability studies from 2001 that Gnome devs took to heart.. but another large push, ironicly, comes from Novell.

Look at the work that Novell has put into the Gnome desktop from their purchase of Ximian. For instance on Evolution file manager they open sourced the Exchange compatability plugin, they improved it considurably and it is now the default e-mail client for gnome.

They also hired cairo and glitz hackers, which is paying off by vector graphic support in the latest gnome gtk toolkit and eventually opengl acceleration thru glitz in up and coming releases. A large part of that is due specificly to Novell's paid developers. (they'd do it without novell, but it's nice to have the money and resources)

They bought Suse and open sourced Yast, which people have been waiting on for a long time.

They developed mono, which leads to great developments with C# applications on Gnome such as F-spot, Tomboy, and Beagle. Lots of that got started from help from Novell.

Previously Gnome desktop was a C-only affair. Now we have fast C bindings and toolkits for a couple of rapid-application-developement style languages in the form of Python and C#. (of course python stuff doesn't have anything to do with Novell, just commenting on the progress)

Also from Novell we have Hula, which is going to provide open source implimentations for the 2 most usefull aspects of MS exchange: email and calendering. Not to mention protocol support for calendering thru a veriaty of email clients including evolution.

And most recently they launched betterdesktop.org and seeded it will actual videos of users on desktops. Invaluable stuff to desktop application and UI developers.

They have actually done Windows to Linux desktop transitions, studied the problems, documented the pitfalls. They have scripts, documentation, and support software that clients can use to circumvent the issues.

All sorts of stuff like that.

They have groupware that is good that supports linux, even if it's currently closed source. They have directory services. All this works for Linux right now, were as open source competators to Exchange are still lacking. Also they have a loyal following that beleive in them/trust them thru the long time netware user crowd. They have support systems around the world, they have many existing contracts with numerous governmental/educational/industrial orginizations around the world.

They have all these advantages. Trouble is that they can't compete with Microsoft enteprise software when the desktop platform is Windows. How the hell do you have a chance to win when your competator owns and designs the only viable desktop platform?

Don't get me wrong. Redhat is great. There servers are great, the software is great. Thru IBM, SGI, and others, as well as numerious kernel developments of their own they made Linux 2.6 ready for enterprise markets and increased the scalability of Linux's performance immensely. They gave us GFS, they gave us big improvements in ext3 over the years, they've offer good enterprise support and a stable long term platform for businesses to use for 5-7 years without distruptive upgrades. They've have helped out Gnome. They've open sourced a top-line LDAP directory server as well as management tools, netscape directory services. (now calling it redhat directory and fedora directory services).

If you want to build a high-aviability highly-scalable distributed linux file server cluster using a combination of open source technologies such as Lustre and GFS and combine it with support for Oracle databases and Veritas storage technology... You go to Redhat. I sincerly doubt that Novell can realy deliver like Redhat/HP/IBM can in the server.

But personally I've gotten the impression that Redhat engineers haven't touched a 'real' enterprise desktop that people actually use since Windows 98. Bluecurve was a interesting thing, but in a crippled puppy sort of way.

But if you want a Linux desktop roll-out to replace a large number of Windows clients... I'd take a long hard look at Novell and Suse.

The way things stand is that Novell will either make Linux on the desktop a marketable item for big business, or they will die trying. It maybe to late for them, maybe they hung onto netware to long, but I hope they make it. And if they can't make Linux on the desktop a profitable item, I doubt anybody can.

Posted Oct 25, 2005 9:39 UTC (Tue) by nix (subscriber, #2304) [Link]

linux and linux-related system *development* is increasingly becoming a paid-affair

Excellent. This means that people who aren't developers are getting the chance to improve the system in ways they find useful (if only via paying Red Hat to do it, or via whatever mysterious market-research consultancy osmosis IBM uses to find out what features users want). This is a large part of the point of the `free' in `freedom': even if you don't program, you can pay someone to do it for you.

This does no harm to those of us who are free software developers either, no matter how lunatic the desires of the bunch paying the bills might be. Think about it: even if a billion programmers materialized out of empty air and started hacking on Linux for pay, this would not make the existing people who hack on it for fun vanish, nor would it stop them forking projects that the billion programmers moved in some direction the free hackers considered undesirable. All it would do would be to attract some of that billion to free software hacking, and that cannot but be a good thing.

(Yes, I know my numbers are ridiculous. That's the point: even if far more programmers than there are on Earth started hacking for pay, it wouldn't diminish the contributions of those who hack for the sheer joy of it one whit.)