Slashdot Mirror
Interview With Ximian's Nat Friedman
Sheepish writes "OSNews features a long and interesting interview with Nat Friedman, of Ximian fame. Nat tells all and talks about the upcoming Ximian Desktop 2 and its differences from Gnome 2, the difficulties of developing the MS Exchange Connector, Linux as a desktop, Mono and plans for Gnome integration, the hundrends of OpenOffice.org changes made to make OOo like a Gnome2 app, and how Ximian feels... about Apple's business. Four screenshots of Ximian Desktop 2 are included too."
Quote: "- Uses MSFT file formats by default, reflecting the reality of most of the documents you will receive. No longer tells you you're about to lose all your data when you save in an MSFT format. "
I'm pleased that XD2 is striving for complete UI consistency. This is something I've always felt was lacking in the overall user experience for linux. Having a coherent set of human interface guidelines ala Apple's materials for OS X can do nothing but help.
I was left salivating after viewing the screenshots and reading about the far reaching changes that were made, especially integrating OpenOffice with the rest of the desktop.
And they're looking into migrating several hundred thousand desktops, especially in Europe. Damn!
Goodbye Bluecurve, Hello Ximian Desktop!
Newsfollow.com
Well I gotta be honest and say.. from the shots I have seen, and from what I have read, I can't really see what the Ximian Desktop offers Red Hat users over the superb BlueCurve front end on the most recent versions.
Antialiasing, clean & well organised style, custom icons, and specially developed management tools. I really really rate what Red Hat have done, and I could never see myself paying for something like Ximian Desktop to replace BlueCurve.
"Hey! Unless this is a nude love-in, get the hell off my property!!"
It's kinda strange... OSS with it's release-early-release-often idea almost makes it seem like improvements come so slowly, because they flow in a discrete trickle rather than the major leaps that come much further apart (emphasis on "seem")... Ximian's been working behind a black curtain for so long, it makes XD2 seem like such a gargantuan improvement...
Though significantly delayed, XD2 was released when Ximian got everything right... and they have... finally I have a desktop environment that I can proud to show to my consulting customers as a viable option...
-jag
http://starboard.flowtheory.net/
Hopefully the week of the June 9, they will also release the complete source code to everything they use to build XD2.
I have 1 Gentoo system at home and 2 Built from scratch machines, and it would be soooo cool to have a ebuild for Gentoo, or at least a Garnome type build script. Especially for their OpenOffice.org version.
After reading this interview I really Can't wait.... Even tried their ftp site, but the XD2 directory is not browseable by an Anonymous Coward!
From their desktop support general info:
/etc/redhat-release
Slackware Linux is a well-respected Linux distribution, and has a dedicated, fierce following. It is possible that Ximian may support Slack in the future but we have no idea when that might happen. Slackware support is likely to come after BSD support, Debian PPC support, and SuSE PPC support. Right now, we have plenty of work supporting the distributions we already support.
The things that prevent Ximian from supporting Slackware are partly technical, and partly market based. Technically, Slack has a package management system which has substantial differences from other distribution's package management systems. Dependency checking, for example, is absolutely necessary for certain Ximian services and features (the installer and the updater, in particular), and is not fully supported by Slackware. Slackware's architects have a well-defended disdain for dependency checking, and we can understand their arguments. But without it, Ximian Desktop can't figure out what to install, what to upgrade, and what to leave alone.
That means, basically, that it's a lot more work for us to add really good support for Slack than it is for us to add good support for, say, Conectiva, which is based closely upon the Red-Hat model. Not only that, but there aren't a lot of distros based upon Slack. From our support for Red Hat, it's a quick jump to other rpm-based distros. If we support Slackware, it's working with an entirely new package system just for one Linux distro.
Another market force is the profile of the typical Slack user. Slackware users often compile stuff themselves. They know how to install software at the command line. They know their dependency trees themselves, and don't trust or need package management systems. They're hackers in the best sense of the word, and we respect them deeply for that. They don't need things like the Ximian Desktop update service, or the graphical installer. Ximian is about making free software easier to use, and Slackware users don't tend to need any help.
So, what can you do, elite Slackware user, ignored by market forces and business types, if you want the prettiest, bestest desktop in the Linux land? You can download pre-rolled tgzs from the variety of Slack software mirrors, or get the binary rpms or source rpms from the Red Hat directory at our ftp site, and install by hand with rpm. Or you can convert them to slack packages with rpm2tgz. And, in a brave trick of hackery, you can fool the graphical installer into thinking you're a Red Hat user. The command:
echo "Red Hat Linux release 7.2 (Enigma)" >
has been reported to make the installer work, although you're likely to have difficulty with one or another dependency somewhere. Official Ximian support of this method is not available, and we cannot give you any guarantees.
Feh, my Atari 2600 walks all over that Pentium 4 AND your Windows 2000 - combined!
I recall back when Ximian first started to come out with some slick looking stuff they were much nicer, asthetically speaking, than any linux distro out there. With Bluecurve and the maturation of Gnome 2.xx it seems the need for Ximaina is greatly diminished.
By the looks of things here I see no need to upgrade from RedHat 9.0 with the exception of getting Evolution 1.4. (And actually if it's faster than the butt slow 1.2 version that would be a good upgrade, now that I think about it.)
Quote: KDE has way more options (the clock properties dialog has five tabs!),
Actually it has 6 in KDE version 3.1: General, Timezones, Plain Clock, Analogue Clock, Fuzzy Clock.
For some reason I find that amusing. If you're going to drop some FUD, at least get your facts straight.
I'm guessing 3.2 will have 12 or 24 depending on it's mode.
Just wondering, but has Ximian made a KDE version of their new industrial theme?
Ya see, the beauty of the KDE/Gnome thing is that some KDE apps you can't live without, and some Gnome/GTK apps you can't live without. Gaim and K3b/Kmail spring to mind right off the bat.
I like KDE themes like QTCurve and Keramik/Geramik because it makes the GTK/Gnome/KDE applications look the same. If using this Ximian desktop means that my KDE apps will look out of place, then it doesn't really seem that appealing.
A theme like this seems like it'd be simple to do, so I'd be very curious to see if Ximian has really done a complete job of it.
Sure, even older.
Ximian Desktop 2 is not a complete distro. It's software for a variety of operating systems.
It is (or will be, upon release) available for download free-of-charge. Source is/will be available for all open/free components. Patches are being and will be submitted upstream to maintainers.
Purchasers ($99) get extras including 3rd party (proprietary) software, PLUS 30 days support, PLUS a year's Red Carpet Express high-speed updates.
a.
Some interesting claims made for Mono:
/. a number of times, ActiveX, CORBA, DCE etc. have all made claims like this and have met with limited success.
.NET: http://www.zope.org/Members/Brian/PythonNet/FAQ.ht ml.
.NET, managed code cannot generally use classes implemented in Python.
.NET would certainly be doable and useful - but it would also be a lot more work than the current approach."
1. Mono can be the universal component hub, allowing you to use C objects from Python, C++ objects from Perl, and so on.
We've certainly been here before. As has been pointed out on
First there is the inefficiency introduced by constantly translating data (where equivalents exist at all), second the impedance mismatch of languages with quite different call models.
Yes, there's some capability here for scripting code written in low-level languages, but that's quite a different thing from claiming to provide universal, peer-level interoperability.
Note that this isn't the same argument that says that bytecode level interworking is doomed - one is still limited to a rather C#-like subset of features, just as one is to a Java-like subset in a JVM.
Nat goes on to give an example of how Mono is changing things:
This is possible because C#'s language features make it trivial to automatically bind C# objects into other languages. Check out Python Scripting for
OK, let's see what Brian thinks this new Python Dotnet is bringing to the table:
" While a solution like Jython provides "two-way" interoperability, this package only provides "one-way" integration. Meaning, while Python can use types and services implemented in
A Jython-like solution for
Hardly a ringing endorsement of Mono here. Perhaps the last reference will be the proposition that we can't refuse?
Nat says:
There's also a Mono-based JavaScript compiler in the works (MS already has one, of course).
Doesn't the Java world have one of those too? Yes, in fact, it's had one for five years. Rhino is a full Javascript compiler, interpreter and debugger, released by Netscape in April 98 and still developed under the Mozilla banner. Not some also-ran knock-off here, but something used in quite significant products such as the Resin web app server.
So, draw your own conclusions about what real new capabilities Mono will bring to the OSS world.
And don't forget that there is at least one company that will definitely gain from this all this free marketing and "innovation".
We don't have a fundamentally new file selection dialog, but we added some quickbuttons to the stock Gtk one that jump you to your desktop, documents or home directory. This makes it a bit easier to use.
Owen Taylor is allegedly developing a new file manager in Gtk 2.4 that should be much easier to use, and that we expect to see adopted across GNOME very quickly.
You are jealous that your sister is looking at other guys?
Creepy +1!
Aside from multiple distribution support, is anyone taken aback about how many companies are essentially peddeling what an admin can do with apt-get/cfengine (with updates to cfengine configs via rsync/ssh) with Debian?
:) Since I follow woody and roll in some of my own updates alongside other users updates, its quite easy to have a 'modern' gnome2 system that has been updated against major security issues.
I too support hundreds of machines, and I find my worst experience is making sure i've got a decent, up to date for bleeding edge kernel handy and a discover database to match it. Nevermind X.
Having a nice automatic installer (autoinstall, heavily hacked, ask for source if you care) and good remote mass administration tools are the two things that make my life easier.
Be weary of supporting these companies, I just don't think they have many peoples best interests in mind if you have a clue handy. Ximian is supporting propretiary file formats (doc!) now, redhat is selling 2 year development cycles (wasn't that a debian complaint a ways back?), and many of them are only selling their 64-bit installers for nearly $1k a pop.
-- dieman - Scott Dier
[rant]
Sure, trade RPM dependency hell (which is really bullshit if you use a modern distribution) for the apple monopoly / shareware hell. Right. With macs and osx, you are forced to either shell out $30 to $100 to do _ANYTHING_ remotely useful, like encoding video, burning DVDs, or backing up your files, or pirate the abovementioned software. Sure, you can use free software, but then you have to mess with porting it and compiling it for PPC and OSX -- a major pain in the ass. That's pretty much the reason why I dumped Windows -- it's not stability or security. I'd say that XP is about as stable as OSX. Both are less stable than my Linux box.
Also, the simple solution to your Linux problems would be to either use packages compiled for your distribution (which is rather simple with URPMI) or to download and compile the source or source RPMs. I don't think you've used linux "for years". More like a week. Anyone who used Linux even for a month would know that packages built for Suse won't work well on Mandrake, which is probably what you were trying to do.
Besides, I would much rather use windows than go for vendor lock-in with apple. I thought people had enough of that with proprietary unix boxes. I have a severe problem with having to buy all my hardware and most of the software from one overpriced company that also actively prosecutes anyone selling compatible hardware. I don't know what planet Apple is living on, but a 1GHz machine with a small hard drive, outdated video card and hardly any RAM should not cost $1500 in this day and age.
[/rant]