Slashdot Mirror
Havoc Pennington on GNOME 3's Future
An anonymous reader writes "Havoc Pennington, lead developer of GNOME, wants to fork GNOME 3.
'So the forces of existing userbase, the easiest-to-reach future userbase, cross-platform applications, and funded development efforts are strongly pulling GNOME 2 toward conservatism. I think GNOME 3 should be a fork for that reason.'" This has been a common practice for not only many open source projects, but proprietary systems such as Solaris for major revisions, so it's not as tumultous a change as the word "fork" may imply.
Darn those pesky users for making us stablize things instead of hacking cool new features! I mean, which would you rather have, a foot menu that works or spatial Nautilus?
Javascript + Nintendo DSi = DSiCade
The name 'Gnome 3' is reserved for the core Gnome product.
If you're going to fork the core product and possibly make an incompatable branch, please give it another name.
94% of Repubs and 21% of Dems voted to renew the Patriot Act
If they want to fork, let them. If it becomes any good, it'll be used
Show this to your friends and family that don't know what a real hacker is
Why doesn't he just join up with KDE? Theres enough bulk in KDE that it should fit his needs nicely.
There is no fork.
- None can love freedom heartily, but good men; the rest love not freedom, but license. -- John Milton
That has got to be one of the coolest first names ever.
oh wait, there is no spoon
Do not try to read the dupe, thats impossible. Instead, only try to realize the truth
What truth?
There is no dupe
Havoc is NOT talking about breaking out of GNOME because he doesn't like the current way.
He is talking about forking off development for GNOME 3, because it would be too disruptive to move everyone onto GNOME 3 immediatly.
Basically GNOME 2 would continue as is, with incremental changes, while someone starts hacking on GNOME 3 for a future release. They would diverge quite heavily after a while, but when GNOME 3 has started getting momentum, GNOME 2 can be closed down.
They should do that! At least, if they are not able to have better functionnalities than KDE, they can claim they have a bigger revision number! :p
As a 90% user of Windows (ya I hate it) and 10% user of Linux (I hate it too), I think the biggest problem with linux is its incessant forking. At least Windows develops some sort of standard and things look about the same on all computers. It's like Windows is monogomous.. It's boring, but reliable.
With linux, every distro breeds more distros.. every project breeds more projects.. They're forking like rabbits!
--- We need more Ron Paul!
This is a great strategy on his part. I view this as analogous to the great gcc2->egcs->gcc3 "fork", which was quite successful.
11*43+456^2
"Let's not make any configuration options available to the users at all! No settings, preiod. If they think they are smart enough to reconfigure the product, let them read through the source and figure out what cryptic Gconf keys they need to hack. Yeah!" Seriously though, is forking such a good idea? I can't say I've run into too many gnome bugs (and I use it everywhere.) What gnome really needs, like a number of large-scale open-source products, is to have all the features 'finalized.' It seems that some things are just not quite finished, or some things hint at integration, yet it is not as complete as it should be. Finish the main features, then fork off that. Why not?
Just because Havoc says something doesn't mean it's true. Havoc doesn't own the Gnome project, and doesn't have the authority to make a big, sweeping change like this.
In the past, sometimes his plans for Gnome have been in conflict with other members of the team.
If Havoc wants to fork the project, fine. But don't call it 'Gnome 3' unless it has been designated the 'Gnome 3' project by the board.
Now, if this was a message from the Gnome Board of Directors, I would feel differently.
94% of Repubs and 21% of Dems voted to renew the Patriot Act
I hope Novell gets involved and has a Mono core dependancy. RedHat won't like it, but they're not the only game in town now that Novell bought Ximian.
.NET libraries that will be written now and in the future.
Mono has the benefits of being able to run Python, C#, Java, C, C++, VB, and a whole slew of other languages that the JVM is incapable of. Not only do you get the benefit of automagic bindings to various libraries, but you get tons of
You still write core parts in C, but more and more in managed code. You'll need a beefier machine, but time marches on. Was 640K enough for everyone?
As far as the legalities of Mono are concerned, I'll leave that to the FUDsters who are better at cowering under the covers instead of embracing good technology.
I know this will sound rude. But I feel like saying it anyway. Gnome has very much been focused on becoming more userfriendly in Gnome2 and it has done this by a less-is-more approach. This has, for me, made it a lot more user-unfriendly. The simple file dialog boxes are a very good example of what I mean: They now by default open up half-opened so users will not be confused by the more advanced options in them. But the problem for me is that the advanced options are things I use every time, meaning an extra click or keyboard press every time I need to use them. There is no good reason for them to appear half-open, it is just done to make it simple. The result of this is only extra time spent using them every time to make it easier to use the first time for complete idiots. Something similar is also done with the features to make it more user-friendly: If a feature is to advanced for a beginner, they are simply removed or placed where they are completely unavailable or require a great deal of effort to use. Gnome2 has come user-friendly to the extend where it is almost impossible to use productive on a day to day basis. I seriously hope Gnome3 will be better. Not that I think I will ever use it as a main desktop again, but as I use a lot of Gnome2/GTK2 programs (like I also use KDE programs in my fluxbox) this annoy me very much.
9/11: Never forget it was a false-flag operation
It seems that the browser-now-known-as-firefox did this. They forked mozilla, created something quite a bit better, then worked out the kinks, and eventually it'll merge back and become mozilla.
I have no problem with gnome forking and creating a cutting edge version that's unstable for a year or two. It might reinvigorate the project so they can make something that gives kde a run for it's money.
Pennington isn't proposing anything. He's merely examining the current discussions on the future of Gnome and exploring possible options. From TFA:
where the comment ends and sig begins
First of all, some xcompmgr support would be nice. Gnome has a few BIG problems with using that program, which is unfortunate because using it on my computer speeds up the sluggish Gnome.
Another thing would be better wireless support. Unlike KDE, there is no app that can do what Kismet can. The network app. lets you connect wirelessly, but no part Gnome lets you scan. In this department many good programs have appeared that would fix this problem. I like- Wifi Radarand this applet
They only need to be incorporated (or packaged with a Gnome distro for the love of diety).
Many people think that Gnome's biggest problem is RAM usage, and they might be right. 256mb feels VERY different than 512mb on the same machine. I personally believe that this problem was made worse in the last release, not made better. I think that 2.12 has intentions on fixing this, so I care more about Gnome 2's interface problems.
Open Source Sushi
I seriously am envious of anyone able to code anything decent. I have developer envy.
In any case, I was wondering some of the same things that you put forth in the blog about fractally "petrifying" the GNOME codebase. It was my first DE i'd used with linux (rh5.2, I think CDE [ugh] was default) and loved it; but it never seemed to add features.
KDE 3.2 grabbed me after a few years of just using black/fluxbox and seems to really have the upper-hand in "creating value" when upgrading from previous versions. Do you feel GNOME has been losing ground for this reason? I'm looking forward to trying Beagle and seeing where Seth's OpenGL hacks lead to for metacity, but outside of that, I can't think of why I'd want to run GNOME anymore. Can you try to give me some insight into why I should be looking forward to future GNOME2 versions, let alone GNOME3?
put the what in the where?
Even though I depend on GNOME libraries for my projects (specifically PyGTK), I think this is a good thing.
The reason why is that having a bleeding-edge version that integrates things like Cairo, xcompmgr, more eye candy, etc will give us who like to have a system with all the eye candy a chance, without having to worry about adding them to GNOME 2.x and possibly disrupting users who want a no-frills desktop. When GNOME 3 becomes stable, it can replace the old version.
But moreover, the Linux desktop is at an inflection point - we're just starting to get the kind of nifty eye candy that other desktops have. GNOME 3 should be a chance to get GNOME ready for the future of the Linux desktop - using Cairo to render the GTK widgets, using Luminosity as the next GNOME window manager, etc.
Sometimes it's healthy to fork off your code and rethink some of the assumptions you made rather than having to deal with the cascading problems that can crop up when you try to muck about and fix those messy hacks we all seem to create.
Forking isn't always bad - sometimes it's necessary to eliminate cruft. If the end result is a better desktop, then that's what should be done.
Seems to me that Havoc wants to create a platform to try out new features that may or may not be accepted by users, and another platform that is more "traditional". When a feature is found to be really useful then they could merge it into their traditional version.
The problem with this, of course, is when the underlying libraries like atk, etc, are altered fundamentally. In that case, things will become a right mess.
XML is like violence. If it doesn't solve the problem, use more.
Yes, lets all fragment our efforts to kingdom come, then try to compete with proprietary software.
...and then lets wonder why Linux isn't taking off on the desktop.
Lets have 3000 different distros, and no clear leader. Lets make sure every distro has slightly different configuration tools. While we're at it lets force users to compile their source, (including the kernel otherwise their hardware won't work).
Lets have 3 different kernel firewalls, in about as many years.
Lets have 300 desktop managers, none of which quite work or interoperate.
Lets have 3 different office suites, none of which quite translate MS Office stuff quite right.
I'm getting goddamn sick of this, and I'm a developer. I'm also damn tired of defending it. I've had comp sci students roll their eyes at me when I had to recompile my kernel to add support for a printer so we could print data off in Linux. I've also had Astronomy Masters students feel overwhelmed with Linux - avoiding it or dumping it out of frustration early.
Lets decide whether we're doing cool techy geeky play stuff, or whether we want to produce something real and tangible and useable by everyone. Lets make up our minds on any given project what our goals are (or what the goals for our group are). Lets contribute to existing open source instead of starting our own little pet project that does no better than anything that came before it. Lets get a bit of unity back into open source, before it goes the way of the dinosaur!!!
These posts express my own personal views, not those of my employer
Back in the "old days," I remember the pgcc/gcc split. The old version of gcc was in dire need of an update, but was relied upon by many projects and users. The pgcc effort came online and made remarkable improvements. The old gcc and pgcc coexisted for years. People wanting the fastest compiled code used pgcc. Them finally pgcc was deemed stable enough and became gcc.
Revolutionary work can be done in a fork and I surely wouldn't discourage it. It will make distributions a little more complicated and may cause compatibility issues, but there is a clear benefit here. If the whizbang new stuff is worthwhile, people will use it, patch the bugs, solve the compatibility problems, and use it.
(Indeed, it's still not that great, or you'd be seeing a lot more i786, -p3 and -p4 RPMs out there. Not many people use an actual i386 these days, except in the space industry.)
PGCC's optimizations were, IIRC, largely rejected as EGCS's working group didn't like Intel's way of doing things. For similar reasons, again IIRC, a lot of the approaches used in Intel's C compiler aren't used in EGCS/GCC either. Well, I can understand that. It's better to use a good design, if you can. On the other hand, if you're writing something for a processor, it would seem to make sense to allow for the faults of that processor.
I seem to remember a similar argument, when SGI issued a whole load of speed-ups for Apache 1.x - the Apache group rejected them because they were not the way the Apache group wanted Apache to work. SGI caved in, eventually, and stopped maintaining the patches. A royal pain, because I quite liked them.
It's a small world and it smells funny; I'd buy another if it wasn't for the money; Take back what I paid (SoM)
And at your foot?
ROMANES EUNT DOMUS
Let's add Evolution to the mix for kicks...that makes GNOME/OO.org/Mozilla/Evolution...whose acronym is GNOME!
Obviously this is a sign from the Gods Of Recursive Elegance (GORE? He's in on it too?) that we're on the right track.
-- Microsoft is the most expensive commodity operating system and office suite vendor in the marketplace.
Not all distros do. Pick an old-school, stable distro. Don't go with some flavor-of-the-month. Try something like Slackware, or shit if you want stable, Debian stable is rock solid. Linux gives you the freedom of choice to pick the right distro for you.
It has: .desktop extensions (like Windows does with .lnk files, but without the arrow telling you it's a shortcut) allowing them to spoof regular documents with icons and everything.
* No menu editor.
* Hard coded un-overridable mime-sniffing that gets lots of things wrong (because it's foolish to even try to anticipate every single file format and code to handle them all) and then forces its will on the user (won't open some of my text files in gedit for "security" reasons).
* A file browser that defeats all that paranoid mime-sniffing "security" by hiding extensions
* Menus that scroll like win95 when very full. A menu editor and/or overflowing into columns would help a lot.
* And a continually decreasing level of configurability.
I suppose aside from that it's very good. It's the desktop environment I'm using now, and the one that I keep coming back too after repeatedly trying to dump it in favor of the alternatives.
There are many good things worth keeping in GNOME, and many worth changing. Some are foundational issues, and the best way to handle those foundational issues (such as getting Storage implemented and suchlike) is within a fork. I love GNOME (and use KDE under GNOME, rather than the other way around) and wouldn't like to see the GNOME 2 line disturbed too much in the name of progress, yet I wish to see that progress happen.
On a related topic, I'm not up to speed with the details of programming GNOME: in which order should I learn my way round the libraries?
John_Chalisque