Slashdot Mirror
Sun Drops Sawfish for Metacity
Cardhore writes: "According to this article, Sun's and Wipro's developers are now working on Metacity, instead of Sawfish. Metacity and Sawfish are two window managers for the GNOME desktop, and Sun has decided to use Metacity over Sawfish for GNOME 2. This decision has been based on issues such as accessibility, maintainability of the code [1], documentation, multi-head support and a general eagerness from the community to commit to Metacity in the future." Here's a brief description of Garret LeSage's experience with Metacity, which is described here as a "boring window manager for the adult in you." Anyone with Metacity screenshots, please post below :)
Metacity and Sawfish are two window managers for the GNOME desktop
/., where potentially-unfamiliar terms are defined. Time after time I've encountered some unexplained reference in an article and wondered, "Am I the only person who doesn't know what this is?"
Thanks for explaining, and I hope this is the start of a new policy on
Work on e17 certainly hasn't stopped. It's just always slow because they're a bunch of people working in their spare time for free, yadda yadda.
I tested it out about a month ago and it was freakin' incredible. If what I poked around with is any indication, it's going to have the best themeability of the lot (and a great theme-writing program, too!). It has a lot of great things going for it. But it's quite a ways off, I wouldn't expect even a beta this year.
Everybody has the source, and it is apparently quite usable since many people sue it. If someone like Sun additional features or bug fixes, they can make them and publish them. The fact that a single person has moved on to doing something else makes little difference for open source software.
Choosing Metacity may be the right thing for Sun to do anyway, but the departure of even the main developer of Sawfish would not be sufficient reason.
I have used both window managers frequently for about 4 or 5 months now (sawfish2 and metacity that is) and find both of them fast, stable, and great products overall from a user perspective.
.. until then sawfish rulez!
Configurability is easily in favor of sawfish right now but that is only because there is not a gui configurator for metacity currently afaik. However, i knew how to make the modifications I wanted and everything works identically to sawfish so no big worries there.
Port over Crux to metacity and you will have another convert
The BIGGEST factor keeping me from using metacity full time is that the Crux theme has not been ported over to it and I cannot figure out how to make metacity themes (or sawfish themes for that matter) and I really hate the look of the default metacity theme when combined with the Crux gtk and gtk2 themes.
I'm actually surprised that they ever went with Sawfish, since it has all sorts of nifty extras (differently themed windows, for example). From the two screenshots I was able to find of Metacity, it looked like a bland Gnome. Given that Sun was a major purveyor of CDE and olwm, I'm not the least bit surprised that they've switched to a tamer wm. I still think they're missing out, but I guess the philosophy behind the decision is "these machines are made for work, not glitz." Not for me...I use Gnome + E .16 at work....single monitor (makes me wish for my dual-head box at home...) with the same desktop look and feel as my home desktop (see more recent shots).
What is your Slash Rating?
Could programming language ignorance or bigotry be at least partially at the root of this? Probably not, but one wonders anyway.
The fact that a single person has moved on to doing something else makes little difference for open source software.
This can be a problem for niche open source software. Some packages are never developed/driven by more than one person. When that person moves on, it's real easy for the package to drift apart. Sure "anyone" can use the source and built it, fix it, maintain it, and further develop it but usually the further you go from just using it towards further development, the greater the skill required which increases the chances that the package will just get orphaned.
I like the KDE/Win32 style alt-tab window list (small window pops up with all available windows listed; alt-tab selects between them).
Very user friendly and very quick to pop between a large collection of windows. No need to mess up your stacking order plowing through umpteen windows to find the one you're looking for.
Why wasn't such a feature implemented in Sawfish? General unpopularity with the feature? Too similar to Windows?
Does Metacity have a similar window list? Or does it use the annoying Sawfish style?
The world needs one full-assed solution, not 400 half-assed solutions. That's the eternal problem with free software / open source development. Quit bickering about which one is better and which one to use; pick one, stick with it, and get it done.
I absolutely agree with you. I get so discouraged when I run into things like the 90 items listed under "Window Managers" on Freshmeat, and not a one of 'em especially useful.
That's the problem with the current state of open source development. Rather than putting 10,000 brains on one project, you put one brain each on 10,000 projects. Net result: almost zero result for a vast amount of work.
Maybe the only way to get programmers organized is to get a bunch of them in one place and wrap a company around them.
When they say "unmaintainability", this is code-word for "Programmed in Lisp", rather than "Programmed in a sloppy messy spaghetti-like fashion", or "The primary developer is no longer working on it". Most likely, the Wipro programmers don't have much experience with lisp/scheme/rep, and a decision was made to dump it for Metacity, which happens to be written in a language they speak (c, that is).
If you read the metacity source code, at least on earlier releases, Havoc had written things like "I won't implement idea X, because it is crackrock. Tough luck." Things like making metacity play nicely with XMMS. Of course, this was when it was his pet project and not being considered by Sun/Wipro. One wonders if there will be a Sun fork of the project, or if Havoc will turn over development or make compromises that Sun will inevitably require.
While I think metacity is a pretty cool project, Sun's decision is probably one of these management mistakes that have been talked about in all the sociology of software development books. Think of all the little bugs that have been sorted out over the years in Sawfish that will have to be solved again. Things like maintaining focus of window when changing desktops using keybindings; or dual-head setups that have different monitor resolutions while using multiple workspaces and desktops. These things will all have to be sorted out again.
Frankly me and probably 99% of other GNOME users don't give a crap what WM they're running as long as it doesn't get in the way of GNOME. It should be as unobtrusive as possible and limit its features to window-manager-y things.
I suppose E would be a good fit if you didn't want to run GNOME, or could put up with the bloat, or wanted to run kewl gigeresque desktops with metal knobs and shit, but for the rest of who just want to run some GNOME apps, then Sawmill is a perfectly usable and functional WM.
Ultimately I'd like to NOT KNOW what WM I'm running. I don't really care that much as long as it moves windows around and is reasonably skinnable. If Metacity is a move in that direction then that fine by me. The sooner I don't need to know what WM is running the better.
I think there's some missing the point going on here. From Sun's perspective (indeed, from a sysadmin's perspective), the lack of its own setup tools, relying on a command interface to change settings is a plus.
Metacity gives GNOME a chance to address one of its manageability flaws, the confilct between a desktop environment and the window manager. Which controls wallpaper? Screensavers? Why are there separate themes and theme settings interfaces for window chrome and the window contents?
It's because some power users high up in GNOME and window manager development--who usually aren't responsible for any machines beyond their own personal ones--like the flexibility of mixing and matching, and like pushing the bounds of what each component of their system can do. So overlapping--and conflicting--features get built.
This isn't the end of the world, but it does make a GNOME system more unwieldy than it has to be. KDE can run with several window managers, but it comes with one of its own that leaves configuration matters to KDE. GNOME hasn't had this yet. Enlightenment, sawmill and sawfish have been progressively better fits, but Sun and others who are moving to Metacity probably see it as a simpler route to getting a decent (GTK+ 2, anti-aliasing, multihead, accessibility-enabled) window manager seamlessly tied into GNOME than revamping Sawfish--and subsuming all of its configuration into GNOME--would be.
GNOME with Sawfish is a much tougher sell to a simplicity-minded CDE administrator than GNOME with Metacity will be, I suspect.
I recently got tired of sawfish too, so I switched to fluxbox, which is a new fork of blackbox with some nice features. One of its new features is that the user can change the button order! So I have the close button on one side and the minimize and maximize buttons on the other side, as they should be.
The window management in Windows is better than anything I've seen in Linux. I'm sorry, but it's true. I don't care if you can make windows "roll up" into the title bar and you think it looks cool - what problem does that solve that wouldn't be handled better by minimising the window and showing it in the taskbar? Really, I'd be interested if someone could tell me the advantage.
I'd like to see a better way to handle multiple windows, but sadly it seems we are stuck with things that look cool rather than anything useful.
These are the problems that need to be solved, I reckon:
Do you have anything to support that statement? With metacity 2.3.337, switching to a desktop with 14 NEdit windows and 2 xterms takes quite a while, well over 1 second. You can quite easily see the individual NEdit windows being mapped, bottom to top. The workstation is a Dual Xeon with 512MB main memory. A dual fucking Xeon and I have to watch my windows get mapped 1 by 1.
It seems to me people are parrotting the "metacity is fast" line without really checking it out. Probably a groupthink preference for C implementations.
If you are running a Sun server chances are you won't have any of this shit running - well at least if you have a clue. Why use system resources and have services running, and also providing the machine with more ways to be compromised if it is a server?
If I am running a production server, there won't be shit for a GUI on there. Who needs it...
--Jon
I recently on a whim tried out the Oroborus window manager and was pleasantly surprised to find that it is a "boring" wm that does nothing but manage windows, has no menu, icons, pager, or anything. It's also Gnome compliant. It looks really cool by default with a green window border somewhat reminiscent of the qnx gui.
The thing that bugs me about Gnome is that it doesn't have its Very Own window manager. Well actually, it seems like it doesn't have a lot of things of its own, like a file manager, to name one. Everything is someone else's project. Gnome will adopt Metacity, and then, like with Enlightenment and Sawfish before it, the developer will head in some other direction, leaving Gnome in search of a new one.
You've got Gnome with gmc, you've got Gnome with Nautilus. Which one is the real Gnome? Why doesn't the Gnome project unify and maintain its own components? To me it seems that they're really lacking in this area. I like how organized KDE is. The wm and file manager are built as part of the kdebase tarball. All one neat package.
This is not meant to fan any kind of KDE vs. Gnome flames, however. I think Gnome is pretty neat, but I just keep waiting...and waiting...for it to "get there".
You had me at "dicks fuck assholes".
Enlightenment was briefly part of gnome, but the dependencies and politics killed that. At that point E ran on a variety of platforms, and the gnome people of the time didn't have any short term plans to move off x86 hardware and linux. Raster et al more or less had a choice between personally porting the rapidly moving target of the gimp tool kit (gtk) to Solaris etc, or just keeping the window manager seperate. Gnome at the time was sadly dominated by politics over functionality, but thankfully moved on to where it is now. There were actually arguments at the time over whether it should ever be ported to any kind of commercial OS for idealogical reasons. In hindsight, the Enlightenment project was better off without that, and other themed window managers were developed to work with gnome and kde. E v0.16 of course works with both.
E was always about "kewl fx" as well as funtionality anyway - the alternatives were fvwm (not fvwm2) which looked pretty horrible and was time consuming to configure, and windowmaker, which had a few cool features like the dock.
That, I believe, is the long range plan. E at the time was simply a window manager with icons, menus, and a pager. The filemanager etc comes seperately, as whatever one you pick from kde, gnome or myriads of unconnected projects.I'm no expert at this, so i'm going out on a limb here, but i think they only 'stole' the interface.
The underlaying (is that an English word?) multimedia platform (gstreamer) seems to be very original and innovative and something we will probably hear more about in the future.
Sure, it looks like iTunes, but under the hood it's a completely different thing. In my opinion the inteface is just a thin layer on the surface.
It just depends on how you define 'stealing'. It's a blurry discussion. iTunes isn't the first audio player so you could say they stole the idea from earlier audio software.
Every spreadsheet and wordprocessor looks alike. But that's far from saying that they stole from each other or that they are the same.
Furthermore rythmbox is not a finished product so who knows. Maybe it will be skinnable in the future.
just my two cents.