Submitted by The PC-BSD project is developing its own desktop environment from scratch! The ultimate plan is for Lumina to become a full-featured, open-source desktop environment that may ultimately replace KDE as its default desktop environment.
A Phoronix reader, Ryan Bram, wrote in to share word on this new desktop environment being developed by the PC-BSD crew, the popular desktop-focused derivative of FreeBSD. This new desktop is called Lumina and is being developed as a home-grown desktop environment catered toward this BSD operating system.
While it’s obviously cool, I wonder if it’s a wise idea to undertake such a huge endeavour. I honestly doubt PC-BSD has the developers, testers, and users required for creating, maintaining, and improving an entire desktop environment.
I hate to be “that guy”, however, it seems a bit silly to develop a whole new desktop environment at this time when there are more pressing matters, especially since the choices of DE’s these days isn’t the problem.
If they need some guidance, just a quick list:
1) We don’t have anything like Adobe creative Suite.
2) All DE’s seem to be missing some basic but common GUI based Setting (Yast comes close though, but OpenSUSE, has some other issues).
3) There seems to be serious performance issues with most DE’s, distros and gaming (Gnome 3 is the only modern DE I could find at the moment, which works at native frame-rate speeds)
4) Even Bitwig on Windows took 2 clicks to get going, but, on Linux, getting realtime audio running is a WHOLE other story.. Last I tried on BSD, my audio didn’t work at all (but it might be different with my new sound interface now).
What’s holding back Linux or BSD isn’t the desktop environment (I haven’t heard anyone complain about the interface). What’s holding us back is the small stuff which we are missing, the small pieces lacking integration and professional applications (for audio, we now have bitwig, but we don’t have anything for design).
If even 1 distro, instead of working on a new DE worked on a free, well-integrated Adobe Creative Suite alternative, it would do FAR more for their cash influx/donations, and for the global *nix environment than saturating the environment with another DE.
Open source projects are not like a company where you can direct resources to anything you want. Whoever the authors are of this new DE, there is no way you would ever be able to tell them to write an Adobe CS competitor instead.
That said, why not Razor-Qt? http://razor-qt.org
Lightweight, Qt based, and already a lot of the groundwork is in place.
You seems to confuse “Open Source” with “community developed”, there are a lot of Open Source projects where the key, or even the most, developers are hired by one or more companies.
Of course a company can hire people to work on a FOSS project and tell them what to work on.
Razor is merging into lxde as they’re dropping gtk.
Razor-qt is being merged with LXDE: http://wiki.lxde.org/en/Merge_LXDE-Qt_and_Razor-Qt
Edited 2014-04-25 12:20 UTC
Umm, that’s exactly what you did, in two out of four talking points:
Anyway, I’m not saying you’re wrong, especially on your third point; I find OpenBox to be the most usable WM by far, and it’s one of the most basic. Even on my beast of a workstation, Gnome and KDE are slow, and the latter is just way too complicated for my taste. But that’s all personal issues.
Objectively speaking (as much as is possible on a subject like this) I think it makes sense for a BSD provider to roll their own DE that isn’t a GNU/Linux port. A native DE that is licensed under BSD rather than GNU GPL, and developed by a BSD team, would be a better fit from both technical and philosophical standpoints.
What I meant was the interface layout, the actual Design, since from a user perspective, that would be the main reason why you would start a new project.
We have a history of constantly making major changes to the existing GUI’s, but then major options go missing in the major updates (which is why its never 100%).
Open source developers can choose to do what they want, but, I’d be hesitant to donate to PC-BSD, if they don’t choose to instead improve existing solutions. This is simply more fragmentation, which will be a subset of existing solutions (rather than fixing the remaining problems with the existing ones)…
Edited 2014-04-25 03:31 UTC
I think it’s perfectly reasonable for a BSD team to develop a BSD project instead of borrowing GNU projects. Whether it’s the WM/DE or another component, I think BSD in general shouldn’t be so dependent on GNU projects. I’m happy that the two ecosystems do get along so well, and can share resources, but there’s absolutely nothing wrong with “rolling your own” if you feel it’s a logical alternative.
But that’s just my take, and not a very popular one among GNU fans, as I’ve found over the years.
If you’ve never tried to port Linux-centric code over the BSD, then you have no idea how incredibly ridiculous the process is. It’s by far much easier to develop a solution (from scratch) that’s tailored for the BSD’s, than it is to port something from Linux & try to force it to fit. The best thing about it is that it’d only have to be developed once for the BSDs. It’s easy to port system software between the BSDs. Why should they be forced to stick with Linux solutions? The BSD’s had their community before Linux even existed; so, it’s about time the BSDs go back to developing solutions for themselves. This little sharing experiment has run it’s course. I think a lot of people have forgotten that Linux & the BSDs are no more tied to each other than AIX & Solaris are -the BSDs would survive just fine without sharing with Linux. I wonder what a BSD solution to replace the whole X stack would look like…
It is also sometimes(not always) true to any software development. In this case, writing from scratch is much easier rather than using existing DEs tailored for Linux.
Funny how as the years have gone by, the “we don’t have” has turned ever more esoteric.
I think you are spot on! The desktops are great, the integration is lacking. I can get 90% of my work done with a text editor (plenty of them out there) and a browser, but to make my work sing, I needed support for that last 10%.
Though whis undertaking seems to be obvious stupidity (KDE/GNOME/XFCE are out there, requires resources, etc), from BSD perspective it is not os obvious:
1. KDE, GNOME and XFCE are increasingly Linux-only. Among BSDs, only OpenBSD succeeded porting GNOME 3 (FreeBSD is close, but not there yet). And this port is done by exmployees of M:Tier, who provides commercial support for OpenBSD, including commercial desktop deployment. KDE appears to be an easier task, but, as OpenBSD port maintainer said, porting KDE means developing KDE. XFCE is yet easier, but the port still lacks some features of Linux version. In fact, neither BSD has desktop environment on par with Linuxes. There is LXDE, but…
2. Everything comming from Linux tends to increasingly depend on systemd. While ubiquitous on Linux, systemd doesn’t fit BSDs in any way. Frankly, much of the desktop stuff that comes from Linux doesn’t make sense on BSDs.
3. Building desktop environment is not as difficult task as it was say 10 years ago. Lots of already BSD-enabled libraries that can be used. I won’t be at all surprised if building desktop environment, that would be on par with GNOME on Linux, requires less effort then porting desktop environments.
4. PC-BSD is backed by iX systems, rather well-off company that earns its money selling BSD hardware and services. They pay BSD awareness, FreeBSD events, etc. They largerly benefit from any positive news about BSDs. Given that desktop environments are one of the most problematic areas of desktop BSD adoption, most likely they would sponsor the effort.
Good points about Linux only (like systemd and etc.). Same goes for Wayland for example.
IMO, Systemd is the real problem. It seems to pretty much be only built for one user (Red Hat) and for one purpose, rapid boot.
Best i can tell, this so that RH can be used to spin up server instances in cloud services similar to Amazon EC2.
Systemd is not a problem for Linux. But it’s not portable to BSDs or Solaris / illumos for example, so for them it’s a problem if systemd becomes a hard dependency for the desktop environment like KDE or Gnome and etc.
Edited 2014-04-25 18:21 UTC
I think it’s because now that everyone has (finally!) realized what a enormous clusterf–k SysV init is RH needs something else that is “enterprise” (another world for “overly complex and complicated”) to woo their customers.
As much as I dislike SysV init, systemd is an even worse solution. Jesus, all we need is service management and that was solved years ago in daemontools/runit (and even Upstart) by running the daemon in the foreground. No need for this cgroup nonsense to track daemons that “backgrounds” and blah blah blah. Just write your daemons correctly to begin with.
Edited 2014-04-26 07:53 UTC
So what existing init do you like? Or should I say, what existing init do you hate the least?
Personally I’m partial to the BSD-style init that Slackware uses; I just “get it” but that may be because I’ve been using Slackware off and on since around 1999.
The ones I like are Upstart and daemontools/runit. OpenBSD’s init style is also nice and simple although perhaps a bit rudimentary.
Good post. I was worried about the whole “developed form scratch” thing until I saw that it’s being built on top of Qt, so no it’s not really being built from scratch, and porting applications won’t suffer form it.
In related news, IBM enthusiasts will be designing a new carrying case for the golf-ball type-writer.
Hopefully it is a great desktop, and as long as it stays away from anything related to GNOME I’m all down for it.
I agree, i am not a GTK3 fan.
Anything touched by RH seems to turn into sewage these days.
They already have investment in a Qt stack. Why not collaborate on the development of the Hawaii desktop? http://www.maui-project.org/
Does any of the BSD’s have the required kernel infrastructure to support Wayland and systemd? As that is what Maui is targeting.
The way things are developing, the BSD’s may become the standard bearers for SysV init, X window system and the continued availability of an X based desktop environment. I think that is a good thing. Disclaimer: I’m a Linux desktop user.
>BSDs
>SysV init
Dumbass, BSD has NEVER used SysV init, even when it still came from Berkeley. They use their own BSD style init, which is simpler but still shell based.
As for the Wayland port, OpenBSD was considering some preliminary work for it as part of a GSoC 2014 idea.
FreeBSD has moved from rc to rc.d, beginning somewhere in v5, if I remember correctly. In rc style init system, /etc/rc will execute other scripts like rc.local, rc.network, rc.firewall, rc.foo, rc.bla and so on. A comparable initialization mechanism has been used in System III if my memory serves me right. The newer rc.d system uses shell scripts located in the /etc/rc.d directory which contain keywords in order to have rc determine the correct order (PROVIDE and REQUIRE). Those scripts act according to parameters like “start” and “stop”, they can be used in combination with the “service” command.
However, FreeBSD can emulate SysV init (see “man init” for details) when called as a user process, for example “init 1” or “init c”.
Each BSD actually has it’s own init style and all of them gloriously are free of the idiotic runlevels.
Since those projects will continuously evolve with the Linux kernel & constantly change as the Linux kernel interfaces change, there’s no way for the BSDs to have the required kernel infrastructure. Linux has, historically, changed infrastructures as fast as infants’ diapers are changed. Really, it’s not a matter of having the infrastructure, it’s more-so of stabilizing the interface to the infrastructure. The BSD systems don’t change interfaces & infrastructure on a dime. They also don’t expect to have to continuously change their software for the sole purpose of an API change. Engineer a solution, then add it; rather than patch a problem & spending the next 2 decades engineering solutions for the problems that the patches cause.
I wrote a desktop for QNX some years ago, never released it, but it was functional. If you’ve already got a decent toolkit (which QNX had, with bells on), then a desktop is not really that big an endeavour. It’s probably less than 1% of the work of making something like Adobe’s Creative Suite, and probably less than 1% of the work of just making a web browser.
If this is going to be just another rip off of the Mac or Windows, then I don’t see a lot of point. But if it’s going to be a genuine attempt to make something innovative and different, then I wish them all the luck they need. Will be interesting to see how it progresses.
Gnome and KDE “has the developers, testers, and users required for creating, maintaining, and improving an entire desktop environment” and they have created and improved one piece of shit.
Edited 2014-04-25 03:08 UTC
…what you all are saying is simply crap. Virtually ALL of the Desktop Environments (DEs) are tailored for Linux. Trying to use those DEs on any of the BSDs seems fine, until you come to the area of device management or any other part that deals with any of the kernel-land interfaces. Once you notice that, you start to notice other nuisances. To be completely honest, the BSDs have always needed their own X Windows server, DEs, & 3D driver solution. The fact of the matter is that many of the developers of these projects focus on Linux, while support for the BSDs languish. Trying to consolidate resources & share these projects is a noble goal; however, it simply doesn’t work out well for the BSD platforms. The integration just isn’t seemless -something that you wouldn’t notice, if don’t use any of the BSDs.
I agree with you on that. Every time I try a more popular DE on FreeBSD/PC-BSD it works alright until I start playing with settings. Then problems and rough edges keep popping up.
About damn time
Why don’t they just fork something existing, and work with that as a base?? They could just keep it GPL if they can’t find anything suitable that is licenced BSD.
Great to see PC-BSD doing well. PC-BSD is what got me into *nix operating systems. I had an old Windows 98 PC I was using for a webserver and I wanted to run something better and more secure. I read about Linux and the BSDs and decided FreeBSD is what I wanted. But after struggling to install and configure it without success, I found PC-BSD. Great OS for beginners, and a full LAMP package was very helpful.
I’m guessing the “new” DE is probably more like a custom KDE implementation, using QT and such. I always found their KDE implementation very good, so I have high hopes.
Having used PC-BSD, I agree with Thom — the team just doesn’t have the resources (eg, developers) needed to pull this off without making critical compromises elsewhere.
In my experience, the project has trouble keeping up with fixes/changes as it is, without taking on this major new development project.
If they go ahead with it, I wish them well and hope I’m wrong. Kudos to the PC-BSD team.
I’m curious about why they didn’t start with Enlightenment. It’s under an MIT license, and it has lots of effects that don’t require accelerated graphics.
The more the merrier in my opinion.
We need not only more choices, but also to remember how to create software from scratch.
Eventually we’ll all be too abstracted from how computers work at the lowest level. In other words we’re loosing control over how we got here.
I would like to see more people building computers from scratch, operating systems from scratch and even doing electronics with brass tacks and real wooden breadboards. We can’t forget the fundamentals. The kids of today are growing up fully abstracted to how any of it works. We’re becoming too consumer oriented and need to get back to creativity.
So yes, I think this is a good idea. Who cares if it doesn’t seem ‘efficient’ to some of you. We need more people willing to re-invent the wheel in my opinion. It’s not just about how far we can get. We must remember how we got here.
I’m a huge fan of KDE3, I’m using it everyday at work but It’s difficult to find new software for it, very difficult to mantain… I’d love to use a new updated dekstop simple, lightweight and KDE/QT based. A dream come true.
Go go PC-BSD!!! Hope you the best.