Do FreeBSD developers "eat their own dogfood"?

vall

Member

Reaction score: 6
Messages: 50

I read a lot of messages like "OpenBSD developers eats their own dogfood in time when FreeBSD developers just uses MacOS X".

I watched a lot of videos and photos from BSD meetings where there are really many Macbooks, but many Thinkpads with FreeBSD too and other laptops running Windows.

For now I only found this message, that can approximately tell about real situation, but not enough. I found (in releable source (twitter :p)) that Allan Jude, Hiroki Sato, Baptiste Daroussin, Ed Maste, George Neville-Neil runs FreeBSD on their laptops.

And I can't go for now to any devsummit for look to the hundred laptops and ask "do you run FreeBSD exclusevly or just today?"

So can anybody who knows the things tell something about dogfooding in the FreeBSD developers society?
Thanks a lot?
 

ronaldlees

Aspiring Daemon

Reaction score: 333
Messages: 767

Probably many of them have day-jobs that keep them from running FreeBSD on their laptop - unless they have two laptops. Since all FreeBSD devs are poor, they only have one laptop.

Relative to the dog food - my dog eats quite a bit better than I do. Mostly human grade stuff gets put into his bowl, and some days I'm envious. I'm not a BSD dev though - so maybe they only have dry kibble.
 

vermaden

Son of Beastie

Reaction score: 1,972
Messages: 3,183

Probably many of them have day-jobs that keep them from running FreeBSD on their laptop - unless they have two laptops. Since all FreeBSD devs are poor, they only have one laptop.
If they would be poor, they would use FreeBSD on $100 ThinkPad T400 instead of Mac OSX on Macbook ... if its employer laptop, then it should not be used for personal things.
 

Trihexagonal

Son of Beastie

Reaction score: 2,497
Messages: 3,057

If they would be poor, they would use FreeBSD on $100 ThinkPad T400 instead of Mac OSX on Macbook...

Just because someone uses a T400 doesn't mean they're poor. I've got a W520, T61 and an X61 and use them out of a love for vintage Thinkpads, not the balance of my bank account. I wouldn't trade any of them for a brand new Mac.

As far as that goes, I'm thinking of buying a T23 to play DOS games if I can find one in good shape.

Not everything boils down to the almighty dollar, or złoty as the case may be.
 

ralphbsz

Son of Beastie

Reaction score: 2,692
Messages: 3,553

Why do you want to force a developer who works in a remote part of the system (for example the guts of ZFS, or some ethernet device driver) to waste his time on installing and managing the FreeBSD user interface? Let that developer use whatever is convenient for him.

For many years, I was a developer in a big-name commercial file system, which runs on Linux and AIX (plus a few other OSes that aren't very relevant). I didn't store my own personal data in that file system (couldn't have done that legally anyway, since it requires a rather expensive license to use), and I used a company-owned Mac as my desktop machine. Some of my colleagues use Linux desktops, others use Windows machines. The real work gets done on hundreds of rack-mounted or virtual machines that one logs into remotely, so what does it matter what OS is handling the keyboard and mouse events?
 

tankist02

Well-Known Member

Reaction score: 47
Messages: 326

It matters because it would encourage developers to improve FreeBSD for desktop. As of now FreeBSD may not be adequate for people who expect modern desktop features with nice graphics.
 

ralphbsz

Son of Beastie

Reaction score: 2,692
Messages: 3,553

So, if there is a developer who wants to work on (to continue my example from above) the internals of ZFS, or the driver for some Ethernet chip, then you are suggesting that they instead invest time into trying to improve the desktop experience?

Two comments:

A: That's not how free software or open source works. People who volunteer their time get to pick what they want to work on. You can ask, beg and cajole, but trying to force them is going to end badly. Now, if you actually were paying for this product, and had a support contract (like for example RHEL, were customers give RedHat good money), then it's different: Then you could actually demand value for your money, or take your business elsewhere. But that's not the case here.

B: Improving the desktop at the expense of the basics is foolish. Because an operating system with strong basics (good hardware support, an efficient kernel, few bugs, fine file systems, ...) is useful, for example for servers. But an operating system with a fancy desktop with nice graphics is useless, if the underlying base system doesn't function well enough.

I think your suggestion is highly counterproductive.

In the end, what to use for the desktop is a personal choice. And there I ultimately agree with you: FreeBSD (and other Unix flavors) are not adequate. Which is why I always end up MacOS (preferred) or Windows (2nd choice) for my desktop environment, both at home and in the office. That's in spite of having access to many Unix machines, and the fact that I set up GUIs on *unix environments several times a year. But I also have many friends and colleagues who disagree, and run various Unix environments as their personal workstation. To each his own.

Ultimately, the fix for better personal environments will have to come from the user community, in the form of more investment (of time and money). Let me paraphrase and extend Fred Brooks' old statement about the cost of software: If it takes 1 unit of effort to develop a program of a given complexity, then it will take 3x effort to develop a system program (today we call that a kernel), also 3x effort to develop a program system (today we call that a package or product), and 9x effort to develop a system of system programs (today we would call that an OS, which contains both kernel and many packages). To add a GUI to it takes 27x the effort. For example, if you compare the amount of work that went into Xwindows (paid for by MIT and various computer companies in the 80s), KDE and Gnome, I think it is vastly larger than the amount of work that went into the Linux kernel itself (initially the product of a few dozen part-time college students) and the Gnu compiler, emacs and tools (a few MIT people and volunteers).
 
D

Deleted member 9563

Guest


It matters because it would encourage developers to improve FreeBSD for desktop.
I totally agree.
As of now FreeBSD may not be adequate for people who expect modern desktop features with nice graphics.
I'm not sure what "adequate" means, but I've been using KDE which has "modern desktop features with nice graphics". :) Yes, there are sometimes some glitches, and the same old bugs that never get fixed exist in the Linux version as well. I'm not sure what it takes to win, but in the meantime my experience is adequate.
 

UnixRocks

Member

Reaction score: 21
Messages: 86

I am not a developer, I am a Unix systems administrator. However, I'm using FreeBSD on my desktop at work and as a home file server / desktop at home. These are not laptops, they are full size tower systems. My netbook has a graphics chip that is not yet well supported in FreeBSD, so it has Devuan Linux 1.0 installed. My company provided laptop for when I work from home is an Apple DongleBook (MacBook Pro 15") with OS X. As the Apple is a company managed resource I am not allowed to swap out the OS on it. However, if I were allowed to I would probably put FreeBSD on that too.

Preferred GUI is Xfce4.
 
Last edited:

Oko

Daemon

Reaction score: 799
Messages: 1,620

I could maintain few ports or write a chapter of the Handbook and get the freebsd.org e-mail but that doesn't make me a developer. Developers in the definition above are people who are actually hacking on the kernel. I have no problem with number 409 of people involved in FreeBSD project just like typical OpenBSD release is signed by 100 people. However incredibly tiny fraction of that number hacks on things like KARL

https://marc.info/?l=openbsd-tech&m=149732265506347&w=2

Beyond the point I don't want to get involved in discussion about FreeBSD desktop.
 

fnoyanisi

Active Member

Reaction score: 103
Messages: 225

Let me start with a disclaimer; I use FreeBSD as my primary OS (no dual boot or whatsoever) on my Toshiba R830 (pre-owned but highly modified after I bought it); I like it, enjoy it, and hack it...no issues so far. We also have a pre-owned MacBook Pro A1278 (with a bit of hardware modifications and yes, FreeBSD folks are not rich) for my wife. For software that run on Windows only, I run a VM on MBP (the VM is just sitting there, in case we need it, has never used it though).

Now I can start writing about the topic...

I also have seen some youtube videos in which FreeBSD developers use an Apple laptop for their presentations...This instantly took my attention and I thought "hang on a sec, this guy claims the system he is developing is more superior than system X, giving some examples, talking about kernel data structures and comparing those with the ones in system X, saying how bad these kernel internals in system X are architected or designed in the first place, etc., yet, he uses a Mac"...and I AM A FreeBSD user. I am sorry guys, but this is, obviously, not a good way of passing the message to your audience. Let's face the cold truth....How could you encourage people to switch to FreeBSD or make them believe/think that FreeBSD is better from system X, if you are not using it personally? (no bigots saying "we do not need them" please)

It matters because it would encourage developers to improve FreeBSD for desktop.
So, I totally agree with this! +1

The fanatic "FreeBSD is designed for servers, bugger off desktops" slogan we all hear does not make sense at all. The more users you have, the more contributions you will get, the more third party software will be ported to FreeBSD...This is a while(1) loop that will repeat itself, simple as that.

Why do you want to force a developer who works in a remote part of the system (for example the guts of ZFS, or some ethernet device driver) to waste his time on installing and managing the FreeBSD user interface? Let that developer use whatever is convenient for him.
Nobody asks ZFS developers to halt their work on the kernel and start working on XFCE panel widgets. It is about changing the mindset and this does not mean using systemd or trying to make FreeBSD a tablet OS.

A very good example could be Ubuntu Linux...It may not be the best gnu/Linux distro from a technical or design perspective (no need to argue this statement here), but it contributed a lot to gnu/Linux world by making people aware that such an OS exists, and encourage them start using gnu/Linux and contributing to it.

And looking at what is happening in our close-minded world...a typical, one step further from where we are, example could be NetBSD operating system. It has a very good source tree, good quality code (I have it on my machine and I love browsing it with OpenGrok), and nice documentation. But looking at where the project is heading to with ever decreasing community.... I would afraid there will be no NetBSD after 10 years?

B: Improving the desktop at the expense of the basics is foolish. Because an operating system with strong basics (good hardware support, an efficient kernel, few bugs, fine file systems, ...) is useful, for example for servers. But an operating system with a fancy desktop with nice graphics is useless, if the underlying base system doesn't function well enough.
Being a "technical minded person", I agree with this, but have to say that this is does not reflect how the real world norms work. Unless you embrace a wider user community, and thrive to catch up with where the demand is, you will perish eventually.
 

Rand0m

Member

Reaction score: 8
Messages: 67

Let me start with a disclaimer; I use FreeBSD as my primary OS (no dual boot or whatsoever) on my Toshiba R830 (pre-owned but highly modified after I bought it); I like it, enjoy it, and hack it...no issues so far. We also have a pre-owned MacBook Pro A1278 (with a bit of hardware modifications and yes, FreeBSD folks are not rich) for my wife. For software that run on Windows only, I run a VM on MBP (the VM is just sitting there, in case we need it, has never used it though).

Now I can start writing about the topic...

I also have seen some youtube videos in which FreeBSD developers use an Apple laptop for their presentations...This instantly took my attention and I thought "hang on a sec, this guy claims the system he is developing is more superior than system X, giving some examples, talking about kernel data structures and comparing those with the ones in system X, saying how bad these kernel internals in system X are architected or designed in the first place, etc., yet, he uses a Mac"...and I AM A FreeBSD user. I am sorry guys, but this is, obviously, not a good way of passing the message to your audience. Let's face the cold truth....How could you encourage people to switch to FreeBSD or make them believe/think that FreeBSD is better from system X, if you are not using it personally? (no bigots saying "we do not need them" please)


So, I totally agree with this! +1

The fanatic "FreeBSD is designed for servers, bugger off desktops" slogan we all hear does not make sense at all. The more users you have, the more contributions you will get, the more third party software will be ported to FreeBSD...This is a while(1) loop that will repeat itself, simple as that.


Nobody asks ZFS developers to halt their work on the kernel and start working on XFCE panel widgets. It is about changing the mindset and this does not mean using systemd or trying to make FreeBSD a tablet OS.

A very good example could be Ubuntu Linux...It may not be the best gnu/Linux distro from a technical or design perspective (no need to argue this statement here), but it contributed a lot to gnu/Linux world by making people aware that such an OS exists, and encourage them start using gnu/Linux and contributing to it.

And looking at what is happening in our close-minded world...a typical, one step further from where we are, example could be NetBSD operating system. It has a very good source tree, good quality code (I have it on my machine and I love browsing it with OpenGrok), and nice documentation. But looking at where the project is heading to with ever decreasing community.... I would afraid there will be no NetBSD after 10 years?


Being a "technical minded person", I agree with this, but have to say that this is does not reflect how the real world norms work. Unless you embrace a wider user community, and thrive to catch up with where the demand is, you will perish eventually.

Totally agree, I personally use FreeBSD as my main OS on 2 of my Acer laptops. It is solid, stable, fast system even on the most humble devices. All what is needed IMHO is more support for desktop application (porting and updating).
I know it can be quite tricky job sometimes as most of the third party software is written to Linux, but I think with more people using FreeBSD as desktop we will find more and better quality codes.
 
OP
V

vall

Member

Reaction score: 6
Messages: 50

The things that 'kill' FreeBSD desktop are bugs like these with GLIB that takes YEARS to fix:
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=221679
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=199872
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=214338

... and without fixing them CAJA or THUNAR just crash on directory deletion ...

let's mothefacka create our own DEs, graphical servers and apps!!
a joke, but our team seems more focuses on porting than on creating new "BSDish" stuff.

I'm just learning everything about Unix (including internals), but I think main thing that can push developer to invest his/her free time to the open-source project is a lovе of this project and thirst for research. And maybe I don't aware, but what are the FreeBSD's ideals? Linux's ideals are to be everywhere and to do everything; OpenBSD's ideals are code correctness, simplicity, research and friendship of developers.

What about FreeBSD? Are they pragmatic sysadmins or their feelings are beyond of practical use? It has direct relationship to who developers are and what programs do they run. I'm not talking about fanaticism but if I've created app, I've mainly created it for myself and I will use it everywhere because for me it's most usable app in the world. If I don't use it - I abandon it.
 

Oko

Daemon

Reaction score: 799
Messages: 1,620

let mothefacka create our own desktops, graphical servers and apps!!
like a joke, but our team seems more focuses on porting then on creating new "BSDish" stuff.

I'm just learning everything about Unix (including internals), but I think that main thing that can push developer to invest his/her free time in open-source project is a love to this project and thirst for research. And maybe I don't aware, but what are the FreeBSD ideals? Linux ideals are to be everywhere and to do everything; OpenBSD ideals are code correctness, simplicity, research and developers friendship.

What about FreeBSD? Are they pragmatical sysadmins or their feelings are beyond of practical use? It has direct relationship to who developers are and what programs does they run. I'm not talking about fanaticism but if I create app I mainly create if for myself and I will use it everywhere because for me It's most usable app in the world. If I don't use it - I abandon it.
Dude you really need to work on your grammar.
 

fnoyanisi

Active Member

Reaction score: 103
Messages: 225

The things that 'kill' FreeBSD desktop are bugs like these with GLIB that takes YEARS to fix:
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=221679
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=199872
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=214338

... and without fixing them CAJA or THUNAR just crash on directory deletion ...
Among the available FOSS desktop environments, I would rank x11-wm/xfce4 as one of the most portable ones. Looking at their Xfce 4.12 Tour page, they are proud of providing screenshots from a system running OpenBSD-Current (see "A note on Xfce's portability" at the bottom of the page).

Unfortunately, free software (this is "free" as in freedom as Stallman stresses out), especially the major desktop environments such as Gnome or KDE (hey, those who use dozens of tiled terminals as their GUI, thank you, but you do not need to yell out how you can live without these desktop environments, and we actually do not need them, save it for yourselves) are now heavily dependent on software such as systemd (maybe no surprise here as Gnome is mainly driven by RedHat), hence hard to port any system but linux.

I wish I started using FreeBSD in 1998, when I first installed RedHat Linux 6.1 (no RHEL at that time) and I might have been a better contributor to the project (with a family and full time job, it is not easier now). However, as the proud users of FreeBSD, we should make as much contribution as we can (donations, bug reports, patches etc.). Just complaining about lack of X, does not help.

Want to advocate FreeBSD? use it and show that you actually use it, recommend it, hack it, contribute to it or wear a FreeBSD t-shirt :)

What about FreeBSD? Are they pragmatic sysadmins or their feelings are beyond of practical use? It has direct relationship to who developers are and what programs do they run.
See Section 1.3.2 here
 
  • Thanks
Reactions: Oko

Trihexagonal

Son of Beastie

Reaction score: 2,497
Messages: 3,057

I must be the luckiest guy running a FreeBSD desktop because in 12 years of doing so I've never had a problem I couldn't figure out either by working with it, google-fu, or searching the forums given enough time (This is a disclaimer), or had hardware that wasn't supported.

Maybe it's that I run older hardware, but it suits my purposes for everyday general desktop use and is what I prefer to run. (Has your brand machine flown on the Space Shuttle, Mir, or International Space Station?) There isn't one thing I want to do on my FreeBSD boxen that I cannot do and could be doing on a Linux box, or a Windows machine for that matter.

I currently have 3 FreeBSD laptops and 1 OpenBSD laptop, and with the exception of one of my FreeBSD laptops that had a HDD failure a couple weeks ago are the only computers I have in use. All of them run x11-wm/fluxbox and the same limited number of programs I found over time to do the work I need done, except OpenBSD doesn't have multimedia/xmms so I use multimedia/audacious on it.

Not everybody wants a DE that is reminiscent of Windows and has a plethora of programs included with it. Window has Linux now, if that's what you want you may well be happier using it.

Now in all honesty, and the reason for my disclaimer, I did struggle with Optimus on my Thinkpad W520. But I looked through the forums and saw a photo from a guy who had the same multiple renditions of the bootscreen and knew to use a Mode option during boot to bypass that issue. I did manage to get to the desktop but it had 640x480 resolution when native is 1920x1080 and at the time I was too tired to mess with it any further and put it off till I had more time to figure it out.

And embarrassing as it is to admit, having never had hardware issues I was not familiar with the xrander command and hadn't read about it the Handbook or I could have resolved the issue right then and there. It was not long after I heard about new security innovations being introduced in OpenBSD 6.2 and built it on my W520. I know what to do now if I want to switch it back though.

Want to advocate FreeBSD? use it and show that you actually use it, recommend it, hack it, contribute to it or wear a FreeBSD t-shirt :)

I can only guess how many screenshots I've posted over the years, I've even seen some of them on a google search for FreeBSD wallpapers. To the point it's almost boring for me since the only thing that ever changes is the wallpaper, let alone everyone else. so I don't flood the thread but have posted them in other forums.

Wrote a tutorial for beginners on how to set up a FreeBSD desktop, too. Spent my own money on a domain for a site to post it on before returning to the forums and hand coded it in valid XHTML and CSS.

If FreeBSD Mall ever sends it I've had an order in since August 3rd for a Power To Serve T-shirt, FreeBSD baseball cap and sheet of stickers. I already have a Classic FreeBSD T-shirt. :)

I tried talking the Assistant Executive Director of the apartment complex I reside in to switching their office machines to FreeBSD but he didn't even know what FTP was...
 

rufwoof

Active Member

Reaction score: 80
Messages: 235

I must be the luckiest guy running a FreeBSD desktop because in 12 years ...
Its my understanding that you do a fresh install at each upgrade? Do you burn a CD/DVD for that or use some other approach for the installation medium? So far I have 10.3 and 11.0 CD's and I used 11.0 to upgrade to 11.1 (suspect it might have been quicker/better to actually have download and burn a 11.1 CD). Wondering what others do. TIA.
 

Trihexagonal

Son of Beastie

Reaction score: 2,497
Messages: 3,057

Its my understanding that you do a fresh install at each upgrade? Do you burn a CD/DVD for that or use some other approach for the installation medium? So far I have 10.3 and 11.0 CD's and I used 11.0 to upgrade to 11.1 (suspect it might have been quicker/better to actually have download and burn a 11.1 CD). Wondering what others do. TIA.

I've never upgraded from one version to the next, not even a version bump, and always do a fresh build from ports. I have 2 ancient 1GB Flash Drives I dd the amd64 and i386 RELEASE versions to and use them.

I did burn FreeBSD 11.1-RELEASE to a DVD and it seemed to take a LOT longer even to get to the setup screen than with a USB, so I nixed it and went with the UBB.
 

rufwoof

Active Member

Reaction score: 80
Messages: 235

Thanks Trihexagonal.

Between my last post and this one I've (casually i.e. not hard at it, but more just doing it in the background) downloaded 11.1, burnt it to a CD, booted, installed, added xorg, jwm, pcmanfm, libreoffice, firefox-esr ... along with installing nvidia and more or less configured it from scratch ... trying to do so from memory (excepting the firewall bits). Had to reference some, but getting better at it, and relatively quick too. Bearing in mind I just install the lib32 option at installation, and use pre-built binaries (pkg), not ports.

Just finishing up on the sizing of fonts etc. so near totally done, along with adding a few extras (htop, leafpad, volumeicon, xload) ... which are just about done as I finish writing this.

Whoops, forgot scrot for screen capturing ... done

 
Top