Are FreeBSD developers ever going to start focusing on user facing stuff?

[kpa]

Yes, sad.


That is why FreeBSD sucks in 'desktop space'. If developers would instead sit down and nail all these inconvenient problems one by one, then FreeBSD would be very usable as a desktop, but its easier to give up and use Mac OS X.

Many of those inconviniences are a result of outside developments that are not in the hands of FreeBSD developers at all. The X windowing system has been slowly rewritten piece by piece using components that have been initially done with no regard for portability and FreeBSD seems to have the hardest time adapting because of de-centralized development. FreeBSD still has many strengths but desktop usage is not one of them unless thing change drastically in the near future.

 

[kpedersen]

and there is just simply too many features I need to give up.

At the moment this is no problem but I am convinced that the future of Mac OS X and Windows are locked down with vendor stores like iOS and Metro. It would be nice to have FreeBSD ready for when that time comes

 

[vermaden]

@kpa,

This is not an excuse.

Is anybody forcing FreeBSD developers to NOT work on X11?

Take a look even at OpenBSD and their Xenocara work, they have a lot less resources than the FreeBSD project and yet, they did it (and still do).

Same for binary packages, FreeBSD recently got PKGng but that still does not solve package 'versions' with some features enabled/compiled or not, OpenBSD a lot earlier had pkg_add -i where in interactive mode you could select which binary package with what features enabled you want to install.

From what I know OpenBSD developers use OpenBSD as their 'desktop'.

 

[Crivens]

@kpa,

This is not an excuse.

Is anybody forcing FreeBSD developers to NOT work on X11?

Take a look even at OpenBSD and their Xenocara work, they have a lot less resources then FreeBSD project and yet, they did it (and still do).

This might be the result of some kind of paradox where less resources get you forward faster. There is a line where your resources start to be insufficient to follow the Xorg pace of $THROW_IN_THE_NEWEST_STUFF (which may be a strand of the Linux' NIH virus, but I digress), so you are better off with a fork. That is what the OpenBSD camp did, and it served them well. They need some efford to sometimes merge things from the X.Org trunk, but the rest is able to continue without interruption.

On the other side, the FreeBSD developers have enough resources to have a chance to stay ahead of that line, so they try. But this eats up a lot of resources, leaving fewer free to do new things when compared with OpenBSD. So we may spend most of the manpower playing the "keep up" game, but only going places where others want to be. We may have different ideas about what we want.

I for one would prefer not to have KMS but a small, fast X11 which can be run without being root or without blowing the kernel security levels out of the window. Is Xenocara available in ports? Or maybe I should try out some medieval XF86 from 10 or 15 years past?

Disclaimer: most of this is just guesstimation on my part. Someone who knows more, please correct me if I am wrong.

 

[throAU]

I for one wish persons suggesting that the desktop usage is not worth one's while to give specific examples so maybe improvements could be made; I've been using it as my principal desktop ( on computers and laptops ) for almost a decade, without hardly ever a divergence to any use of any other operating system, and still have original files from the 2004 install here and there, despite many hard disk failures [some involving the freezer trick data saves...]

OK, I'll have a go. I'm aware some of these are not easy or maybe not even possible, but that doesn't matter to me as an end user, I'll use what works. I have money and will pay for what works, so $0 is not a limiting factor for me. This is by no means a comprehensive list, just a few off the top of my head.

• Automatic wifi connection to known networks as I roam to different networks. I should be able to pick from a list and enter details in the box to connect, when it is seen.
• Easy set up of L2TP/IPSEC VPN connectivity
• Driver updates from third parties via binary module without needing to recompile the kernel or parts of it.
• Automatic back ups (something like time machine)
• System-wide scripting in the GUI - like folder actions via automator in OS X. Even better if it is point and click for workflow creation like automator.
• Pervasive client-observer architecture. An example: you make a change to a folder in OS X. All finder windows are automatically updated, as changes send a notification to anything observing it. No need to hit F5 to refresh.
• Gesture support for my trackpad
• GUI consistency between applications. In OS X, every single app uses cmd+q to quit, for example. The free Unix desktop is a mish-mash of different paradigms with little UI consistency and different toolkits.
• Full CODEC support for whatever media I want to play.

And no, these aren't all FreeBSD problems. I don't expect all of them to ever be fixed. I'm a pragmatist. For desktop use I'll use something that works. For server use, I use what works.

The desktop hardware I buy (Mac laptop) is nice anyway and OS X is already there. As far as I'm concerned, FreeBSD focusing on those issues is a bit of wheel reinvention, trying to solve a problem that is already solved.

I think more benefit would come of focus on what FreeBSD does best (the server side), and attempting to make THAT easy to integrate into other homogenous environments.

The desktop market is already shrinking anyway.

 

[vermaden]

The desktop hardware I buy (Mac laptop) is nice anyway and OS X is already there.

I date to doubt that you would bought the same hardware if it came with Windows.

I used Mac OS X for more then a year and its quite nice (probably least PITA from commercial desktop OSes) but that keyboard insanity with CMD keys keeps me away from it.

 

[kpa]

@kpa,

This is not an excuse.

Is anybody forcing FreeBSD developers to NOT work on X11?

Take a look even at OpenBSD and their Xenocara work, they have a lot less resources than the FreeBSD project and yet, they did it (and still do).

Same for binary packages, FreeBSD recently got PKGng but that still does not solve package 'versions' with some features enabled/compiled or not, OpenBSD a lot earlier had pkg_add -i where in interactive mode you could select which binary package with what features enabled you want to install.

From what I know OpenBSD developers use OpenBSD as their 'desktop'.

I'm the last one to make excuses on behalf of those in charge of the development of FreeBSD. I'm constantly annoyed by the rather haphazard manner of development and lack of central management that would keep tabs on what the individual developers are doing and would set concrete goals for the projects.

 

[vanessa]

Hey guys, I can agree with most of your arguments except to give up FreeBSD as a desktop OS and use it on servers only. Where do you draw the line between a server and a desktop? Servers also have graphics, new chipsets, or maybe a video surveillance card inside.

The lack of drivers is a huge problem for every OS, because not everybody can choose the hardware depending on the OS, and not every OS can be shipped with its own hardware (or the other way around) like Apple does.

Then, driver programming is one of the most difficult areas of development. Many people can code a desktop application, even more a PHP one. But go find someone with enough experience in driver development and expect him to have the same hardware as you!

I can only agree with @thorbsd that FreeBSD just has to have driver support for a Dell(!) laptop nowadays. If you go and count the number of those trackpads, you'll certainly come up to hundreds of thousands all over the world, which could be powered by a common driver. Heck, we are talking here about a basic input device like a mouse!

The only viable solution for this problem is what @da1 already suggested. There must be a way to initiate and coordinate paid development for each and every area of interest. I am also one of the many willing to pay for certain features, especially drivers, as I don't have the know-how to develop them myself.

FreeBSD would only profit from such paid projects, as the number and quality of ports would rise rapidly. And think about the many talented developers without a regular income! They will be happy to deliver whatever piece of software the community needs.

 

[mix_room]

I am also one of the many willing to pay for certain features, especially drivers, as I don't have the know-how to develop them myself.

Remember that drivers have the added problem of being hardware dependent, which means that you need to know how the hardware works. This means that the hardware vendor needs to cooperate. It is not only a problem of money.

I agree with the OpenBSD mantra of 'eating your own dog food', as not only does it mean that you are more likely to find the problems, but they are more likely to affect you, and thus you are more likely to fix them.

 

[vanessa]

Remember that drivers have the added problem of being hardware dependent, which means that you need to know how the hardware works. This means that the hardware vendor needs to cooperate. It is not only a problem of money.

Sure, however there are for sure many missing drivers available for Linux in source code, which can be ported to FreeBSD without vendors' support. Let's start with them ...

 

[jb_fvwm2]

I perceive a number of posts wishing there were more persons working on FreeBSD code. Back in the days of shareware and DOS, I was fond of flowcharts if something was complex. I continually wish for flowcharts additionally to other sources of information...

.....................................................
   w 7 .................................> FreeBSD
........           ^a^                      ..........

For instance, if that "a" was a small-print printable box-to-box flowchart(s) of a Windows install migrated to FreeBSD, (the flowchart large and complex...) it might attract more users to FreeBSD from Linux and from Windows than other methods. It might even help FreeBSD users migrating from MBR to GPT, to CUPS, to Wireless, etc if those sections were included. I even imagine if were unfoldable to several yards-on-a-side, folded into booklet form, and sold by FreeBSD retailers...

 

[throAU]

I date to doubt that you would bought the same hardware if it came with Windows.

I used Mac OS X for more then a year and its quite nice (probably least PITA from commercial desktop OSes) but that keyboard insanity with CMD keys keeps me away from it.

CMD + Apple keys are just like Windows and Menu keys on a regular keyboard.

Took about a week to get used to it.

edit:
Not saying necessarily to give up on desktop development, but there is PC-BSD for that, and most of the software involved (Gnome, KDE, X11, etc.) is not written by the FreeBSD team anyway.

The angle I am coming from is that resources spent by the FreeBSD team in this area are effort already being made by PC-BSD, and would be resources diverted from improving FreeBSD as server, which it is a far more realistic option for. Again, I am aware people do use it for a desktop, but you really do have to make some serious sacrifices to do so.

On drivers: I truly believe that until there is a stable ABI for either Linux or FreeBSD, getting driver support is going to be an uphill battle that will never be won. We can either accept that, and live with limited driver support, or accept the fact that some companies don't want to provide programming specs or source code (the actual reason is entirely irrelevant really, the end result of not getting info is what matters) and provide an ABI.

It's been 20+ years now for both Linux and FreeBSD and the driver situation hasn't changed much. I don't think it's going to any time soon.

 

[CoTones]

It seems some people here don't understand desktop importance. OK, just an example: bankrupted SUN with excellent and expensive Solaris for servers versus always defective and best by general usability desktop Windows from multimillionaire Microsoft; Linux distributions everywhere - desktops, servers, firewalls, routers, media centers, mobile phones versus BSDs with limited capabilities for new hardware; and situation (only slightly better because of more free developers here) like Open Solaris will happen without generous anonymous donations and support.

 

[vermaden]

CMD + Apple keys are just like Windows and Menu keys on a regular keyboard.

Took about a week to get used to it.

Nope. if it was ONLY that, then I would be using Mac myself.

On Windows its CTRL-C / CTRL-v / CTRL-T / CTRL-W / CTRL-* and WIN- for some Windows features (move windows/open Explorer/...).

On Mac there is no CTRL-C for copy, its CMD-C, shortcuts that were 'bound' to CTRL now are 'bound' to CMD ... but not all of them, some are still CTRL-* and some are CMD-*, but in Terminal, its still CTRL, not CMD ... I just want to stay away from this mess.

Not saying necessarily to give up on desktop development, but there is PC-BSD for that, and most of the software involved (Gnome, KDE, X11, etc.) is not written by the FreeBSD team anyway.

So when are PC-BSD developers going to write VT code to get the console back after 'visiting' X11? How is their work on ATI graphics cards drivers going? I assume that they already have Nvidia Optimus worked out and implemented, they just do not post it public yet because they want to polish it a little bit more. </sarcasm>

PC-BSD team get what FreeBSD does and make a default install (and installer) from the base system and from the ports.

 

[wblock@]

So when are PC-BSD developers going to write VT code to get the console back after 'visiting' X11?

Not PC-BSD, the Foundation has funded that. I can't find a link right now.

How is their work on ATI graphics cards drivers going?

Again, not PC-BSD, but committed to -head, last I heard. I have not tested it yet.

 

[ondra_knezour]

bankrupted SUN with excellent and expensive Solaris for servers versus always defective and best by general usability desktop Windows from multimillionaire MICROSOFT

You mean the bankrupted Sun with the desktop enviroment bought by the multimillionaire Oracle without any of them? And more to that, which have quickly lost interest in the desktop parts of the Sun like the OpenOffice.org suite and keep their server business like the Sparc hardware and the Solaris software, even if they already have the Oracle Unbreakable Linux etc.? Just to note that having the desktop is not the only requirement to success on the IT market today.

 

[vermaden]

@wblock@,

You obviously missed the </sarcasm> tag at the end of my sentence

@throAU suggested that the PC-BSD team should do 'desktop FreeBSD' and I showed him that these things can not be done by PC-BSD team itself.

 

[jrm@]

I used Mac OS X for more then a year and its quite nice

Was the sarcasm tag missing here too?

<light-hearted-jab>
Or, maybe you like a layer of obfuscation between the UI and the real OS. Or maybe you think Lenovo isn't dictatorial enough and you need real hardware lock-in.
</light-hearted-jab>

Whatever floats your boat.

 

[silicium]

After retiring my main Windows desktop (just tired of XP for daily work), despite having successfully tried PC-BSD, I had to switch to Linux because the deal breaker was the lack of X input driver for my Wacom Intuos2 tablet, which could have been the most expensive part when new (I got used A4 and A6 models and an unused "new old stock" A5, they even work on Windows 7 through complex XP driver installation hacks).

I installed Debian minimal base with x11-wm/openbox and some xfce4 tools instead of its bloated default GUI, and could quickly use my tablet thanks to https://github.com/RoaldFre/wacom_serial5

To bridge the gap between developers with limited time and users wanting drivers, I'd agree to pay developers to write guides how to port drivers from Linux source.

 

[vermaden]

Was the sarcasm tag missing here too?

No I really liked it, it run 130-150 days of uptime (reboot only for updates), if only not that keyboard layout ;p

<light-hearted-jab>
Or, maybe you like a layer of obfuscation between the UI and the real OS. Or maybe you think Lenovo isn't dictatorial enough and you need real hardware lock-in.
</light-hearted-jab>

Lenovo is not better here (and Dell also), the last good keyboard layout by Lenovo was in T420/T520/W520. They spectacularly fscked up the T430/T530/W530 keyboard :\ Same with Dell laptops newer then E6410/E5510/...

 

[ondra_knezour]

guides how to port drivers from Linux source

Basics http://www.freebsd.org/doc/en/books/arch-handbook/driverbasics.html
Book http://nostarch.com/bsddrivers.htm
Review of the book and an example http://forums.freebsd.org/showthread.php?t=31610
Older resources compilation http://forums.freebsd.org/showthread.php?t=1566

 

[throAU]

On Mac there is no CTRL-C for copy, its CMD-C, shortcuts that were 'bound' to CTRL now are 'bound' to CMD ... but not all of them, some are still CTRL-* and some are CMD-*, but in Terminal, its still CTRL, not CMD ... I just want to stay away from this mess.

So how do you manage with the FreeBSD or Linux desktop then?

 

[throAU]

@throAU suggested that the PC-BSD team should do 'desktop FreeBSD' and I showed him that these things can not be done by PC-BSD team itself.

There is nothing (other than the possible developer time/skill constraints) to stop the PC-BSD team from doing this themselves if they wanted to fork and submit patches back (or not).

 

[vermaden]

There is nothing (other than the possible developer time/skill constraints) to stop the PC-BSD team from doing this themselves if they wanted to fork and submit patches back (or not).

Fork is probably the WORST thing they could do, they do not have resources for that.

 

[scottro]

For what it's worth, on an Asus Zenbook with an i5 and integrated Intel graphics card, I can go from X11 to virtual terminals without problems. I can also quit out of X. However, on another machine, with an i7, also with integrated Intel graphics, I can neither quit X (if I run startx I have to stay there till the machine is shut down) nor get a VT. So, while I share vermaden's reservations about Intel, I'd add that it might work, depending upon the system. Note that I'm using that machine with 10-Alpha, I don't remember if it worked with 9.x.

In both cases, only the vesa() video driver works. No luck with Intel drivers.

I think there is some importance to attracting desktop users, mostly because of agenda setting theory. http://en.wikipedia.org/wiki/Agenda-setting_theory

As for the (alleged) topic of the thread, I think (and this is purely subjective, based on nothing but anecdotal evidence), that Ubuntu, and Mr. Shuttleworth's aggressiveness with some of the hardware manufacturers, did a great deal to increase support for Linux, both in hardware and software, and I think it's one reason one sees so much supported hardware and software for Linux.

On the other hand--gosh, I could argue either side of this argument, depending upon my mood--Linux has, again, in my highly subjective opinion, lost something by becoming so user friendly. As it gets more Windows like, on some of the forums, one sees more Windows like problems, such as GUI issues preventing a system from booting.

[EDIT many days later]

Vermaden (as usual) is correct. The machine I mentioned, where I could get back to console or use ctl+alt+Fx to get to virtual console, only works with the vesa() driver. When using the Intel driver, it is as Vermanden describes, one can neither quit out of X nor go to a virtual terminal.