Wireless GUI: "Killer App" for FreeBSD?

[kpedersen]

I have heard a couple of times in this thread that FreeBSD commandline users / developers should give GUI applications a go and maybe start to appreciate it.

This is quite amusing because I am sure that command line users have used GUI applications and have more experience in both than the GUI users who have tried using the command-line.
So I turn the question around... Perhaps people wanting (and seemingly needing) to add superfluous GUI cruft to FreeBSD should actually give the command line a go. It might open some doors for them

GUI software does have it's place (once a portable library and CLI version of the software has been made). What would be a bit silly is trying to get BSD developers to waste their time on it when instead they are more qualified to do important stuff anyway.

But anyway, leave the GUI stuff to the Gnome project. They seem to be doing a good job (*snigger*)

As for the absolute mess that is web browsers and HTML... yeah, perhaps we should go back to the text only gopher protocol in the command line. Less stupid flash adverts. And give me Mutt over gmail or hotmail web any day of the week!

 

[CreativeGPX]

"Command line" is far too imprecise to mean anything in a discussion about usability or user efficiency though. All that really means is that you are deciding to use characters. It doesn't say whether there are a ton of those characters or a few. It doesn't say if those words are cryptic or readable/memorable. It doesn't say how much study and specialized knowledge is required for a user to come up with those characters. It doesn't say how finicky the input processing is. It doesn't say how human-readable the output is. There are incredible command line interfaces (Git :f ). There are also lots of terrible ones. Apparently, we have a terrible command line solution to the scenario that was mentioned. Regardless of how that is solved, the problem is the lack of solution not the nature of the solution. If a good CLI solution existed already, making the GUI version would be trivial. If good GUI solution existed, making the CLI version would be trivial. The challenge is in the automation of the actions performed, not the ability to solicit input and present output.

 

[tzoi516]

So I turn the question around... Perhaps people wanting (and seemingly needing) to add superfluous GUI cruft to FreeBSD should actually give the command line a go. It might open some doors for them

Why do I care about fuel economy, cruise control, power steering, locks, and brakes? It's all cruft on a car because all I need is for it to take me from point 'A' to point 'B'.

I don't think a GUI is better, but I don't think it's inferior either. It's like buttons on a phone - they're going away whether people like it or not, what's progress today is the lost love of the NES tomorrow.

 

[shepper]

Selecting a wireless interface at boot is getting some attention in OpenBSD

 

[nanotek]

A final postscript from me as OP.

Thanks to all those who contributed, and particularly to those who made positive suggestions and even volunteered to contribute something: I wish you good luck with this.

To those who got all hot and bothered with the suggestion that a GUI tool might be useful, all I can say is: if you are reading this on a browser like Firefox or Chrome, then shame on you: what is wrong with lynx?

:e :e :e

Touché! Very good call, my friend.

Lots of people who have commented seem to have willfully either misunderstood, or be misinformed. Relatively few people have commented on the mobile broadband issue, yet I have said a couple of times that this is the thing that is particularly difficult in FreeBSD: there is no point commending tools like wpa_supplicant for this as they do not do that job (and are not intended to).

Don't let the fanboys and deliberately obtuse bother you, you were more than eloquent, not to mention respectful, in your OP. The fact is, FreeBSD is behind the curve in this department and (as far as I know) no developers commented in this thread so the vacuous remarks don't reflect the ambitions or sentiments of those that matter.

I have given up and gone back to Debian. I can tell you that I do *not* have either Gnome or KDE installed, I am still using FVWM, but I do have nm-applet and it works: I can connect to mobile boadband with a few clicks and then get on with the things I actually want to work on. For all I know it might be troublesome in a large corporate environment, but I would not use it there anyway. For connecting a laptop to a mobile broadband service on a train passing through areas of good and bad connectivity, it works beautifully.

Maybe dual booting with Debian and a shared /home partition might be an idea till FreeBSD catches up?

If Microsoft can do it (and do it well), FreeBSD will...eventually.

 

[kpa]

If Microsoft can do it (and do it well), FreeBSD will...eventually.

Microsoft has been able to do what they have done simply because the emphasis has been on the GUI from the very start of MS Windows and the GUI is an integral part of their operating system, it's not called "Windows" for nothing. On FreeBSD the GUI parts have always been on optional addon status and it has always been assumed that the GUI parts come from the X11 windowing system (with no real alternatives to speak of) with some 3rd party window manager/DE suite. Furthermore it has been assumed that the functionalities that the window manager/DE provides is up to the user him/herself to configure and the OS itself forces no policy whatsoever on what type of applications or utilities the DE suite (for example) should contain.

All I'm trying to say with this is that development of such graphical "killer app" is quite out of scope for the FreeBSD OS itself. It can make it easier for someone to develop such app by improving the programming APIs that are used to configure the network interfaces and fixing the problems with the wireless networking.

 

[kpedersen]

FreeBSD will...eventually.

There is no reason to assume it ever will. This sort of GUI requirement is low priority for most of the FreeBSD userbase.

Luckily you GUI guys have Windows, Mac OS X and countless distros for that

Oh... and PC-BSD if you do preferably decide to stay with BSD.

 

[roddierod]

I think this is relavent to note when people talk about userbase and more user equates to more contibutions of code: http://spectrum.ieee.org/computing/soft ... ting-linux

Also there is a FreeBSD Desktop mailing list http://lists.freebsd.org/mailman/listin ... sd-desktop and there wiki page https://wiki.freebsd.org/DesktopIntegration

There is not much traffic on the list, but I think it purpose was to discuss things such as this. Maybe if we could get more traffic things like this would get made.

 

[nanotek]

If Microsoft can do it (and do it well), FreeBSD will...eventually.

Microsoft has been able to do what they have done simply because the emphasis has been on the GUI from the very start of MS Windows and the GUI is an integral part of their operating system, it's not called "Windows" for nothing. On FreeBSD the GUI parts have always been on optional addon status and it has always been assumed that the GUI parts come from the X11 windowing system (with no real alternatives to speak of) with some 3rd party window manager/DE suite. Furthermore it has been assumed that the functionalities that the window manager/DE provides is up to the user him/herself to configure and the OS itself forces no policy whatsoever on what type of applications or utilities the DE suite (for example) should contain.

All I'm trying to say with this is that development of such graphical "killer app" is quite out of scope for the FreeBSD OS itself. It can make it easier for someone to develop such app by improving the programming APIs that are used to configure the network interfaces and fixing the problems with the wireless networking.

FreeBSD will...eventually.

There is no reason to assume it ever will. This sort of GUI requirement is low priority for most of the FreeBSD userbase.

Luckily you GUI guys have Windows, Mac OS X and countless distros for that

Oh... and PC-BSD if you do preferably decide to stay with BSD.

You're missing the forest for the trees: it's not about a GUI in the base system. If you read the OP and, in my case, my posts we've not explicitly requested any such thing. It's about basic functionality existent in other CLI/TUI environments that FreeBSD lacks in its default installation and the absence of trivial features in ports for both graphical and text environments.

I can take my Windows 7 or OpenSUSE notebook to the airport with me and use mobile broadband in the cab on the way, seamlessly pick up the WiFi hotspot at the terminal while waiting for my plane. Jump back on mobile broadband after arriving interstate and taking the train to my final destination where I switch over to my clients wireless AP. And do the same in reverse on my return.

And, given OpenBSD is looking toward improvements in this area, you can bet FreeBSD will follow suit.

 

[kpedersen]

I can take my Windows 7 or OpenSUSE notebook to the airport with me and use mobile broadband in the cab on the way, seamlessly pick up the WiFi hotspot at the terminal while waiting for my plane. Jump back on mobile broadband after arriving interstate and taking the train to my final destination where I switch over to my clients wireless AP. And do the same in reverse on my return.

To do all this is trivial in FreeBSD. When I have taken my laptop to the airport, it is very easy to swap networks (probably quicker than opening up any program). Depending on how your mobile broadband works, again, that should just be connecting to a different ssid with wpa password? Same with the client's AP right? Where are you finding the difficulty exactly? When using Linux, I tend to use wpa_supplicant directly anyway because it *is* an easy TUI tool, the only difference is that Linux calls the WiFi interfaces stupid names.

The more important issue in my opinion, is FreeBSD's slightly unreliable suspend resume, meaning that moving around the airport usually involves shutdown / restart.

So like I said. If you want a WiFi GUI program in ports, thats obviously fine. But also try out PC-BSD, that comes with a WiFi GUI tool by default with no extra work required.

And, given OpenBSD is looking toward improvements in this area, you can bet FreeBSD will follow suit.

From the misc@ mailing list, it seems like an OpenBSD user is attempting to put some wifi selection in the boot scripts. I dont quite think this means OpenBSD is looking at including this in the actual OS and there is no evidence as such.
Besides, if you may notice, FreeBSD includes far less in the base than OpenBSD. The goals are not the same between these two operating systems. Otherwise we would have X11 in base (though not necessarily Xenocara).

But that said... I really don't want to be "that guy" that puts a negative spin on any project idea. So if you (or someone else) wants to develop an easy to use WiFi tool. May I suggest using the Motif GUI toolkit? I know the CDE project is lacking a WiFi tool and this project might be able to hit two birds with one stone.

 

[Deleted member 13721]

FreeBSD will...eventually.

There is no reason to assume it ever will. This sort of GUI requirement is low priority for most of the FreeBSD userbase.

Luckily you GUI guys have Windows, Mac OS X and countless distros for that

Oh... and PC-BSD if you do preferably decide to stay with BSD.

Tell me, how would desktop FreeBSD turn out any differently than the already existing PC-BSD? I'm still trying to figure that out...

 

[kpa]

FreeBSD will...eventually.

There is no reason to assume it ever will. This sort of GUI requirement is low priority for most of the FreeBSD userbase.

Luckily you GUI guys have Windows, Mac OS X and countless distros for that

Oh... and PC-BSD if you do preferably decide to stay with BSD.

Tell me, how would desktop FreeBSD turn out any differently than the already existing PC-BSD? I'm still trying to figure that out...

No difference at all because the desktop parts wouldn't come from FreeBSD, they would still be third party software pre-configured and tailored for FreeBSD. So why re-invent the wheel when there is already PC-BSD?

 

[roddierod]

Tell me, how would desktop FreeBSD turn out any differently than the already existing PC-BSD? I'm still trying to figure that out...

No difference at all because the desktop parts wouldn't come from FreeBSD, they would still be third party software pre-configured and tailored for FreeBSD. So why re-invent the wheel when there is already PC-BSD?

I think the main difference is in the packaging system, those pbi or apps or whatever the current term is. Are they all still self contained binaries? I don't follow PC-BSD development closely. I installed version 10 and I could not get software to install using that app store thing.

 

[robspop]

To do all this is trivial in FreeBSD. When I have taken my laptop to the airport, it is very easy to swap networks (probably quicker than opening up any program). Depending on how your mobile broadband works, again, that should just be connecting to a different ssid with wpa password?

Sorry but no, it is not easy. I can go from pub to hotspot to home to friend's house and swap wifi using wpa_supplicant quickly and easily. Mobile broadband uses a modem dongle that requires a ppp configuration script, these are not quick and easy to set up. I will accept that doing that is largely a one-off task, but it is by no means trivial. And the problem became much bigger when FreeBSD 10 came along and refused to even recognise the dongle as a modem but insists on treating it as a mass storage device. I can spend as long as I like trying to configure this using any method from command line to willing the laptop to do what I want via the power of thought alone, and it still won't connect to the internet.

And I wish you would actually read my posts if you are going to comment on them: for the nth time, I have no problem using the command line: almost all my work is done there and at any given time I probably have a browser and maybe a dozen xterms running. I never asked for a gui as such, just gave one as an examle of an easy-to-use tool that FreeBSD doesn't have.

Let me ask you a question directly: if FreeBSD's attitude to suggestions like mine is "go away, we like it as it is", where do you see its future lying?

 

[arp242]

FreeBSD has a graphical control panel, it's called vi.


There are really 2 problems:
1) FreeBSD lacks *some* system for easier configuration.
2) PPP is bloody hard to setup

I think point 1 has some merit. The smarter programmers automate stuff (that is, make it easier)*, yeah, I know how to setup my wireless NIC, but I really don't want to. If possible, I want the blighter to `just work'™ with as little fuzz as possible. I'd much rather go out and get a date than muck about with wpa_supplicant.
The *really* smart programmers automate stuff, yet make sure it can still be modified or hacked if need be. That way, you have the best of both worlds.

Making a graphical (GUI or curses) WiFi scanner/setup script is fairly easy, and you can just write the chosen settings to /etc/rc.conf and /etc/wpa_supplicant.conf. Any half-witted programmer could probably do it in a week.
It doesn't even need to be `graphical' as such, the OpenBSD installer just asks you a bunch of questions, but it works very well. The old XFreeConfig did the same.
There's no need for NetworkManager, or other `magic' services, which are really just a pair of complicator gloves.

CentOS does this fairly well, you have a set of system-config-* tools you can choose to install, which just modify various config files in /etc/. I never used them much, but my less experienced co-worker used to use them (until he got instructed in the true ways of UNIX, and became enlightend). It's a good idea, but somewhat poorly executed on CentOS.

AFAIK FreeBSD has one tool, which is tzsetup(8).

As for point 2, I know little about PPP. I once tried to setup a mobile dongle, and stopped when blood started to come gushing out of my nose. I spent 2 weeks in recovery.


*: An example of people who really don't seem to understand this, are (some of) the ArchLinux devs. About a year ago they removed their perfectly functioning installer, you now have to install ArchLinux using bash + fdisk + tar + praying you don't do anything wrong and hose your entire system.

 

[roddierod]

So as I am seeing here the issue is more of hardware support for a usb modem wifi dongle...something I don't have or have ever used.
The other being the "easy of use" for configuring PPP - something which I have done since I bought a shiny new 56K modem way back when in the FreeBSD 3.x and 4.x days.

Being I don't have one of these mobile wi-fi things, I don't think there is anything I can do to help create something easily switch PPP connections. Too bad there isn't some kind of virtual device.

 

[jb_fvwm2]

Due to a recent broadband downtime, I scripted a reboot-to-wifi.(sh) and a reboot-to-dsl.(sh) each of which copies seven! files from the script's staging directory to /etc /boot (etc)... but I had a laptop already setup with wifi. Envisioning as I type this a hundred such users each posting to a thread here their scripts for a GUI (ncurses) programmer to program into a frontend app; seems all it would take is time. But there are specific (firewall...) (ndis-or-not) files within that script that makes it problematic... maybe if/when someone ever works on such a frontend one could check to see what Linux does in that regard (I think they have only two or so default firewalls in common use IIRC...)

 

[CreativeGPX]

Any half-witted programmer could probably do it in a week.

You're not proposing that anything that the individual user of FreeBSD could code a solution to in a week need not be included with FreeBSD, are you?

 

[kpedersen]

Mobile broadband uses a modem dongle that requires a ppp configuration script, these are not quick and easy to set up. I will accept that doing that is largely a one-off task, but it is by no means trivial.

Since the OP topic is Wireless GUI, I assumed you connected to your mobile broadband via a WiFi interface (as seen in slightly less crippled Android phones).
Yes, PPP is awkward but frankly, I am more confident I can set it up on FreeBSD with the excellent handbook (https://www.freebsd.org/doc/handbook/ppp-and-slip.html) rather than some short lived [GUI] tool (like on Gnome/Linux) that will probably be replaced next month with some other time consuming mess.
If users are not prepared to go through the handbook and set this up, FreeBSD really *is* the wrong tool for them.

the problem became much bigger when FreeBSD 10 came along and refused to even recognise the dongle as a modem but insists on treating it as a mass storage device. I can spend as long as I like trying to configure this using any method from command line to willing the laptop to do what I want via the power of thought alone, and it still won't connect to the internet.

If FreeBSD 10 recognises it as a mass storage device, then lets stop wasting time thinking about perceived ease of use and lets spend time on getting the important stuff working. As you know, it is quite unlikely a nice easy to use script or GUI tool is going to help with this sort of issue.

I never asked for a gui as such, just gave one as an examle of an easy-to-use tool that FreeBSD doesn't have.

Since you seem to have experience in using *nix, my responses were not aimed at you. There are others in this thread that seem dead set on the "GUI is the most modern and only way" attitude that I was trying to get through to.

Let me ask you a question directly: if FreeBSD's attitude to suggestions like mine is "go away, we like it as it is", where do you see its future lying?

I am going to push back on this one. I see the future of FreeBSD in a much better position than the absolute fsck up that is the average "user friendly" Linux distro with their self indulgent desktop environments, complete with binary configuration blobs, stupid interface names and completely unusable commandline tools. The way that mainstream Ubuntu has turned out really does make me question open-source usability "experts".
There is a reason why FreeBSD is becoming attractive to Linux users recently. For example, we have less than a fraction of developers and... well, FreeBSD is pretty darn great isn't it?

 

[tzoi516]

the problem became much bigger when FreeBSD 10 came along and refused to even recognise the dongle as a modem but insists on treating it as a mass storage device. I can spend as long as I like trying to configure this using any method from command line to willing the laptop to do what I want via the power of thought alone, and it still won't connect to the internet.

If FreeBSD 10 recognises it as a mass storage device, then lets stop wasting time thinking about perceived ease of use and lets spend time on getting the important stuff working. As you know, it is quite unlikely a nice easy to use script or GUI tool is going to help with this sort of issue.

Most wireless broadband modems have a dual feature: flash storage and wireless broadband. Generally the drivers are installed on the flash part for ease of installation compared to the earlier modems. My opinion is FreeBSD recognizes the flash storage, but I'm sure if he was to run pciconf it will show the wireless broadband as being unrecognized.

 

[wblock@]

From u3g(4):

     In some of these devices a mass storage device supported by the umass(4)
     driver is present which contains Windows and Mac OS X drivers.  The
     device starts up in disk mode (TruInstall, ZeroCD, etc.) and requires
     additional commands to switch it to modem mode. If your device is not
     switching automatically, please try to add quirks. See usbconfig(8) and
     usb_quirk(4).

 

[kpa]

If the 3g modem worked on 9.2 but doesn't work on 10.0 it's certainly a regression that needs to be reported via a PR so it can be fixed for 10.1. Just waiting apathetically that someone else reports the issue and it gets fixed is a non-starter idea with FreeBSD.

 

[robspop]

Since the OP topic is Wireless GUI

Fair point, but that was not my title for the post, it got edited to encourage more people to read it. My title was just Killer App and the suggestion was for something that would manage all connections. I have no problem with either wired or wireless via wpa_supplicant, my point about the linux thing was that it manages ppp as well, which is rather handy imho.

Since you seem to have experience in using *nix, my responses were not aimed at you

Apologies if I seemed to get a bit huffy!

well, FreeBSD is pretty darn great isn't it?

Actually, yes, I think it is: I would tend to describe FreeBSD as real unix and most of the Linuxen I have tried as a real mess. Personally, I would like to see more people using FreeBSD, and that is why I started the thread: I think if it was all a bit easier on a laptop then more people would use it.

I have personally solved the problem by changing my dongle to a mobile wifi hotspot so no need for ppp any more.

 

[CreativeGPX]

The way that mainstream Ubuntu has turned out really does make me question open-source usability "experts".

Ubuntu is aiming to be usable to a wide audience and usability is audience specific.

In FreeBSD, being able to say "read the manual" to most questions is a strength. The system tries to be central and consistent so that a motivated and computer literate person who has free time can basically figure out how to use or change any aspect of the system. In this case "usability" means something like, "for a well trained user, what can they do and how quickly?"

Comparatively, in Ubuntu (and Mac and Windows), having to say "read the manual" is a weakness. The system tries to be something you can just jump into and start using intuitively without an explanation. You're not supposed to need time and effort to learn it. The interface is supposed to help you bump into features and settings that you didn't know about rather than you having to stumble across them in a book or a manual so that you can learn by doing. In this case "usability" means something like, "for an inexperienced user, what can be done without help and how hard is it to break things?"

Neither idea of usability is inherently better. I think the more widely used definition of usability is the latter which corresponded to intuitiveness. The former corresponds more to power and elegance. Without a doubt, the latter will appeal to way more people because most of the population isn't an expert, dedicates little time to becoming an expert and doesn't think it's their job to become an expert. FreeBSD's notion of usability works well for the smaller group of experts as long as we keep it so that the effort in reading that manual keeps allowing people considerable advantages over other OSs.

The question that this whole conversation seems centered around though is whether these two types of usability are contradictory or not.

 

[Cthulhux]

I think if it was all a bit easier on a laptop then more people would use it.

Why would we profit from more people using it? FreeBSD is already peeking at becoming more "Linux-y", probably driven by Linux refugees who think "this is too much tech, let's add shiny GUIs and wizards for basic tasks".