Connect to the Internet in More Than 10 Clicks (BUG)
(As requested by my audience, I wish to make a distinction in pointing out that my post is mistaken in thinking that the described features of KDE are a default. They are, contrarily, a BUG with my own installation. Nevertheless, I will hold the secondary method of connecting to the internet as something that can be revised and improved.)
As KDE worked on the new Network Manager for version 4, they decided to make the interface more powerful and through the use of the network manager widget, more accessible to users. The interface is indeed powerful and full of features aimed at working out every corner of customization on a given network. Wireless or wired network, they both have their particulars when it comes to connecting to the internet.
However, seeing at different operating systems, KDE's network connection manager seems cumbersome and hard to work with. Nevertheless, the disappointment comes from the first run of this software, not from the daily usage of it. Once everything is set up correctly, network manager will not even ask you to connect to the internet. It will do it automatically. Automatization however, comes at a cost of more than 10 clicks and a little extra knowledge about the particular network you are trying to access.
I really like the way that Mac OSX does it. They have an icon sitting in the system tray, as you click it, a drop down menu displays all the available networks. Users select one and you would be connected to it after typing a password if necessary. The next time you are around such network, Mac OSX will connect to it automatically unless you decided to run on a different network.
Another method out there is very similar to what Mac does. Ubuntu tries to connect to the internet in a simlar way. They also feature a network icon in the system tray. If you click on it, you get a drop down menu with available networks as well. Selecting and entering a password are very much the same as in Mac, except Ubuntu (and KDE) have an added feature of password protection through Gnome's Keyring and KDE's KDE Wallet.
During the first time that a user selects a network that is password protected, he is asked to create an extra password to store the password just entered into the wireless network setup. At first, this was very confusing to me, I didn't know if my password for the keyring or wallet had to be my root password, or a different password. I also ran into the mistake to placing the wireless network password into the keyring, thinking that it was the password for the wireless network being requested. Obviously, this can be solved by being smarter ;) But it does show that there is a learning curve that is not so smooth when working with networks. I think people who are trying to connect to the internet on a Mac also run into confusion as well. An icon sitting at the top bar that does not connect to the internet unless you click on it also represents a learning curve but one that takes less time and effort to do what we do most on a computer, browse the internet. Mac does not have a keyring or wallet that pops up upon login when you are trying to connect to a specific network. I believe these settings could change to make it easier on users to connect. If security seems to be an issue, probably Linux can get by easily with it. openSUSE could also benefit from a more seamless connection to networks by simplifying the process, it would stand out as the distribution with the easiest connection learning curve.
This is what one currently sees as you try to connect to the internet under KDE 4.6
You connect not by clicking on an available network, you do it by clicking network manager.
Then you are presented with a screen that contains the first network setup designed for wired networks. However, in my case, I need to add a wireless network, so I click on the second tab.
After that, if this is the first time, then I do not see any available networks. I need to click on add.
A rather complicated form shows up in which a tiny button called "scan" will eventually show me how to find the wireless network I want to connect to.
Once clicked, I am greeted with a very tiny window that shows all available wireless networks around me. If you are in the place where I am, it is really hard to see where I am to click. Instead I choose "Details" which organizes the view in a list.
I select the one I am to connect to and hit OK.
Later, to finish the process, I need to hit OK a couple of times and then I get back to the desktop. But I am not connected yet. I need to go back to the Wireless icon on the system tray and now I am able to see the new connection I added through the prior steps. Once I click on it, the setup will finally end by connecting to the wireless network.
I think this is really too long. It could be made short and there are plenty of good options out there that have simplified this process. openSUSE will benefit from a simpler way to connect to the internet if they choose to create a more simple method. My idea would be to simply show a drop down menu with the available networks, select, enter password and connect. If a keyring or wallet is necessary, allow this services to unlock passwords at system startup without having to enter it everytime.
By the way, this is the way I have my current KDE 4.6 with openSUSE 11.4. I am liking it so far.
As KDE worked on the new Network Manager for version 4, they decided to make the interface more powerful and through the use of the network manager widget, more accessible to users. The interface is indeed powerful and full of features aimed at working out every corner of customization on a given network. Wireless or wired network, they both have their particulars when it comes to connecting to the internet.
However, seeing at different operating systems, KDE's network connection manager seems cumbersome and hard to work with. Nevertheless, the disappointment comes from the first run of this software, not from the daily usage of it. Once everything is set up correctly, network manager will not even ask you to connect to the internet. It will do it automatically. Automatization however, comes at a cost of more than 10 clicks and a little extra knowledge about the particular network you are trying to access.
I really like the way that Mac OSX does it. They have an icon sitting in the system tray, as you click it, a drop down menu displays all the available networks. Users select one and you would be connected to it after typing a password if necessary. The next time you are around such network, Mac OSX will connect to it automatically unless you decided to run on a different network.
Another method out there is very similar to what Mac does. Ubuntu tries to connect to the internet in a simlar way. They also feature a network icon in the system tray. If you click on it, you get a drop down menu with available networks as well. Selecting and entering a password are very much the same as in Mac, except Ubuntu (and KDE) have an added feature of password protection through Gnome's Keyring and KDE's KDE Wallet.
During the first time that a user selects a network that is password protected, he is asked to create an extra password to store the password just entered into the wireless network setup. At first, this was very confusing to me, I didn't know if my password for the keyring or wallet had to be my root password, or a different password. I also ran into the mistake to placing the wireless network password into the keyring, thinking that it was the password for the wireless network being requested. Obviously, this can be solved by being smarter ;) But it does show that there is a learning curve that is not so smooth when working with networks. I think people who are trying to connect to the internet on a Mac also run into confusion as well. An icon sitting at the top bar that does not connect to the internet unless you click on it also represents a learning curve but one that takes less time and effort to do what we do most on a computer, browse the internet. Mac does not have a keyring or wallet that pops up upon login when you are trying to connect to a specific network. I believe these settings could change to make it easier on users to connect. If security seems to be an issue, probably Linux can get by easily with it. openSUSE could also benefit from a more seamless connection to networks by simplifying the process, it would stand out as the distribution with the easiest connection learning curve.
This is what one currently sees as you try to connect to the internet under KDE 4.6
You connect not by clicking on an available network, you do it by clicking network manager.
Then you are presented with a screen that contains the first network setup designed for wired networks. However, in my case, I need to add a wireless network, so I click on the second tab.
After that, if this is the first time, then I do not see any available networks. I need to click on add.
A rather complicated form shows up in which a tiny button called "scan" will eventually show me how to find the wireless network I want to connect to.
Once clicked, I am greeted with a very tiny window that shows all available wireless networks around me. If you are in the place where I am, it is really hard to see where I am to click. Instead I choose "Details" which organizes the view in a list.
I select the one I am to connect to and hit OK.
Later, to finish the process, I need to hit OK a couple of times and then I get back to the desktop. But I am not connected yet. I need to go back to the Wireless icon on the system tray and now I am able to see the new connection I added through the prior steps. Once I click on it, the setup will finally end by connecting to the wireless network.
I think this is really too long. It could be made short and there are plenty of good options out there that have simplified this process. openSUSE will benefit from a simpler way to connect to the internet if they choose to create a more simple method. My idea would be to simply show a drop down menu with the available networks, select, enter password and connect. If a keyring or wallet is necessary, allow this services to unlock passwords at system startup without having to enter it everytime.
By the way, this is the way I have my current KDE 4.6 with openSUSE 11.4. I am liking it so far.
18 comments:
Why are you doing it that way? Can't you just click on the applet, click "show more", then click the wireless network you want? That is three clicks.
While KNetworkManager is not the easiest tool out there, 10 clicks are too many!
- Click on KNetworkManager icon, which is the equivalent of the icon in mac (1 click)
- Click on WLAN Interface on the left hand side of the panel (2nd click)
- A list of networks will be show. Select yours (3rd click)
- You will be asked for your connection passphrase. Insert it, and click OK. (4th click).
The configuration of the keyring has nothing to do with KNM :-)
I actually tried your suggestions, except it would not show any networks that were available, and once I click on whatever was there, nothing would happen. No passphrase popup or anything. Maybe it's just my computer, but that is why I went with the other method.
If that is the case it sounds more like a bug than a design problem. I can't test it myself until tomorrow.
I also still have problems with knetworkmanager/plasmoid, mostly because of creating new connections. for example I am still unable to create "system connections", because that checkbox is greyed out, and there is no possibility (perhaps except login as root) to activate it.
so I still go through the painfully process everytime to install nm-applet (some wired icon-dependencies). Why there is not a kde-clone of nm-applet, this applet is really fine!
The network management stack is very complex. This makes this class of problems harder to fix than necessary.
Instead of writing long illustrated blog entries about how bad it sucks, some people have decided top actually do something about it, and pitch in.
You, my friend, are actively discouraging these people (including me). That's probably not your intention (not sure what you're intention indeed is), but it's definitely its effect.
Thanks :(
The network management stack is very complex. This makes this class of problems harder to fix than necessary.
Instead of writing long illustrated blog entries about how bad it sucks, some people have decided top actually do something about it, and pitch in.
You, my friend, are actively discouraging these people (including me). That's probably not your intention (not sure what you're intention indeed is), but it's definitely its effect.
Thanks :(
[In case anybody who reads this feels encourage do improve the interaction schemes, feel free to drop me (sebas - kde - org) a mail and I help you getting started.
@sebas
I don't see this post merely as a 'this sucks' type of commentary. I think it is constructive enough in highlighting better possible ways of doing things. Whether or not you see it as valid or want to take the points on board is your choice, but trying to silence all such bloggers is not very democratic.
In open source there is an inherent gulf between the developer community and the user community and their objectives, and the 'intention' you seek here is I guess simply a desire to contribute in some way that the user best knows how. Please realise that not everybody is or can be a coder and hence can 'pitch in' in the way you would prefer. By working as an open source developer one must appreciate community involvement in the many forms it can take, coding, commentating or otherwise. If anything, ordering all potential contributors to code or shut up only creates an image of developers as cliquey and high and mighty, and puts off more users from becoming developers!
One can find a myriad deconstructive bloggers or critical comments around the internet that are evidently unhelpful or discouraging towards open source developers. In this case, I might not agree with all this poster's opinions but I think they are nonetheless provoking some constructive dialog from the user perspective.
As other commentators have said, the design is really much better than it appears in your blog entry.
I do sympathize - I had a network-manager bug in 11.3 that made it frustrating as well. I'm sorry you've had a hard time of it. Still, it's important that you realize that the behaviour you're describing doesn't apply to everyone.
I suggest you look at the userbase article [1] and ask about these issues on IRC to resolve them.
I would also respectfully suggest that you edit your blog entry to indicate that what you're seeing there is due to a bug, and not design.
[1] http://userbase.kde.org/NetworkManagement
The Mac design is actually even better than what you describe. The password *is* securely stored in a keyring, it is just that that keyring is transparently unlocked when you log in. There has been years of discussion about making that sort of thing work in Linux as well, but it will probably be another few years before it happens.
I would definitively do what you guys suggest. Do not get discourage, I do not mean to sound rude. If I knew how to code, I would definitively help. Hopefully you also realize that much software in the open source world is in a constant sequence of development, meaning that there is always room for improvement. I have only pointed out something that as a user, seemed hard to do for me. There has to be some value in working with people who run into problems and that is what I am part of. You guys are doing a work that is simply beyond me and I praise your efforts because of that. Nevertheless, this does not mean, in any way, that such results are perfect, neither are my posts. Therefore, I will change my entry with a (BUG) line on the title. Just don't take it personal. I love KDE and openSUSE. If I wasn't genuinely interested in them, I would not even have a blog.
Keep up the good work guys.
Hi Andy,
I'm just a bit bemused having tried to post a comment on your blog. It doesn't let me login properly with my Blogger account ID, producing some error message, only to show me as logged in when I press the back button to my comment. I can then preview and publish the comment, even receiving an email confirmation (as below), but it doesn't show up. I had the same problem posting to another Blogger post a few weeks ago.
Before I contact Google, can you just confirm you never saw this post and cannot access it anywhere in your account or email?
--
gumb has left a new comment on the post "Connect to the Internet in More Than 10 Clicks":
@sebas
I don't see this post merely as a 'this sucks' type of commentary. I think it is constructive enough in highlighting better possible ways of doing things. Whether or not you see it as valid or want to take the points on board is your choice, but trying to silence all such bloggers is not very democratic.
In open source there is an inherent gulf between the developer community and the user community and their objectives, and the 'intention' you seek here is I guess simply a desire to contribute in some way that the user best knows how. Please realise that not everybody is or can be a coder and hence can 'pitch in' in the way you would prefer. By working as an open source developer one must appreciate community involvement in the many forms it can take, coding, commentating or otherwise. If anything, ordering all potential contributors to code or shut up only creates an image of developers as cliquey and high and mighty, and puts off more users from becoming developers!
One can find a myriad deconstructive bloggers or critical comments around the internet that are evidently unhelpful or discouraging towards open source developers. In this case, I might not agree with all this poster's opinions but I think they are nonetheless provoking some constructive dialog from the user perspective.
@Andy(gumb):
I think you misunderstood sebas's point.
I believe sebas was pointing out that creating an illustrated blog entry takes a certain level of effort, and that for a similar or lower level of effort, the blogger could have contacted other users or developers, or filed a bug report. These actions could either resolve the problem or just inform developers that the bug existed. However, creating a blog entry doesn't actively engage developers, and appears to be only critical.
Think of this engagement in a different context. What would you do if something you bought brand new didn't work properly? You might take it back to the store, or just call and ask/complain. But talking about it in a park where the store owner or product designer may or may not hear you is not going to solve the problem.
The really cool thing about open source is that we are all empowered. If something is broken, it can be fixed. But the way to get things fixed is by engaging in the process - asking on IRC, filing a bug report, or even modifying the software code.
Like your Desktop theme.
What is the name of the theme that you are using ?
It is qt-curve with an ambiance setting from kde-look.org. Faenza icons, and the wallpaper is called DVNO. The plasma theme is Fushigi.
I tried it myself. It popped up the password dialog no problem.
"I have only pointed out something that as a user, seemed hard to do for me."
The problem is that rather than asking whether there was an easier way to do it, you just jumped right in and asserted it was broken. By simply asking someone if there was an easier way you could have avoided wasting a lot of your time and a lot of our time. In the end, your proposal is pretty much the same as how things actually work.
This is not the first time you have done this, nor is it the first time I have pointed it out to you. On several occasions you have criticized developers for things that weren't even true.
This is part of the reason we have specific channels for these sorts of suggestions. bugs.kde.org, the brainstorm forum, and the mailing lists all have many more readers than your blog and would almost certainly have gotten you an answer more quickly.
If you are going to set yourself up as someone pushing KDE development forward, you need to at least do your basic homework. If you routinely make pronouncements that are rendered meaningless because they are derived from basic factual errors, people are simply going to stop paying attention.
"There has to be some value in working with people who run into problems and that is what I am part of."
The value of working with someone is directly proportional to the value of that person's input. If you want your input to be valuable, you need have at least some understanding of how the situation really is.
The KNM in openSUSE 11.4 is useless, so's the plasmoid afa my computers and the mobile UMTS are concerned. I find the KNM/PNM to be some kind of buggy work in progress.
I managed to run smoothly nm-applet from GNOME on a freshly installed openSUSE 11.4. My guide is here:
http://susestarter.blogspot.com/2011/03/step-2-need-mobile-broadband-use-gnome.html
If you got wifi/UMTS problems, first to do is deinstalling NetworkManager-kde4-libs package and isntall Networkmanager-GNOME. If nm-applet still has problems, first then kernel or firmware can be blamed.
I am disappointed that openSUSE 11.4 has been released this way with a useless KNM/PNM that confuses thousands of new users who will have serious problems with the most important thing in the avarage laptop/computer these days - network connectivity. I vote Networkmanager-GNOME to suse's default, even in KDE!
openSUSE users can open YaST > Sofware > Software repositories > Add... > Community List > KDE Updated apps. (http://download.opensuse.org/repositories/KDE:/UpdatedApps/openSUSE_ + your distro version and update their plasmoid-networkmanagement package from there.
If you want even more recent snapshots you can find them at your distro version's playground repo.
Post a Comment