Two of my housemates knocked on my door together a while back. They wanted help getting their computers on the house’s WEP protected wireless network. I pulled out my Ubuntu 6.10 laptop and sat on the couch with them, after writing down both the passphrase and the 26-character hex passphrase on a piece of paper.
My laptop uses NetworkManager, which as far as I can tell, has three functions:
- Select a wireless network and associate with it
- Lock up the entire computer and force you to hard reboot
- Lock up and prevent the computer from hibernating, sleeping, or associating with any wireless networks
Fortunately for me, NetworkManager seemed to be focusing on task #1. But on to my housemates. First, I read them the passphrase. My housemate running OSX 10.2 selected our wireless network in the menu at the top of the screen, put in the passphrase, waited a moment, and was then on the network. It took her about a minute.
My housemate with a brand-new Windows XP was still fiddling. There was nowhere to enter the passphrase. Various confusing dialogs were buried in processes like: Right click => “Properties” => “Advanced” tab => “Details” => “Configure” => “Apply”. There were two separate groups of dialogs to manage the wireless card – one accessible from the system tray, and one from the wireless device icon in Windows Explorer. From time to time, one or the other of these would announce that another program was in charge of this wireless card and it was throwing up it’s hands in defeat.
There were at least three dialogs that wanted the 26-character hex passphrase, and two of them used a password field, which shows bullets for the characters, and required you to type the 26 characters twice. Hello? The point of asking for a password twice is so that you can be sure the user didn’t type it wrong as they were changing it. What is the point of asking for a passphrase twice before checking it? If the user types it wrong, you can check that as soon as it’s typed in once. The person who thought that the average user would be able to type 52 seemingly random characters correctly, without any feedback from the interface, needs to be shot.
What’s more, there was a checkbox that switched the field type from a password field to a regular text entry field, adding to the complexity of the interface and completely bypassing the (dubious) purpose of using a password field in the first place – to protect from prying eyes. And none of these dialogs ever remembered the password, so each time we entered a password, pressed “Apply”, and the application announced that it was, in fact, not in charge of the wireless network, all 56 carefully entered characters were gone forever.
The three of us sat in the living room for an hour. My housemate with OSX was blithely reading webpages and chatting. My housemate with the brand-new Windows XP machine sat there, clicking “Apply” and “OK” and entering, and re-entering, the same 26 characters in the same five boxes over and over and over again. And I sat there checking my email and watching the logs on the wireless router, where I could see my housemate’s Windows machine request an address via DHCP, and the router send a DHCP address in response, and then, ten seconds later, see another request for an address from the Windows machine.
After an hour, I told my housemate that I was sorry and she would have to call technical support. I don’t know how, but she did eventually get on to the wireless.
Results of the impromptu OS wireless comparison:
- Best: Mac OS X; <1 minute, 2-3 clicks, enter password once
- Works sometimes: NetworkManager under Ubuntu; <1 minute, 3 clicks, enter password once, unreliable
- Better luck next time: Windows XP; >1 hour, too many clicks and password entries to count