Things have changed a great deal since 1982. These days, having a computer is almost synonymous with having access to the internet. In truth, a computer that cannot access the internet today is almost quaint. Sure the average Linux distribution comes with enough software to run an entire office out of the box, but if you can’t communicate with your customers via email, fill in forms at online websites, or look up information on the web you’re SOL.
For the home user, instant messaging has become almost as much of a “critical app” as email, family members are sharing photos and videos online, and much of the younger set is blogging or otherwise keeping friends up to date on their activities at various social networking sites like Twitter, MySpace, or Facebook daily.

While you can certainly buy a new PC with Linux pre-installed and configured (hopefully such that all your shiny new hardware is working properly), installing Linux on a PC that shipped with Windows will invariably require downloading drivers from a manufacturer’s FTP site or scouring tech forums for some clue about how to go about getting your hardware to work well, just work, or even be recognized at all. While there are still many reasons to have a PC even without internet access, getting Linux installed on it at all may actually require it.

For laptops, and an increasing number of PCs in environments where wired ethernet isn’t possible or is just too much of a hassle to set up, folks are routinely using wireless cards to communicate with a wireless router in order to connect to the internet. This is presenting some daunting challenges for Linux adoption at least among the home user crowd who might be looking at Linux to provide some fresh life for their older hardware. When dealing with laptops or PCs the problem is generally the same except that on laptops the wireless card may already be built in, though its true that many older laptops didn’t come with wireless at all. Cheap wireless cards sold in consumer electronics stores everywhere seem to work right out of the box for Windows users on PCs and laptops alike, but almost always present some kind of problem for a Linux user. While techies know that this is just a function of the fact that the companies are churning out Windows drivers for their hardware and keeping those drivers updated regularly, the average home user may not realize this until *after* they bring their shiny new wireless card home and pop it into their machine. Once they’ve plunked down $40 to $60 on a card, they just want to pop it in and have it work. Unfortunately, since Linux is not officially supported by nearly any manufacturer, it often does not - and for whatever reason if the user does not have another method to reach the internet, either through a wired connection to the router or another machine already connected to the internet, they will be unable to access the documentation and help forums they might be able to leverage to get this new hardware functioning. I see this as a major stumbling block to broader Linux adoption.

If by now you’ve guessed that I’ve recently had some run-ins with this problem, you’re right! Futzing around with getting a Linksys PCMCIA wireless card to operate at all in a Thinkpad (eventually succeeding with bwcutter) and a Belkin F5D7000 PCI card to even be recognized at all in an old Dell PC (also eventually succeeding with ndiswrapper) both under Ubuntu 8.04 was enough to convince me that unless the average user is very lucky, getting to the internet wirelessly in Linux is likely to be a problem. My eventual success was bittersweet because the only way I got these devices to work at all was to use the proprietary Windows drivers in a way that wasn’t intended by the developers. Some users of Linux would balk at the thought of using proprietary drivers at all, preferring to be completely free of “non-open” software. My wireless devices still don’t work natively under Linux, and as long as users can get them to work in this manner, what incentive is there for manufacturers to create a native Linux driver anyway?

It wasn’t my intention to explain the reasons for this problem - the constantly changing chipsets used by wireless card manufacturers, the entrenched Windows monopoly, the relative “newness” of the technology in general (as compared to ethernet), but rather to point out a weakness in Linux folks should be aware of. Before running out to buy a wireless device, check for Linux compatibility! Some chipsets are better supported by the two methods I mentioned. Don’t plunk down good money on hardware that won’t work at all in Linux, even if you’re running Windows today. If you decide to try Linux in the future you’ll wish you had.

The first open phone is already out (for developers) called the Neo Freerunner. By “open” I mean that the OS is fully documented, and customizable by those with the skills to modify it. It is using Openmoko which is the competing open mobile operating system (OS) to Google’s Android mobile OS (still no phones with that yet as far as I know). Both projects are leveraging the linux kernel which makes this exciting to me. You can use the new $400 device (which is currently sold out) on the AT&T network, but its mainly going to Universities for use in computer science classes right now.

To be fair, these types of devices are really a lot more than just a “phone”, they are really true mobile computing devices with touch screen input, nice screens (relative to the little Motorola RAZR screens if not the iPhone and enough OS horsepower to actually be useful for real computing. While I don’t expect anything as amazing as the iPhone to secretly emerge, I would be more interested in an open phone than a closed source one on principle as well as potential. (disclaimer: I don’t currently have a cell phone).

Two good websites to watch for news as the 700MHz spectrum opens up next year are:
http://www.openmoko.com/
http://code.google.com/android/documentation.html
We will either see a whole lot of new open phones appearing next year, or a whole lot of nothing since the ‘open spectrum’ was gobbled up by AT&T and Verizon (Google didn’t get any), and neither company probably has any real interest in allowing the other’s customers to operate over their network.

ThinkPad R32 2658-N3U
Original description: P4-M 2GHz, 256MB RAM, 30GB HDD, 14.1 XGA(1024×768) TFT LCD, 8x-3.3x DVD-ROM, Modem(CDC), Ethernet(LOM), Li-Ion battery, WinXP Pro, Wireless upgradeable

I like to try out laptops once in a while and put Linux on them to see if how easy it would be to do everything I need to do without paying a Microsoft of Apple tax. Since I “switched” to a Mac (from Linux) in 2002, I still use Linux at work, and have a family PC set up with Linux and I’ve been thinking it might be time to switch back. I’m not a big fan of Apple’s move to Intel, and while Macs are really slick, setting up house in Linux is a lot more fun. This is not a comprehensive report, but just my personal notes as I attempt to use the machine.

I’ve taken this machine home before and put FreeBSD on it (I had never played with that OS and it offered a good excuse). A Pentium 4 mobile, it’s quick enough for what I do with a computer. It’s missing some things I’ve grown accustomed to though. I have firewire devices (external hard drive and lightscribe DVD burner), this machine doesn’t have a firewire port. I like to plug my powerbook into the TV frequently to display slide shows of images, or watch Youtube videos, etc. This machine has no TV out capabilities. The battery may be shot as I only get an hour and a half of useage out of a charge. It runs hot. It has no touchpad, it just has the standard little red IBM nubbin in the middle of the keyboard (basically a strain gauge), and I think I’m the only one at work that actually likes that at all anyway.

Video
There’s nothing in the xorg.conf file about what driver X is using, but the R32 has an ATI mobility M6 and it seems to be autodetected. The display only does XGA, and so can only do 1024×768 which is a bit too small for most folks today (are our eyes getting any better?), but then my G4 powerbook is only XGA also, so it’s actually a treat for me since the screen is a lot bigger (14″ screen as compared to 12″ on the powerbook).

Wireless
Since the last time I had this laptop at home it’s been sitting in a closet and because it doesn’t have the optional built in wireless card, nobody really wanted it. I actually went so far as to start to order an internal wireless card for it, but the BIOS only accepts certain types and for whatever reason that didn’t work out through work so I eventually just brought it back in and there it sat. Recently I had purchased a Linksys wireless adapter for another machine at work, and luckily for me and unluckily for that machine’s user, it stopped working. I decided to bring it home to “test it” in the old Thinkpad. It seems to work fine (with the setups described below), which indicates the problem is with the other machine. Hopefully I can get that person another machine so I can hang onto the card, though to be honest I’d rather have an internal wireless card than the PCMCIA.

At first I tried to get the Linksys adapter working in Redhat Fedora 7 (before I returned it last time I wiped FreeBSD and tried out Fedora for abit) but I was stymied quickly. After figuring out what I had to do to get the Broadcom chip in the Linksys to work, I must admit it would probably have worked just as easily in Fedora, but I ended up installing Ubuntu 8.04 from CD, with gnome desktop leaving a 4G partition for Windows 98 which I need to put on later to run some windows only programs. There’s about 20G for Ubuntu. A PC2700 512Mb ram chip that I had recently replaced in my G4 Powerbook fit in this machine without issue and brought it up to 768Mb. This actually was the move that made it possible to load Ubuntu in the first place, because before I added the extra memory I was trying to boot the Ubuntu live CD and it was just reading from the CD for what seemed an eternity and I eventually just gave up.

The Linksys WPC54G PCMCIA card didn’t work out of the box in Ubuntu either. Network manager in Ubuntu was nice enough to let me know that the driver wouldn’t work till I had aquired the proprietary firmware which was the key piece of info I needed. I plugged in a regular network cable, and issued “apt-get install b43-fwcutter” which downloaded the open-wrt firmware, and cut out the files needed, and copied them into the right place. The network-manager icon then showed my local wireless and I could join. Connection speed is reported as 1Mb (abyssmal) though I haven’t tested it and it seems snappy enough for normal work. Others have complained about the same thing though, so will have to keep an eye on that thread for possible solution. There was another option (ndiswrapper) but my eyes glossed over reading up on that route.

[update 5/7/08: found a post in Bug#205114 that suggested
aptitude purge b43-fwcutter
aptitude install bcm43xx-fwcutter
Speed tests at speakeasy.net are now reporting somewhat faster download speeds, but not sure if it's just faster tonight or if something really improved. Update2: nah, it was no faster. Im back on b43 which gives me anywhere between 80KB/s and 120KB/s (down) where I usually routinely get 4 to 500KB/s on the Powerbook]

Shutting Down
Getting the wireless working introduced a new problem shutting down. Shutting down normally actually brought me to a black screen after the process, but didn’t shut the laptop OFF actually. To shut off all the way I’d have to hold in the power button for 5 seconds. I found that turning off the PCMCIA slot first allows Ubuntu to shut down completly. F Just have to add “/sbin/pccardctl eject” to the shutdown scripts: I wrote a little bash script
#!/bin/bash
/sbin/pccardctl eject
and saved it in /etc/init.d (and chmod to 755 so its executable)
Then in /etc/rc0.d, I made a symlink to my script named S22ejectcard
with ln -s ../init.d/ejectcard S22ejectcard

I have no issues with Suspend which works just like it does in windows. The little green Thinkpad moon lights up and it goes to sleep. I haven’t tried just closing the lid yet (I’ve been too busy wanting to keep working on it

Mouse buttons
The R32 has three thumb buttons right below the space bar, in perfect location if you’re using the little red rocket pointer gizmo. The middle of these three special thinkpad buttons didn’t do anything out of the box. However, I remembered that I was able to get it to act like the middle mouse button or a scroll wheel (of sorts) in FreeBSD and I had notes from that install. These pointed me over to the xorg.conf file. I modified mine with the following tidbit to get the center button working. I just rebooted after modifying /etc/X11/xorg.conf although I could probably have just restarted X windows I guess.
Section "InputDevice"
Identifier "Mouse0"
Driver "mouse"
Option "Protocol" "Auto"
Option "Device" "/dev/sysmouse"
Option "Emulate3Buttons" "true"
Option "Emulate3Timeout" "80"
Option "EmulateWheel" "on"
Option "EmulateWheelButton" "2"
Option "ZAxisMapping" "4 5"
Option "Buttons" "3"many
Option "CorePointer"
EndSectionAudio

Audio
Ubuntu had no problem detecting the Intel audio. Rythymbox needed gstreamer codecs installed before it could play mp3s (a non-free software requirement of mine for podcasts). Also, I needed to install “bad” codecs for .wmv, .pls, .ra support which are obviously needed since I like listening to live streaming radio and so it’s not really “bad” that it works now. One wierdness was that the .pls for liveireland.com while streaming the audio OK, wasn’t actually dislaying the names of the tunes for very long before it was crapping out but kept displaying the first song. I had to keep restarti ng it so that it would edisplay the song I was listening to.

Printing
My Brother HL-1440 is shared on my Apple Airport (a nice feature of the Airport), but got some strange error when attempting to connect to it. I used Avahi (a rendezvouz browser) and found my printer was at 10.0.1.1:9100, so I just went to System-Administration-Printing and selected New Printer. I picked AppSocket/HPJet Direct. Then I entered the IP address above into the Host box; Port 9100 was automatically entered into the box for port. Clicking forward, I chose Brother and selected the driver for the HL-1250 which was the recommended driver, and clicked Forward. That was it - working good.

Software
Most of what I spend time doing can be found right in a browser, so since Ubuntu came with Firefox 3, I was just a few short addons away from my normal working environment (mainly Gmail based workflows including Document creation and calendar). Del.icio.us plugin brought in all my links.
Pidgin connected me to my instant messaging systems (including work where I chatted with the guy with the broken laptop There were a number of programs that aren’t in the default Ubuntu install that I can’t live without, including qlGo, xchat, gftp, and a CVS client as good as LinCVS. An aside: I found tkcvs, but struggled for a long time with the right CVSROOT setup till I found :pserver:username:password@myserver.org:/usr/local/projects. Installed apache2, mysql-5, and php after I realized I couldn’t work on the sites I just downloaded from CVS! phpmyadmin was just as easy to install as everything else and made setting up my database driven sites easy. I had taken some pictures with my Canon S30 digital camera which was autodetected by F-Spot photo manager, and downloading the images and storing them away was as easy as Picasa is on Windows.

I’ll probably update this post again if I find some more time to play with this machine.

I’ve been using gnupg for years. It’s not that I have lots of super secret spy stuff to communicate to nefarious characters around the globe, but rather that when I send something personal, I’d like to keep it that way if possible. Like a lot of folks I have about a zillion passwords and authentication keys now to all the different websites and programs I use, online or otherwise. Like most people I used to keep all those passwords in a list in a text file someplace (usually right on my desktop!), but I decided a few years back to start encrypting it. I spend most of my day in a browser or in an email client, so years ago I chose to use gnupg and thunderbird with the enigmail add-on to sign and encrypt my email messages. I admit since I started using gmail more often I haven’t been signing my emails as much as I once did. The firefox extension firegpg allows you to use gnupg in gmail, but it’s just not as slick as enigmail is in Thunderbird so I eventually uninstalled it. Over the years I have accumulated a hoard of emails that can only be read when I click the nifty ‘decrypt’ button and enter my passphrase correctly. I’ve got lots of cool stuff squirreled away there including some emails from Wil Wheaton and other less well known folks on various arcane topics. Things worked so well for so long that I had totally forgotten that I set my key to expire in 5 years. Then, one day last week - it did!

I needed to shoot off a private, encrypted email and - no dice. Your key is expired!
This isn’t as earth shattering as it sounds. You can still access all your encrypted emails fine, but you can’t create any new ones with that key ever again. I’ve had several keys over the years for the different email addresses that I’ve had. They’re all tied together under one “umbrella” of sorts - they may be retired now but they’re still associated with each other, the history is all there. But when the key expires, that legacy is done. Everybody else in the world who had your key will eventually get flagged (as I understand it) that your key is dead and will likely remove it from their keychain. All those key-signing parties, all those trust settings!

I couldn’t just change the expiry date with gpg because it complained that my secret key was expired! What’s a man to do?

As it turns out, some nice person posted this information which describes the simple solution. LOL - setting your BIOS clock back to before your key expired allows you to operate on your key again and change the expiry date to some time later than the current date! Now why didn’t I think of that? (I was able to just change the OS date and time - and unplugged from the net so it wouldn’t update with ntp automatically)

After I came back to work after the holiday, there were several mysterious system failures to investigate. The solutions ranged from the ridiculously simple (it was unplugged) to the arcane (no display, no boot-up, no beep codes). Unfortunately this last was on a fairly important machine used for software development in the engineering department. It took just a few minutes to eliminate most of the possible causes for this problem (tightening cables, unplugging USB devices, reseating memory, etc.) and I soon had the machine booting into Windows but the system process was running at 99% and it was basically unusable. The user claimed that there had been a Windows update on the last day before holiday and I suspected that the machine had been turned off in a rush mid-update (or there was some virus at work). I certainly couldn’t prove either yet, and it didn’t matter. The user notoriously forgets to check in code and it was imperative that I get all data off immediately and over to a working machine so work could continue, plenty of time for post-mortem after the user was back in business.

Normally I would just dump the hard drive into another machine and copy the data over directly, but I didn’t have a machine that I could afford to bring down at the time. Also the drive was IDE, and I would have to find a machine I could plug it into easily since most machines I have are SCSI or SATA now. I’d have to use a CD Rom cable on those, and then I’d probably end up wasting time with cables and jumpers as well as potential problems in Windows recognizing it, or drive letters… I just wanted to just boot up from a Live CD and copy the stuff off immediately to the laptop the user was going to use until I had the problem fixed.

I’ve used Knoppix Live CDs for many years to rescue data in this manner, but for some reason I was having trouble with it this time. I didn’t have a nice ssh program (like Filezilla) on the laptop, so I had just started the samba service in Knoppix. I had gotten about a third way through a copy oepration and it crapped out. Also the transfer was taking forever. I didn’t know what was going on, I just needed the data and i wasn’t going to get into a tangent investigation. I was considering how long it would take me to put the drive in one of the machines when I remembered a post I had read on the LILUG’s mailing list over the holiday. I hadn’t ever tried the Trinity Rescue Kit, but a quick read about it suggested it was just the kind of thing I needed. Whereas Knoppix is a full blown OS on a disc, TRK is a lean set of utilities to do those jobs that admins need to do - and more frequently when you have a bunch of Windows machines around.

I had the machine booted up with TRK and serving up the hard drive on the network in just a few minutes. I copied off the entire drive to the laptop and put the user back to work shortly thereafter (after a short lecture about checking in more often

PS: It’s been some time since TRK saved the day, but I’ve tried TRK’s virus scanning abilities on a machine another co-worker brought in from home which was hobbled by viruses (why do they always blame their kids?) with less success - this is probably because I didn’t know what to expect and didn’t use it properly. The virus scanning took a *long* time, and eventually it found 90 virii on the machine. It supposedly saved these to a tarred gzipped file, but although the instructions said I would be given a change to delete these quarantined files, I was instead dumped to a command line. I didn’t know what to do next and shut down. The file was never saved to the disk, so I basically wasted time since they were all there upon reboot. TRK had helped identify the viruses, but not eliminate them. I ended up using some specific cleaning utilites targeted for the viruses involved, and then Windows didn’t work correctly anyway. TRK could have helped get the personal files off the machine, but these were properly backed up already anyway by the user, so that wasn’t needed. I’d really like to figure out how to use TRK to virus scan properly, though since my company has mandated no IE or Outlook (we use Firefox and Thunderbird now) we haven’t ever had a virus problem… well, once we did - when somebody decided they *had* to use IE anyway. That guy learned his lesson the hard way when I had to wipe the machine and he lost data

I’ve lost count of the number of times parents have asked me how to block internet use on their home PCs. Usually it’s because their kids have been caught visiting sites they shouldn’t have been allowed to go to and the completely clueless parents had no idea what their kids were doing until it was too late. Why anyone would allow their kids totally unsupervised acess on an internet connected machine is beyond me, but it happens. I hear some of you yelling, “Get to point Nate, what did you tell all those people who asked?!”

Usually the answer I gave varied depending on who was asking, but almost always it centered around dumping Windows entirely for the machine in question. This is usually NOT what people want to hear because they have a comfort level with Windows - they think of themselves as competent on a computer because they can use Windows and moving away from Windows woulld be to leave that comfort zone. But, if truth be known, it’s the best solution because it costs nothing to implement aside from a little time, allows you to repurpose an old machine (the usual desire anyway), and it offers a *far* more robust and secure platform than the old version of Windows 98 which is usually the only Windows that otherwise runs decently on the old box. (Linux is much leaner and runs really well on old hardware).

By the time I’ve mentioned Linux, the askers have already tuned me out, and have decided to go seek advice from their MCSE friends and family who will certainly offer saner advice than this crackpot Linux geek, but that’s all for the good since its one less person I have to support on my day off when their computer is frozen with viruses again. I don’t pretend to know what Windows solutions are the best choice to do your job as a parent for you and make sure your kids aren’t visiting hard core porn sites while you sip martini’s in the basement with your friends. I *do* know that you can get an awesome solution with free software pretty easily.

There are several things that matter here - is the computer shared? Is it shared between you and the kids, or do just the kids share it? Are the kids all the same age and have the same restrictions or do they differ? Do you want to block *everything* and allow just some sites, or allow *almost* everything and block just some bad sites you know of? Do you want to block just the web, or are you concerned about IRC, nntp, bit torrent, peer to peer sharing networks, VOIP / skype, email access… at this point most of the people I talk to are physically walking away. Most folks don’t know there is anything more to the internet than “www”.

In my case the rules are pretty much the same for both kids, so I’m not worrying about granting more access to the older one (yet). I use the machine occassionally, but largely it’s a ‘kids’ machine, set up in the center of the house where we can all see it - NOT in a kid’s room. The machine was bought new with Windows ME on it which was a dismal situation all told. Since I won’t use illegal copies of Windows, I installed the free ubuntu linux on this kid’s machine (an old PIII 800MHz Dell). They each have their own account with their own preferences and programs installed. If the machine was a laptop or had multiple interfaces (wireless + ethernet), I might have chosen another solution, but since the machine is a desktop with only one interface, I chose firestarter. Easily installed using ‘apt-get install firestarter’, it could just as easily be installed using the synaptic install tool. Configuring firestarter is very easy. Once you explain to someone what an inbound connection and an outbound one is, the rest is fairly straightforward. My kids are currently allowed to visit several internet sites including pbskids.org, noggin.com, and webkinz.com, but not much else. Therefore, I configured firestarter to run as ‘restrictive’ which means I must create a ‘whitelist’ of websites that they can access. Since the list is so short right now it’s no big deal, though I expect that this may become a hassle someday as they get older, but for now it’s perfectly acceptable from a management standpoint.

I spent too much time banging my head against Ubuntu’s vncserver so I could log into a remote desktop (just like Remote Desktop on Windows with Terminal Services), until I realized that I didn’t need to load the whole stinking desktop! That would mean sending the entire screen over anyway which is probably not as fast as just sending the program I need. Since OS X has it’s own X-Windows server, all I had to do was ssh into the Ubuntu box a little differently to “tunnel X”, and then launch the program on Ubuntu I want to run which then appears on my Mac’s desktop. Since I am really lazy, I just ran the entire gnome-panel.

ssh -X -c blowfish my_login@my_ubuntu_box gnome-panel