1. Install Wine

To install Wine in Fedora, we will use yum. However, for other distributions, it should be widely available in your respective package manager or you can install from source.

[root@localhost ~]# yum install wine*

2. Install Photoshop

I installed Photoshop from the CD. I attempted to run the setup files from Nautilus, but it failed and I continued the rest of the installation from the command line. My CD was auto-mounted in the /media/Photoshop CS2/ folder so I will navigate there to open the setup file. Be sure not to run the installation as root.

[user@localhost ~]$ cd "/media/Photoshop CS2"
[user@localhost ~]$ wine Setup.exe

Now the normal Photoshop setup routine will initiate. I selected my region and continued. Next I was prompted for Gecko integration. This is a result of using Wine and is required by the Adobe license screen to display an HTML version of the license in the program window. I allowed it to install the integration, but it never completed. I canceled the integration and it only affected the next screen that asks me to agree to the license – by not displaying the text of the license. This is fine, so I selected Agree to continue on. From here on out you can continue with the installation as you would on a Windows machine.

3. Execution

Everything in the installation went smoothly and didn’t garner any more attention than had I installed it on a regular PC running Windows. To run Photoshop, I need to locate the executable and run it with Wine – just as we ran the setup executable with Wine. When using Wine, it will create a folder that represents the C: Drive in Windows. On a Fedora machine, this is /home/user/.wine/drive_c/ and contains a Program Files directory.

[user@localhost ~]$ cd "/home/user/.wine/drive_c/Program Files/Adobe/Adobe Photoshop CS2/"
[user@localhost ~]$ wine Photoshop.exe

This initiated and started Photoshop just fine. I messed around with a few images and was pleased with the installation. My next move was to create a menu shortcut for launching Photoshop. I did this and then ran it again to insure my menu entry worked right and was presented with the following error:

I found that this is an error that occurs in a Windows installation of Photoshop when there is an issue with the preferences. My current workaround is to wrap the Wine command to execute Photoshop in a shell script that will delete the preference file after Photoshop is closed. This means that any saved preferences will be lost each time Photoshop is used. Hopefully I can find out what is precisely going on here soon and come up with a more eloquent solution, but this will have to do in the meantime. Here is the contents of photoshop.sh which I store in my /home/user/bin folder for easy execution:

#!/bin/bash
wine "/home/user/.wine/drive_c/Program Files/Adobe/Adobe Photoshop CS2/Photoshop.exe"
rm "/home/user/.wine/drive_c/windows/profiles/path/Application Data/Adobe/\
Photoshop/9.0/Adobe Photoshop CS2 Settings/Adobe Photoshop CS2 Prefs.psp"

Be sure the set the script to executable after you write it with:

[user@localhost bin]$ chmod +x photoshop.sh

Now you can create a menu entry or launcher that executes this script.

A Patent Problem

Patenting software has only been well established in the United States since the early 1990s. There is some controversy to the legality of patenting software. It has been suggested that Microsoft began patenting any computer process they could. Being able to patent a process rather than actual code would bar other developers from writing their own code for the same process – and effectively destroy innovation.

Microsoft has thousands of such patents and has offered deals to certain Linux distributors in order to protect that distribution’s users from being sued. Novell, Linspire and Xandros have all bowed to Microsoft and teamed up with them – buckling under the weight of the fear, uncertainty and doubt that Microsoft is so fond of spreading.

Recently, Microsoft brought this offer to Red Hat – the most widespread Linux distribution in corporate America and the handler of the Fedora Project. Red Hat denied the deal. Red Hat director of corporate communications Leigh Day was quoted as saying: “We continue to believe that open source and the innovation it represents should not be subject to an unsubstantiated tax that lacks transparency.”

Popular user distribution Ubuntu has also denied the deal. Canonical’s CEO, Mark Shuttleworth, posted on his blog: “Allegations of ‘infringement of unspecified patents’ carry no weight whatsoever. We don’t think they have any legal merit, and they are no incentive for us to work with Microsoft on any of the wonderful things we could do together.”

I’m glad to see there are members of the community that will stand up to these bullies and fight for free software.

We will be creating a script that, when activated, will call up Firefox and send us to the composition section of Gmail. Then we will set this script up as the default e-mail client and send the target e-mail address to it. So, first we will put together a simple script and name it gmailto.sh. Here is what it should be comprised of:

#! /bin/sh

firefox "https://mail.google.com/mail?view=cm&tf=0&to=
`echo $1 | sed 's/mailto://'`"

I placed this script in my /home/user/bin folder as I do all of my personal scripts and then made it executable with:

[user@localhost bin]$ chmod u+x gmailto.sh

Now all we have to do is specify this script as our default mail handler. In Fedora, using Gnome, this menu is listed under System>Preferences>More Preferences>Preferred Applications.

It is important to place the %s after calling the script in the Preferred Applications window so that it places the target address into your new message.

The problem had always been that there was not a suitable program written for Linux. There were plenty of Xbox utilities for Linux that would extract or extract and upload an ISO, but nothing seemed to handle the repackaging task. I had tried to run C-Xbox Tool under the Windows emulator Wine, but to no avail. So, I just basically quit creating ISOs. Until this weekend when I had encountered the glitch that occurs in Star Wars Knights of the Old Republic II: The Sith Lords. There is a part in the very beginning that requires you to use the game disc to bypass the glitch. I have long since lost the game disc and needed to make another one. This is when I decided to see if I could get a different Windows Xbox utility to work – enter Qwix. This program is developed by the entity responsible for the Avalaunch dashboard for the Xbox. I was able to run this under Wine without any special parameters (see screenshot below) – but you cannot run it exactly the same as you would under Windows.

There were two issues that I had to work around in order to get this to work. First, I could not get the FTP to work in Qwix to get the game files. So I had to transfer them, before creating the ISO, using my regular FTP client – in this case gFTP. Second, I could not get Qwix to open the transferred game folder to create an ISO from it. My solution for this was to move the folder into Wine's simulated C: drive environment. There is a hidden folder under the user directory labeled .wine, and just below that is the drive_c folder. The contents of this directory will look familiar to Windows users. I copied the entire game folder into the drive_c folder and selected it on the Local folder line of the Source section of Qwix under the Create ISO tab from the left sidebar. It created my ISO pretty fast – considering the size – and I was able to bypass the glitch in the game.

I see five different areas of software that I use during development. I first, of course, need a development environment to generate my code. I need software to manipulate images that will be used in my sites. I need a program to encode videos into web friendly formats. After all of this, I need to get my files onto the web server. Lastly I need to be able to view the completed work to see how it turned out and determine where I need to tweak my code.

IDE

jEdit is my IDE of choice. I have tried others and could never find one the encompassed all of my needs. jEdit's robust module system allows enough expansion to cover almost any developers' needs. Most important of these is the FTP plugin for editing code directly on the server. The CSS auto complete feature is a nice plus too, though it was never a requirement that I had set out to find.

Image Manipulation

The GIMP is a no brainer here. It is the standard for image editing and creation in a Linux environment. A lot of people say that is severely handicapped in comparison to Photoshop. I am not a professional designer but have been able to do everything I have ever needed to do to an image with The GIMP. A legal version of Photoshop would cost me over $600, that's a cost I can't really justify for an individual – sans self-employed, professional designers. If you are into design that much, then you are probably a Mac elitist anyway – so why are you using Linux? I have found that I can almost always follow a good Photoshop tutorial I have found on the web and mimic it using The GIMP with equal results. I may have to add a couple more steps to the process, but that is fine considering I did not have to fork over any money to do it. Note: This argument does not acknowledge the option of Photoshop piracy.

Video Encoding

MPlayer's movie encoder is fast on the command line. I am able to take nearly any video format and then resize it and convert it into a Flash video. That makes it have a smaller filesize and also ensures that it will be in a format that is usable to most visitors.

File Transfer

I can edit my code directly on the server with jEdit, but I have to get my images, videos and whatever else I am delivering up on the server. gFTP is the only FTP client I have ever used in Linux. I have never found a reason to look elsewhere because it does everything I need. It allows bookmarking so I can connect to my sites with a click and get my files up.

Web Browser

Of course I use Firefox to test sites with. It is also my main browser for surfing the web. But Firefox is not representative of the users who will be visiting sites. Unfortunately most people still use Microsoft's travesty of a browser – Internet Explorer. Luckily I do not need to keep that OS available as an option on my machines just to use that broken browser. IEs4Linux is a script developed by Sérgio Lopes that enables the download and installation of IE 5, 5.5 and 6 through Wine, the Windows emulator. I do not use all of the versions of IE that he makes available, just IE 6 and 7. To get IE 7 you currently need to use the beta version of IEs4Linux.

Linux Professional Institute (LPI)

The Linux Professional Institute was founded in 1999 and is a non-profit organization with the goal of delivering vendor neutral Linux certifications for system administrators and programmers. It is a Canadian company whose exams are available in almost every country in the world. They provide 3 tiers of certification covering the scope from Junior to Senior Level Administration. The first two tiers consist of two exams each and the last tier consists of only one exam. Each exam currently costs around $150.

Junior Level Linux Professional (LPIC-1) The first tier consists of two exams, 101 and 102 – they are both required to obtain the certification. Exams are designed to verify that you can work at the Linux command line. They also test to see that you are capable of performing easy maintenance tasks: help out users, add users to a larger system, backup & restore, shutdown and reboot. You are also expected to be able to install and configure a workstation (including X) and connect it to a LAN, or a stand-alone PC via modem to the Internet. This tier requires that the candidate be familiar with two package management systems – RPM and Debian. According to their website, for an extra $100 you can become an Ubuntu Certified Professional on top of the LPIC-1 Certification.

Advanced Level Linux Professional (LPIC-2) Requires that you have already attained LPIC-1 status and you must two exams – 201 and 202. This tier focuses on administering a small to medium site and checks that you can plan, implement, maintain, keep consistent, secure and troubleshoot a small mixed (MS, Linux) network, including a:

• LAN server (samba)
• Internet Gateway (firewall, proxy, mail, news)
• Internet Server (webserver, FTP server)

Senior Level Linux Professional (LPIC-3) This is the newest tier that has just become available in 2007. You can choose either the Core or Specialty path. You must have LPIC-2 status and pass one exam to obtain the Core certification. The Specialty path requires that you pass the Core exam and also pass a respective Specialty exam. This tier assumes that you have several years of experience with installing and maintaining Linux on a number of computers for various purposes. It also requires knowledge of Bash or Perl scripting and experience utilizing LDAP. Though I could not find explicit pricing information on this tier's exams, it is assumed that too are around $150 apiece. There was no immediate information to indicate otherwise.

Novell, Inc.

This set of certifications is geared towards the mastery of Novell's Linux distribution – SUSE Linux Enterprise. They offer two approaches: Novell Certified Linux Professional and Novell Certified Linux Engineer.

Novell CLP This route is intended for people interested in being Linux administrators. It does not have any mandatory course work and only requires the passing of one exam – the Novell Practicum. This exam is scenario based and is meant to see if you can apply your knowledge to real world problems. Successful candidates are expected to possess the following skills:

• working with and creating scripts in the shell
• administering users, groups and file permissions
• administering virtualization with Xen
• configuring and using Samba
• configuring network connections

Novell CLE This route is intended for people seeking to master advanced SUSE Linux Enterprise Server administration skills. This certification requires two exams – you must pass the Novell CLP program first and then take one additional exam to attain the Engineer standing. This certification requires that you have a more in-depth understanding of networking services and security.

Each exam will cost approximately $195.

Red Hat, Inc.

Perhaps the most well known set of certifications is offered by Red Hat, Inc. Their certifications are geared towards the use of their Linux distribution – Red Hat Enterpise Linux. They offer four different certifications: Red Hat Certified Technician, Red Hat Certified Engineer, Red Hat Certified Architect and Red Hat Certified Security Specialist. Each of these certifications is valid until after the next major release of Red Hat Enterprise Linux. For example, if you are certified while RHEL 5 is current, then your certification is valid until RHEL 7 is released.

RHCT Their base level certification is intended for entry-level admins and focuses on the administration of a single system and attaching it to a network. There is only one test to pass in order to obtain this certification but is a 3 hour, hands-on lab exam administered at a Red Hat facility. It will cost approximately $350.

RHCE This is Red Hat's intermediate certification. This certification has been the closest to a standard in the Linux certification community. It is designed to test administrators with several years of experience and deals more extensively with services and security. While this is a great credential to have – it is pricey. At around $750, this certification may not prove a possibility for someone who is currently on the job hunt. Again, there is only one exam required.

RHCA It has been rumored that there are less than 100 IT professionals with RHCA status. This is not hard to believe due to the program's requirements. You must pass a 2 or 3 day performance-based live lab exam The fees for this program are very high. There is a total of 5 exams at $750 each, so you are looking at a minimum of $3750.

RHCSS This program consists of 3 exams, again at around $750 apiece. This level is meant to emphasize a proficiency in security. The focus of the exams is on using SELinux and Red Hat Directory Server to meet security requirements. This seems like a highly specific designation. Not all organizations are going to utilize this type of LDAP management and not all distributions are as eager to utilize SELinux as Red Hat is.

CompTIA

CompTIA is well known for their entry level certifications – most notably the A+ Certification. Their Linux certification is designed for people with between 6 and 12 months of Linux experience. It is vendor neutral and focuses on using the command line, user administration, file permissions and the management of Linux-based clients, servers and security. There is one exam to pass that costs about $200, but the certification is valid for life.

Which Certification?

Being that there is no set standard of certification in the market, you can likely use any of them effectively on your resume and during the interview process. The Red Hat certifications are going to be more geared towards someone who works for a company that is already utilizing that specific distribution and will likely be paid for by said company. As a Linux user who is looking to get into Linux administration as a profession, then you should really look at how much experience you have. If you are relatively inexperienced, then you could probably study a bit and obtain the Linux+ Certification without much effort. With more Linux confidence and experience you could climb the LPI tiers or go with Novell. Novell is strong here because they have had years at the top of the certification food chain and have great name recognition in the IT world. Their program also requires fewer exams to obtain their highest credential. Although Novell is not a vendor neutral certification, a competent employment candidate should have little problem explaining how the knowledge can be carried over to different distributions. The key in obtaining one of these certifications is being able to use it to back up claims of your Linux knowledge to a prospective employer and also give you a bit more confidence during the interview process.

I personally use MPlayer as my video player of choice in Linux. It has an encoding program call MEncoder that we will use to convert our video. MEncoder should convert any video that MPlayer will play. If you are using Fedora Core, as I do, then you can find both MPlayer and MEncoder in the Fedora repositories. First we need a video to start with. For the purpose of this example we will use this video. It is a short, 2.4MB video of Built To Spill front-man Doug Martsch doing an acoustic version of “Car” somewhere, I do not recall the specific details. The video is DivX 4 with MP3 audio and is 480 x 360 pixels at approximately 15 frames per second. The following command will convert it to the Flash video (FLV) format. NOTE The backslash at the end of each line is used to break up commands in the shell:

[user@localhost ~]$ mencoder testvid.avi \
-ofps 15 -o testvid.flv -of lavf \
-oac mp3lame -lameopts abr:br=64 -srate 22050 -ovc lavc \
-lavfopts i_certify_that_my_video_stream_does_not_use_b_frames -lavcopts \
vcodec=flv:keyint=50:vbitrate=300:mbd=2:mv0:trell:v4mv:cbp:last_pred=3 -vf scale 320:240

This produces a Flash video that is the same resolution as our source video. One thing you should look out for is matching the frame-rate of your source video. The section in the second line of the command, -ofps 15, tells MEncoder to make the new video have 15 frames per second. You can find your video's frame-rate easily with the file command:

[user@localhost ~]$ file testvid.avi
testvid.avi: RIFF (little-endian) data, AVI, 480 x 360, ~15 fps,
   video: DivX 4, audio: MPEG-1 Layer 3 (stereo, 44100 Hz)

We are also changing the output video's resolution, when we convert it, from 480 x 360 pixels to 320 x 240 pixels. You can omit the last line, -vf scale 320:240, if you do not wish to alter the resolution.

We are finished. You just need to upload your new FLV file to your server and integrate it into your site. I will not cover that process because it will be different for each user dependent upon your particular hosting service and/or CMS. We have managed to reduce the filesize of our video from 2.4MB down to 933.2KB. We have also managed to standardize our video so that it does not depend on the vast array of browser video plugins that our site visitors could be using – just Flash. Here's the video we have produced:

My girlfriend uses a desktop system that I had loaded with three hard drives. Two of the drives are full partitions of the ext3 filesystem and one is NTFS. There is a good reason for this – she was used to using Windows XP and I wanted to gradually ween her off of it. It was easier to do this when Linux was already available on the system via a dual-boot. She has been on Linux for a good five or six months now (at home, work is a different story). She emailed me today asking me to see if I could get a particular file off of the computer that she had not accessed since she was using Windows XP. I was already on the system via SSH and knew that I could get the file without rebooting into Windows or even getting up to go into that room. All that I needed to install was an NTFS reading kernel module. In Fedora, I would type (as root):

[root@localhost ~]# yum install kmod-ntfs

Upon completion of the installation, I needed to determine which drive the NTFS partition was on:

[root@localhost ~]# fdisk -ul
Disk /dev/hda: 20.0 GB, 20020396032 bytes
255 heads, 63 sectors/track, 2434 cylinders, total 39102336 sectors
Units = sectors of 1 * 512 = 512 bytes

   Device Boot      Start         End      Blocks   Id  System
/dev/hda1   *          63      208844      104391   83  Linux
/dev/hda2          208845    39102209    19446682+  8e  Linux LVM

Disk /dev/hdb: 40.0 GB, 40000000000 bytes
255 heads, 63 sectors/track, 4863 cylinders, total 78125000 sectors
Units = sectors of 1 * 512 = 512 bytes

   Device Boot      Start         End      Blocks   Id  System
/dev/hdb1   *          63    78124094    39062016   8e  Linux LVM

Disk /dev/hdd: 40.0 GB, 40000000000 bytes
255 heads, 63 sectors/track, 4863 cylinders, total 78125000 sectors
Units = sectors of 1 * 512 = 512 bytes

   Device Boot      Start         End      Blocks   Id  System
/dev/hdd1   *          63    78108029    39053983+   7  HPFS/NTFS

From the output, I could tell that /dev/hdd was my NTFS drive. To access the drive immediately I would need to mount it. I first create a directory as my mount point and then issue the mount command:

[root@localhost ~]# mkdir /media/drive_c
[root@localhost ~]# mount /dev/hdd1 \\
   /media/c_drive -t ntfs  -rw -o umask=0000

It is likely that she is going to want to access this drive again in the future, so I appended the following line to /etc/fstab to enable it each the system boots:

/dev/hdd1       /media/c_drive  ntfs    rw,defaults,umask=0000 0 0

The options I have used here allow for full read/write access to the drive, so if you intend to boot into Windows again, then be careful with what you delete from that partition – Linux will not warn you about removing critical system files from your Windows partition.

I have used several bootable operating systems in the past: Knoppix, Damn Small Linux, and Puppy Linux. None of these really did much for me. I would boot into them and then I was merely stuck with an environment that was crippled in comparison to my usual environment. I can see how they might be applicable to users who are considering the switch to Linux, but as a regular user it provided little value. I noticed that the only time I was really booting into a live environment and using it was with the excellent SystemRescueCd. It is a Gentoo Linux distribution that is comprised of many useful system utilities – and it had a very small footprint at around 110MB. Here is how I put it on my 512MB USB thumb drive as a bootable operating system.

Before starting, you need to back up the data that already exists on the drive because this process will entirely wipe the drive out.

While the drive was plugged in I had to find out what address it was using. There are several ways to do this, if the drive was being used before it would have been partitioned:

[user@localhost ~]$ df

This will return the available partitions – we are looking for something in the /dev/sd? range. If there is only one USB drive plugged it, then it is most likely /dev/sda. If the drive does not contain any partitions yet, then try this:

[user@localhost ~]$ ls /dev/sd*

Now that we have the address, we need to first remove any partitions. This needs to be done as the root user and the drive should not be mounted:

[root@localhost ~]# fdisk /dev/sda

Press “p” to list the current partitions. Then press “d” to delete each one, you may be prompted to enter the partition number. Press “w” to write the changes to the disk and exit. Now that the partitions are removed we are going to wipe the drive out. Warning! This will absolutely remove everything on the drive! This process will take some time since it is writing zeroes over every bit of the drive.

[root@localhost ~]# dd if=/dev/zero of=/dev/sda

Now the drive is clean and we can set it up. First we need to put an MBR (Master Boot Record) on the drive. I am not aware of a program in Fedora that does this, so I installed ms-sys. If you are on a Debian based system, then you can probably use the install-mbr command. Once installed we will create a Windows 95 boot record on the drive:

[root@localhost ~]# ms-sys--mbr95b /dev/sda
Windows 95B/98/98SE/ME master boot record successfully written to /dev/sda

We will now create one partition to house the SystemRescueCD system and another to use as regular storage. We invoke fdisk again:

[root@localhost ~]# fdisk /dev/sda

This time we will press “n” to create a new partition. Then “p” for primary partition and “1″ to signify it as the first partition. Enter “1″ again to select the first cylinder. Next it requests the last cylinder, I entered “+150M” to grow the partition by 150MB. We need to change the type of partition by pressing “t” and then “e” to select W95 FAT16 (LBA). We need to make this partition bootable, so press “a” and then “1″ again to select the first partition. You can create whatever partition you like with the leftover space. Once done, press “w” to write the changes and exit.

Next, we format the partition:

[root@localhost ~]# mkdosfs /dev/sda1

And then we will use SYSLINUX as a bootloader for the distribution.

[root@localhost ~]# syslinux /dev/sda1

At this point, you may want to reboot into the USB drive to see that it is working and set up correctly. If it is, then you should see a message similar to this:

SYSLINUX 3.11 2005-09-02 EBIOS Copyright (C) 1994-2005 H. Peter Anvin
Could not find kernel image: linux
boot:

If it did not work, then you should see something along the lines of “boot failed”, “Missing operating system” or “Boot error”. If you see any of these, you should try repeating the above processes. If it just boots into your regular system, then you will need to tweak a setting in your BIOS to boot to the USB device. If everything went well, then it is time to transfer the system files. You need to boot back into your regular system and mount your new thumb drive partition:

[root@localhost ~]# mkdir /mnt/usb
[root@localhost ~]# mount -t vfat /dev/sda1 /mnt/usb

You will next need to download the disk image from their download page and mount it:

[root@localhost ~]# mkdir /mnt/rescue
[root@localhost ~]# mount -o loop systemrescuecd-x86-x.x.x.iso /mnt/rescue

And then copy the files over:

[root@localhost ~]# cp /mnt/rescue/syslinux/syslinux.cfg /mnt/usb
[root@localhost ~]# cp /mnt/rescue/sysrcd.dat /mnt/usb
[root@localhost ~]# cp /mnt/rescue/isolinux/* /mnt/usb
[root@localhost ~]# cp /mnt/rescue/bootdisk/* /mnt/usb

That is it. The drive should be ready now. When you reboot you should be prompted by the SystemRescueCd boot screen.

• Kernel 2.6.20-1.2925.fc6
• ATI Radeon 200M – 128MB video card
• ATI driver 8.34.8

I referenced this post at Fedora Forum and acquired some recent source RPMs so that I could build the appropriate x86_64 RPMs for the Xgl server and a library it is dependent upon that is not available in the repositories. The forum link only supplies the x86 versions. I have archived the x86_64 versions of the RPMs here. This post assumes that you already have your ATI card working with full direct rendering using the fglrx driver. First, you need to install some package dependencies. You may or may not need these depending on your current setup. These are the packages I had to install in order to create the RPMs:

[root@localhost ~]# yum install libXfont-devel libXrandr-devel libXrender-devel
libXt-devel libfontenc-devel libdrm-devel libXaw-devel xorg-x11-xtrans-devel
libxkbfile-devel libXext-devel libXxf86misc-devel xorg-x11-util-macros libXtst-devel
libXres-devel glitz-glx-devel libdmx-devel mesa-source libselinux-devel libtool
libXinerama-devel libXxf86vm-devel

Now we need to decompress the tarball from above:

[root@localhost ~]# tar zxf xglrpmsx86_64tar.gz

And install the RPMs:

[root@localhost ~]# rpm -ivh *rpm

If everything went well, you can now install your composite window manager of choice. For Compiz, we first want to make sure there are not any latent packages installed, then we will install the new package:

[root@localhost ~]# rpm -e --nodeps compiz
error: package compiz is not installed
[root@localhost ~]# yum install compiz

For Beryl, my personal choice, run either:

[root@localhost ~]# yum install beryl-gnome

or

[root@localhost ~]# yum install beryl-kde

The above command are dependent upon which desktop environment you use. You can now logout and you should find an Xgl session available for whichever desktop environment you use. Log in to your Xgl session and start the new window manager. For Compiz, type:

[user@localhost ~]$ desktop-effects

And for Beryl, run:

[user@localhost ~]$ beryl-manager

After I confirmed that everything was running well, I added the window manager command to the startup programs. The GUI for this is available in Gnome at System->Preferences->More Preferences->Sessions, under the Startup Programs tab.

I also added the following snippet to my xorg.conf file in the Module section:

SubSection "extmod"
    Option "omit xfree86-dga"
EndSubSection