High Tech Sister

Feed History Problems With Google Reader & Feedburner

kj — Mon, 03 Nov 2008 16:27:19 +0000

Recently I began implementing a RSS feed for my webcomic AStoryADay.com. Creating an auto-updating RSS feed was fairly straightforward , and I could have easily left it at that. Visitors to the site could have simply clicked on the RSS link on my site and subscribed to the feed. However I wanted to keep some feed stats and metrics, and I felt that using Feedburner to manage and syndicate the feed was the easiest way to do this. All I had to do was point feedburner to my original feed and direct my feed link on the site to the new feedburner feed.

Everything seemed to going fine. The feed worked well in all the feed readers I tried except for Google Reader. Evidently google reader keeps all the entries for the lifetime of the feed. This is fine for the most part especially if the feed has been managed by only one person or organization. If I want to delete an entry from both the feed and google reader, I just have to reuse and update the GUID I assigned to it with either new or blank content.

However, it seems that my problems never have such simple answers. First off, apparently someone used the feedburner feed “AStoryADay”, in the past. It seems that the person stopped using the feed about a year ago and removed it from feedburner so the name was available again. But google reader still has a history of the old feed and after it lists my new entries it tacks on the old entries which I did not create or own.

This certainly makes the entries harder if not impossible to remove. So this leaves me with a few choices: (1) I could just forego using feedburner and just use my own feed without tracking it with feedburner, (2) Change the name of the feedburner url. Since I haven’t been a big user of feedburner stats in the past, I may just use my own feed and other tracking software. This would make sense because in all honestly I think it is more useful track newsletter statistics which is something I definitely want to implement. (3) My third option is to work to find a solution the problem without changing the name or feed url. This could be time-consuming and I don’t really want to really invest that much time in this small issue. So I probably won’t take this path.

On the other hand, I do think that Google is going to have to definitely give feed owners more leverage and flexibility in the management of their feeds as it pertains to google reader especially in the light of domain turnover rates.

Kim Jackson

The Case For Multiple Browser Support…(Or Not)

kj — Sat, 18 Oct 2008 16:41:27 +0000

Lately I’ve been discussing my recent dive into the world of web development– design, css/xhtml, and php coding. One of the projects I’m starting with, a webcomic, was mainly chosen to make the process more enjoyable. And to put things in perspective, I am just now completing the css/xhtml phase. It was during this phase that I ran into some issues concerning browser support that I suppose new and seasoned developers alike both face.

During the design phase I really didn’t give browser support all that much thought. In fact my chief concern during this development stage was keeping my Gimp layers separate to allow me to easily use the graphics in css/xhtml markup. My second most pressing concern was making sure I could reduce the size of the images while maintaining quality and improving browser load times. At first I was solely working with jpeg images but quickly turned to png files for some of the graphics due to its superior support for image transparencies. Going in, I had no idea this choice would affect which browsers would properly handle the look and feel of my site. Of course I knew that older browsers would have problems, but I naively thought no modern browser would encounter these same problems.

Needless to say, Internet Explorer is not your typical browser. I encountered problems with virtually every version of IE due to the image file types alone. IE just didn’t handle transparency correctly on a consistent basis. Also some of the css positioning which I used that was handled correctly by most gecko browsers didn’t work in IE. Firefox, while rock solid for the most part, also had a couple of problems. The main one being that Firefox 2 doesn’t totally implement the css statement inline-block. This is not a problem in FF3 or the latest versions of Opera.

Since I develop primarily using the latest open source technology, many of these issues didn’t surface until I booted into the windows side of my machine and started testing my design and implementation there. I also used the fantastic browsershots.org service to test how well the look of the site held up across upwards of 60 additional browsers.

The Implementation Dilemma

So now knowing where some of problems existed, I had to decide if I was going to fix all, some, or none of them. With IE there is a never ending list of hacks one can implement to overcome many of its shortcomings. Also I could have used gifs instead png files to overcome some of the image compatibility issues. However I decided against both of these solutions. Reason #1 is that this a personal project. If I were doing a project for a main stream business, I would probably implement some hacks to provide better IE support. Secondly, I like having clean css without mudding my stylesheet with ever-changing hacks. Finally, Internet Explorer 8 beta 2 works extremely well with the site. Microsoft seems to be going in the right direction with this version.

As far as Firefox goes, FF3 totally supports my design and implementation and version 2 only has problems with the inline-block statement I used. Firefox does have a workaround for this which I may use. I also could implement the feature in question in another way. But I’m leaning towards looking forward and mainly supporting FF3 and beyond. The latest version of Opera– fast becoming my favorite browser– also has excellent support.

The question then becomes: Do you support the old standards or strictly embrace the new ones? In this case I have opted for the latter. Will this alienate some possible visitors? Of course it will, but you can’t please everyone. I think as a site developer and content creator, you have to stay true to your vision for the site. The web is a huge place and people will come.

Kim Jackson

CSS Growing Pains….Sigh….

kj — Sat, 11 Oct 2008 19:36:33 +0000

I’ve been having CSS nightmares since I decided to venture into teaching myself web design. I wake up in cold sweats babbling about relative and absolute positioning. Learning how to deal with transparent png files and repositioning gifs makes my head hurt. The mention of things like elastic layouts makes me want to run the other way. No not really, but I did just finally figure out how css positioning works. Up until now I have been kind of stumbling through, writing css, and hoping it did what I wanted–not the best way to write any kind of code I must admit. Last night things just kind of fell into place. My aha moment came when I was trying to position several divs inside another div that had an absolute position. I’m a C++ and Java application programmer, but CSS really wore me out. It requires a completely different skill set. And on top of that it accounts for my 2nd shortest blog post ever: 168 words!

Kim Jackson

Linux, Seti@Home, And The Search For Extraterrestrial Intelligence

kj — Mon, 06 Oct 2008 18:15:50 +0000

I remember reading Octavia Butler’s book, Parable Of The Sower and particularly the quote:

“The destiny of Earthseed is to take root among the stars…beyond Mars. Other star systems. Living worlds.”

This passage in the book really resonated with me and took me back to the 3rd grade when I had an assignment to watch and document the phases of the moon everyday for a month. Ever since that assignment, I’ve been endlessly seduced by the sky and that which lies beyond it.

I was never too good at using a telescope, but that’s done nothing to hinder my endless fascination with the stars. I can’t help but to wonder what’s really out there. It’s strange to think that a lucky few (astronauts and civilians alike) have gotten to leave our atmosphere and gaze at space and the stars from such a different vantage point. Who knows, with commercialization of space travel on the rise, maybe I’ll get there one day. But until then I’ll contribute by giving my free cpu cycles to Seti@Home.

Seti@Home is a project aimed at finding signs of extraterrestrial intelligence using the collective processing power of computers around the world to analyze massive amounts of data gathered from radio telescopes. I first participated in the Seti@Home effort back in 2000 during my sophmore year in college. Back then I was still using windows as my primary operating system and didn’t know of any other Seti@Home client available at the time. Fast forward a few years and the Seti@Home project is still going strong and clients are available for a range of operating systems including linux.

Installing and Configuring Seti@Home

Seti@Home now uses the BOINC open source client application that is tailored to projects requiring volunteer and grid computing resources. Using the Boinc client you can not only connect to Seti@Home, but also other space related efforts such as Einstein@home and Milkyway@home. There are also a host of medical and mathematical projects too.

Download the Boinc client software from the Berkley Boinc Project.

Once downloaded, I created a directory called gridComputing and copied the the downloaded file to this directory. Then I ran the shell script at the terminal prompt using the following command: sh ./boinc_6.2.15_x86_64-pc-linux-gnu.sh
Running the script resulted in the extraction of the Boinc client files. Among these files is once named run_manager. If you double click on this file, you should be presented with the following dialog box. I just choose “run”.
Next you’ll be presented with a screen instructing you to attach to a project. Click Next.
On the “Choose a project” screen you can choose from a host of projects to join or enter the project url if you want to join a project not listed in the selection box. I chose SETI@home. Click Next.
If you are new to Seti@home or whatever project you are joining, you’ll have to create an account before attaching to the project. Click Next.
That’s it for setting up Seti@Home and the Boinc client. After you have logged in with your account information, you’ll be presented with the following screen.
If you want to specify parameters such as when to run seti, how much hard drive space to use, and how long it can run, click Advanced View and these changes can be specified.

Now sit back, relax, and bask in the knowledge that your would-be wasted cpu cycles are now being put to good use!

Kim Jackson

Dual Monitor Madness Part 1

kj — Wed, 01 Oct 2008 05:00:55 +0000

Over the course of the last couple of weeks, I’ve been working on a few projects and writing a ton of code. It was during one of my breaks — an increasingly rare occurrence these days — that I realized that I wasn’t taking advantage of the real estate that my dual monitors provide.

Back when I was solely using Windows XP as my operating system of choice, I frequently would change my monitor configurations from portrait to landscape depending on which display orientation was best suited for the task at hand. Yet, during the entire time I’ve been using Linux, Ubuntu more specifically, it’s never crossed my mind to try to rig up a similar setup until now.

What’s Wrong With My Current Setup?

During the last year or so that I’ve been using Ubuntu, I’ve been primarily using my monitors as one large widescreen display. This setup has worked well for the most part, and one of its chief benefits is being able to move windows between the two screens. However, since the X Windows system treats both monitors as one large display, I’ve ran into my share of problems here and there using some media applications. Applications such as vlc, mplayer, movie player, and kino segfault and crash when attempting to resize them or crash when loading media files that automatically call for a resize. It seems that these applications and the X system don’t interact well when working with large and irregular workspaces. This is part of the motivation behind me wanting to set up multiple monitor configurations. I simply miss watching movies while coding or drawing.

Is Manually Configuring Monitor Setups Necessary?

There are many ways to accomplish various monitor configurations in linux. I found the best method for my purposes was using multiple xorg.conf files and employing a combination of options such as Xinerama, RANDR, and NVIDIA’s Twinview. I have automated the process somewhat using scripts both during runtime and during the booting process to give myself maximum flexibility.

At the moment, other than RANDR (the X Windows Resize And Rotate extension protocol for dynamically changing displays), I don’t believe that there is any support for switching monitor display configurations in realtime without restarting the X server. I use the RANDR options in some of my xorg.conf files but couldn’t get it to work quite right in the terminal for realtime changes. However, I really don’t need realtime display changes; although it would be a nice feature to have, restarting X and cleaning up a few details is plenty fast for me. Other options would be to use multiple xsessions and/or different window managers. But those are topics for other articles.

New versions of the X windows system have just been released by the Xorg Foundation and are slated to be included in the next version of Ubuntu being released at the end of October. As the X system gets better and linux as a whole improves, the need for manually configuring monitor setups is increasingly diminishing and may eventually become obsolete. However that time has not arrived yet — at least not for me. And since I’m knee-deep in projects, I probably won’t be fully configuring my system with the new version of Ubuntu, Intrepid, until I get some down time during the holidays. Only then will I be able to assess whether I experience any improvements in my system on this front. I’m not expecting any 360 changes.

A Closer Look At My System

Before diving into all the technical details of my monitor configurations, I feel that it is important that I at least provide some of my system specs as any configuration is highly dependent on the hardware that is used.

My Setup:

Operating System: Ubuntu Hardy 8.04 64-Bit

Monitor 1: Dell 24″ Wide-Flat Panel (Digital)

Monitor 2: Acer 19″ Wide-Flat Panel (Analog)

Graphics Cards: 2 NVIDIA GeForce 6600 LE cards running in SLI bridge mode

System Memory: 4GB

Processor: AMD 64 Processor Running at approximately 3GHz

Graphics Driver: I use the latest restricted nvidia drivers

Rotating Display Arms: I ditched the display stands that came with my monitors long ago in favor of adjustable display arms that mount to my desk. This makes rotating and adjusting them a breeze.

A Look At My Display Modes

After assessing the options available for my system, the first order of business was to decide how many different monitor configurations I wanted. After some contemplation, I settled on six. Here’s a preview of each of them.

Wide-Span: This is my current configuration and is particularly good for drawing and design work. Both screens are in landscape mode and windows can be moved between the two of them.
Wide-Separate: In this configuration the monitors will both be in widescreen mode; however, the displays are independent of each other. Each one has its own set of workspaces and windows can’t be moved between the two screens. This makes it very easy to draw on one screen while using the other to watch a movie or surf the net.
Code-Span: I find widescreens very inefficient for any substantial coding. So this configuration puts both monitors in vertical (landscape) mode while still allowing windows to be freely moved between the two.
Code-Separate: The monitors are in the same positions as in Code-Span, however each monitor is independent of the other. I don’t find myself needing to move windows between screens much when coding, so this is my favorite configuration.
Draw-Code: In this configuration, the larger screen is in landscape mode and the smaller display is in portrait mode. Both screens are independent of each other. I like using Gimp on the widescreen and checking XHTML, CSS, and PHP in portrait mode.
Code-Draw: This is my second favorite mode. On days when the coding is not so intense, I like to code or write on the larger display while watching a movie on the smaller one. The large display is in portrait mode and the smaller one is in landscape mode. Again, both displays are independent of each other.

Summary

I’ve provided an overview of my motivations behind wanting to have a variety of monitor configurations available to me as well as closer look at my system specs and the 6 display modes I want to support. Also, I’ve touched on my chosen implementation method. In the coming articles I will flesh out in greater technical detail my exact configurations, the scripts I have automating the process, and expose some of the quirks and problems I encountered during the process.

Kim Jackson

Creating A Webcomic To Learn Web Programming

kj — Sat, 27 Sep 2008 14:16:59 +0000

Recently I’ve been thinking about web projects. Right now I’m working on one — redesigning my drawing site — and that’s going pretty well. When I hit my next milestone for it, I’ll update the status here. After doing the design in GIMP, I took some time to properly learn CSS and PHP. Up until now, I’ve been a hack at both of them — butchering code to get things working. I’m trying to accomplish more elegant solutions now.

Since the drawing site is undergoing a major overhaul, I wanted something a bit smaller and simpler to put on the web first. The first thought that came to mind was some type of small interactive web application. But although I’ve been an application programmer for a while, I don’t have much interest in web applications. So I settled on building a website consisting of dynamic database-driven webpages. It’s nothing new, but it’s a project that will get my feet wet for this whole web programming thing. And since I have quite a few drawings that can be used as a starting point for a webcomic, I decided to use this for the site’s content, thus making the project a webcomic.

I know that I could use wordpress and the comicpress theme for it. But I just didn’t want to go that route. Currently I use wordpress for quite a few of my sites, and it’s terrific software, but I wanted to try something different; create more of a learning experience. I love building things from scratch, and I’ve been itching to do this on the web.

I’m still sketching out the details, but I thing it’s going to be pretty simple for the most part. There probably won’t be support for things such as commenting and forums. I think these features are unnecessary and most of the webcomics I like don’t have them. Well it’s back to work for me!

Kim Jackson

Virtualization As An Alternative To Dual Booting Part 3

kj — Wed, 24 Sep 2008 20:29:03 +0000

Overview

Article 1 covers installing and configuring my virtualization platform of choice, the Sun xVM VirtualBox.

Article 2 covers installing Windows XP on VirtualBox and configuring system hardware and peripherals.

Article 3 covers the pros and cons of using a virtual machine as well as for which types of projects they are suited.

Okay, virtualBox has been installed and configured, and the windows xp virtual machine is running smoothly. Now it’s time to examine why I chose dual booting instead of virtualization. To quickly recap, I only need windows for an art class that I am taking. For this class I need use of my wacom intuos3 tablet and Photoshop CS3. So let’s first look at setting up the wacom tablet on the windows xp virtual machine.

Setting Up The Wacom

After setting up access to my usb devices in the previous article, installing and configuring the wacom tablet is a rather trivial task akin to following the driver installation steps for any non-virtualized windows operating system.

Start VirtualBox and launch the windows xp virtual machine.
Right click the usb icon on the bottom of the virtual machine and click on the tablet device.
Download the wacom driver or insert the wacom installation cd. Then double click the driver and begin the install.
The driver install may require a reboot.
Once windows xp restarts, the wacom should be installed and its configuration utiliities should be available from the start menu.

The wacom tablet works fairly well on the virtual machine. In fact, I only experienced problems when working with large gimp and photoshop files that contained many layers. So for most general purposes, using the wacom with photoshop and/or gimp in the virtual environment is an okay solution. However, since I mainly work with larger files, this was kind of a deal breaker for me. Let’s move on to my next challenge.

VirtualBox & Dual Monitor Support (Well Sort Of…)

I like using photoshop and gimp on my dual screen monitor setup. I use my main Dell 24” widescreen flat panel monitor as my main drawing area. My 19” widescreen flat panel serves as a utility screen where I keep my palette and other miscellaneous windows. Using the virtualbox graphical interface, I can launch a single virtual machine window and resize it to fit my larger monitor. But I wanted to know whether there was any way to use both monitors with virtualbox, so I read the virtualbox manual.

The bottom line is, virtualbox only supports multiple displays if you are using it as a headless server. Once you have your virtual machine running on a headless server, you can then use remote desktop software to log into the server. Since I have two displays, I needed to launch two instances of my remote desktop software, and each instance must state which monitor it is going to use else it will use the main monitor by default.

Although this seems pretty straightforward, implementing it caused me at least one headache! So here is my solution.

Configuring VirtualBox For Dual Monitor Support

First of all, the VirtualBox Remote Desktop Protocol(VRDP) must be enabled. To enable this, you go to the main virtualbox screen and click settings.
Once the settings box appears, click on Remote Display. Then click Enable VRDP Server. Also take note of the server port as it will be used later.
Now from the linux terminal, check to make sure VRDP is on with the following command: VBoxManage modifyvm VMNAME -vrdpmulticon on where VMNAME is the name of the windows virtual machine. In my case it would be winxp.
Enable multiple monitor support from the linux terminal with the following command: VBoxManage modifyvm VMNAME -monitorcount 2

Configuring Windows For Dual Monitor Support

After VRDP has been enabled, windows xp now needs to be configured to use dual monitors. Windows XP can only be configured for multiple monitor support if guest additions, which were discussed in the last article, are installed. Guest Additions provides an alternative video driver that allows windows to support multiple monitors in the virtual machine. I believe windows xp may be the only guest operating system for which virtualbox provides multiple monitor support although it may also do this for vista now. So assuming the guest additions package is installed, configuring dual monitors under windows xp is straightforward. You simply need to go to the display properties menu in windows and extend the display to the second monitor and choose a screen resolution for it. Apply the changes and shutdown the virtual machine.

Using Virtual Desktop Software To Launch Windows In Dual Display Mode

Ubuntu has remote desktop software installed by default. However it is terminal-based. I like to add a graphical frontend. You can install grdesktop via the the synaptics package manager.
Now we need to launch virtualbox in headless server mode. You’ll need to know the local ip address of your system. First launch a terminal and enter the following command based on your system configuration:

VBoxVRDP -startvm winxp 192.168.1.106

You should get the following screen
The windows xp virtual machine is now running but we need to remotely log into it to see the desktop. So launch 2 more linux terminals. Then use each terminal to launch an instance of grdesktop using the following command:

grdesktop
Configure each grdesktop client to connect to the headless server. To get dual screen support, you need to specify the screen number in the domain box. Put @1 in the domain box on the first client and @2 in the domain box on the second client. Change the operating system to windows xp using the dropdown box in each client.
Each client should open its specified windows desktop screen and you should now be able to work between them like dual monitors.

This approach is okay, but of course it’s not quite like having use of two physical displays. For one, each is in a windowed mode although with a little more configuration you could change to seamless mode. A second problem with this configuration is that there is a noticeable lag when doing anything more than basic tasks. VirtualBox can be configured to allocate more video memory and main memory, but the performance still won’t be the same as running the operating system on actual hardware because everything is going through a layer of abstraction. So even when loading gimp or photoshop using this dual window method, I noticed even more lag time than when using virtualbox in its regular mode. This was because the two instances of rdesktop were sharing the already limited video memory. As I said before, I could have changed this, but allocating more resources to the virtual machine would have started draining resources from the linux environment and slowed it down as well. This was the second deal breaker.

The third deal breaker was simply that virtualbox doesn’t support the virtualization of firewire devices. Linux video editing tools leave much to be desired, so one of the perks of dual booting with windows xp is that I can do better video editing.

Summary

VirtualBox is excellent when it is used for testing software, running office apps, or retouching or editing the occasional image in photoshop. However for seriously using media applications such as photoshop, I found that it was simply easier to dual boot. This allowed for better performance and less headaches. As an added perk, I can also use better video editing applications.

Kim Jackson

Virtualization As An Alternative To Dual Booting Part 2

kj — Mon, 22 Sep 2008 20:41:12 +0000

Overview

Article 1 covers installing and configuring my virtualization platform of choice, the Sun xVM VirtualBox.

Article 2 covers installing Windows XP on VirtualBox and configuring system hardware and peripherals.

Article 3 covers the pros and cons of using a virtual machine as well as for which types of projects they are suited.

Alright, it’s time to pick up where I left off in article 1. Now that virtualbox is installed and I have a virtual machine configured to store the windows operating system, it’s time to install windows.

Installing Windows XP on VirtualBox

So from the Sun xVM VirtualBox main menu, I highlight the virtual machine I want to configure and click settings.
Under settings I click the CD/DVD Rom option so I can mount the dvd drive from which windows will be loaded and click ok.
Now I’m back at the main virtualbox screen, and I click Start to launch the virtual machine and begin installing windows.
Now the windows operating system cd should boot and the windows setup installation should start.
The windows setup will detect the 10GB partition that was created to store the virtual machine.
I just format the partition using the Quick NTFS option.
The installation should coast along now.

Getting Control Of the Mouse

Now that Windows is installed and running, it becomes apparent that I don’t have control of the mouse and have to go through the initial configuration screens solely using the keyboard.

When I try to click anywhere in the windows virtual machine, I get the following message. However I still do not have control of the mouse until the Guest OS Additions are installed which provides the virtual machine with mouse integration support.

Installing Guest Additions

Once windows fully loads up, with the mouse, navigate to the virtual machine menu and click the Devices tab. Under this tab, select the install Guest Additions option.
Now I’m presented with the following setup screen. It’s a simple install. I just have to used the keyboard to install it as the mouse won’t be working inside the virtual machine until this software is installed. After the install, reboot the virtual machine. When it starts again, I have control of the mouse.

Now that the mouse is no longer an issue, I must now tackle the fact that I don’t have access to any of my usb peripherals. I don’t know if this is an issue on other linux distros, but it is with ubuntu 64bit hardy. This is basically because virtualBox expects to find the usb devices in a different location than the location where ubuntu stores them. Trying to right click on the usb icon at the bottom of the virtual machines, the usb devices may be visable but grayed-out. I also received some usb related error messages.

Getting USB Devices Working

The first thing I had to do was enable usb support on the virtual machine. From the the main window of the virtual machine, click on settings. Choose the usb option from the menu and enable usb and usb 2.0 controller support.
The next thing I had to do was go to the Users and Groups option in the administration tab located under the ubuntu system menu. From here I unlocked the menu using my admin password and clicked manage groups.
A list of groups appeared and I scrolled down until I saw vboxusers. I double-clicked on the group and found out its group id which I will need for later steps.
Next I had to edit the fstab file.

sudo gedit /etc/fstab
I added the following line at the bottom of the file to mount the usb file system(/proc/bus/usb) on the mount point usbfs where virtualbox expects to find it. The devgid should be changed to whatever group id was determined in step 1.

none /proc/bus/usb usbfs devgid=126,devmode=664 0 0
Now that I have a mount point established as usbfs, I need to actually set it up to mount automatically so virtualbox can use it. To do this, I start by editing the mountkernfs.sh file.

sudo gedit /etc/init.d/mountkernfs.sh
Add the line,

domount usbfs usbdevfs /proc/bus/usb -onoexec,nosuid,nodev,devgid=126,devmode=664

right above the line,

# Mount spufs, if Cell Broadband processor is detected

Save and close the file.
Next edit the mountdevsubsfs.sh file by uncommenting the following lines.

#mkdir -p /dev/bus/usb/.usbfs
#domount usbfs “” /dev/bus/usb/.usbfs -obusmode=0700,devmode=0600,listmode=0644
#ln -s .usbfs/devices /dev/bus/usb/devices
#mount –rbind /dev/bus/usb /proc/bus/usb

to read

mkdir -p /dev/bus/usb/.usbfs
domount usbfs “” /dev/bus/usb/.usbfs -obusmode=0700,devmode=0600,listmode=0644
ln -s .usbfs/devices /dev/bus/usb/devices
mount –rbind /dev/bus/usb /proc/bus/usb

Now save and close the file.
After all this is done, restart ubuntu. Start the windows virtual machine again. Now when right-clicking on the usb icon on the bottom of the virtual machine, I am now able to click on the usb devices I have available on my machine. After clicking on a given device, windows will either automatically install it or require additional software drivers to be downloaded to enable device functionality.

Summary

This article focused on installing the Windows XP operating system on the virtualization platform, getting the mouse working, and enabling usb connected devices. During this stage, I didn’t run into many installation issues. It was pretty straightforward. The focus of the next article will detail my particular experience with the wacom intuos3 tablet and performance issue that come up when using a virtual platform.

Check back soon for Part 3 of the series where I’ll be examining the performance issues that arise with using a virtual platform.

Kim Jackson

Virtualization As An Alternative To Dual Booting Part 1

kj — Fri, 19 Sep 2008 18:00:13 +0000

As a result of my last article series on dual booting, I’ve been asked quite a few times why I decided to dual boot my linux machine with windows xp instead of installing windows on a virtualization platform such as VM Server or VirtualBox. The short answer to this question is that I have used virtual machines in the past, but they do have their own set of limitations. However using virtual machines is a good option in many situations. So in response, I’ve decided to examine the process of using virtual machines and some of their perks and and shortcomings. Below is a quick overview of what I’m going to cover.

Overview

Article 1 covers installing and configuring my virtualization platform of choice, the Sun xVM VirtualBox.

Article 2 covers installing Windows XP on VirtualBox and configuring system hardware and peripherals.

Article 3 covers the pros and cons of using a virtual machine as well as for which types of projects they are suited.

What is a Virtual Machine?

In a nutshell a virtual machine usually refers to a computer software application that acts just like a particular hardware platform such as a pc or a mac. Using a virtual machine, you can run and test multiple operating systems or other software without actually installing them on your physical computer. This also makes it easier troubleshoot software without actually compromising your current system setup. In addition to this, virtual machines can also enable you to run software not designed for your hardware platform.

Some of the requirements and goals of virtual machines were first formalized in the early seventies by Gerald Popek and Robert Goldberg. Popek actually helped develop a prototype of an early virtual machine for the PDP-11 architecture. Implementing and coding virtual machines can be a complex process requiring intimate knowledge of both software and hardware. However, no matter how complicated the development process may be, virtual machines usually strive to meet three goals for their end users: portability, simplicity, and performance.

Installing VirtualBox On Ubuntu

There are a variety of virtualization options for linux including but not limited to VirtualBox and VM Ware. I prefer VirtualBox simply because it was the first one I used on Ubuntu. VirtualBox was originally created by Innotek but is now being developed by Sun Microsystems. The OSE(Open Source Edition) that comes packaged with Ubuntu doesn’t include support for usb devices so I don’t use that one. Rather I use binaries that are freely available from Sun. Here are the steps I followed to install VirtualBox on my 64bit Ubuntu Hardy box and setup an installation environment for Windows XP.

Download VirtualBox from here and launch the install.
If virtualbox was previously installed as it was in my case, you will be asked if you want to delete the old modules and setup files. It’s usually safe to do this especially if you don’t want to reuse anything from the old install.
VBox will then create a vboxusers group. Click forward.
Next you’ll be asked if you want to compile the vboxdrv kernel. Check the box and go forward. The install should complete.
Now for virtualbox to appear in the applications menu, you need to restart the system. In my case I had to completely shutdown my system and power it back up because I kept losing access to my mouse. This was probably due to a persistent state being held in memory somewhere. Nevertheless everything worked well after the complete system shutdown.
After launching virtualbox, I received a message stating that I had a virtual hard disk that wasn’t accessible. This only appeared because I had virtualbox installed previously. And even though I unistalled virtualbox and deleted the drive manually, I didn’t unregister it before I uninstalled virtualbox. This led to the new install finding a reference to it. So I clicked the check button.
Then I’m taken to a screen where I can remove the old virtual drive.
Now I get the normal virtualbox launch screen and here I can create a virtual machine to host my guest operating system (windows xp in this case although I could install just about any operating system). Click New.
The VM Wizard appears, click Next.
Now it’s time to create a name for the virtual machine and select an operating system. After this is done, click Next.
It’s time to allocate the base memory for the machine. This is the minimum amount the machine will be given when it is started although it can use more as necessary. I usually allocate 256MB but only 192MB is really needed.
After creating a virtual machine and allocating memory for it, I now have to create a virtual hard drive that will be used to store the machine. Select New.
The Virtual Disk Wizard Appears, Click Next.
Here we have the option of a dynamically expanding drive for the virtual machine or a fixed-sized drive. Personally I prefer fixed sized drives for virtual machines. Subconsciously I think I have latent fears of a rogue operating system taking over my system! But putting my delusions aside, it really doesn’t matter which option you choose although fixed sized is good if you have limited disk space. Now click Next.
From the Virtual Disk Location and Size screen under Image File Name, you can choose the physical drive where you want the virtual machine stored and how much space to give it. 10GB is the default and generally a good option to pick. Click Next.
Finally a summary screen is presented giving an overview of the virtual hard disk that is about to be created. If something isn’t right, it can be changed by going back. If everything is correct, click finish and the virtual machine hard disk creation will begin.
We’re not quite done yet; we’re back at the virtual hard disk screen, and the path to the machine we just created should be filled in by default. This machine will be used to boot the operating system that is going to be installed. Click Next.
A Summary screen is presented and if no information needs to be changed, click Finish.
Now virtualbox has a machine that can be used to load the windows xp operating system.

Summary

This article focused on installing the virtualization platform, VirtualBox, on a 64Bit Ubuntu Hardy machine. Also, a virtual machine environment was created that could support the installation of a windows operating system. The focus of the next article will be getting windows up and running on the virtual platform and some of the quirks the installation of peripherals presents.

Check back soon for Part 2 of the series where I’ll be installing windows on the virtual machine.

Kim Jackson

Media That Inspires Me: Ken Robinson’s TED Talk

kj — Wed, 17 Sep 2008 07:23:47 +0000

A few days ago when I was visiting my mom, my sister showed me a bulletin that she’d received about preparing to take the ACT test. I was kind of surprised at first seeing as she is only in the seventh grade. She’s not even in high school yet! She has tons of homework and on top of that, she is now studying for this test.

I immediately thought back to when I was in the seventh grade; I hadn’t even heard of the test at that age. Schools seem to have changed so much and also seem to be increasingly focused on test scores. I think I took the ACT for the first time in the ninth grade and a second time in the eleventh grade. I did well both times. However I don’t put much stock in those tests or any tests for that matter because about a quarter of the way through each time I became bored and started choosing random answers. I’ve never had much patience for such tests. So for me I guess, the test measured my aptitude and knowledge for picking random answers. Knowledge doesn’t necessarily equal success and vice versa.

I was thinking about all of this today when I came across Ken Robinson’s Ted Talk. He was presenting a refreshing presentation on reforming the educational system. It wasn’t a new idea but it was, as I said, refreshing to hear him add his view to the ever growing conversation. The general point that I took from the talk was that the educational system needs to pay attention to each child’s interests and method of learning to provide that child with a infrastructure to nurture the talent. Everyone has talent. He argues that if this was done, there would probably be fewer cases of attention deficit hyperactivity disorders and similar diagnosis.

I wholeheartedly agree with his observations. Indeed I believe if I had been in a highly regimented school focused on standardized tests like my sister, especially in my early years, I probably would have gotten lost in the cracks. But growing up in a very small village had its advantages and school was a very creative place for me. I know quite a few schools like this exist. I just wish they were the norm instead of the exception.

Nautilus Segmentation Fault With Blank CD

kj — Mon, 15 Sep 2008 19:53:48 +0000

I guess nautilus has had a bug that causes a segmentation fault when you try to run it as root while a blank CD/DVD is mounted for a while — at least in the Ubuntu Hardy 64 version I’m using. However, it’s been so long since I’ve needed to burn a cd/dvd that I didn’t notice this until yesterday when I had a blank cd in the drive. I was trying to run nautilus as root using a script I had written. However the new window with nautilus running as root didn’t appear. Immediately I thought I had hosed my script. So I went to the terminal and ran sudo nautilus and received a segmentation fault error.

Eventually I got around to removing the cd and found that it was the cause of my problems. After the cd was removed, nautilus ran fine. Without removing the cd, nautilus can be ran as root from the terminal using the command sudo dbus-launch nautilus. DBUS is an interprocess communication api. dbus-launch should start a dbus-dameon running in the background and allow nautilus to run as root.

How To: Dual Booting Linux & Windows On Separate Hard Drives Part 3

kj — Sat, 13 Sep 2008 23:57:36 +0000

Overview

Article 1 covers Introductory Information and Preparations necessary before You Install Any Additional Operating System.

Article 2 covers System information collection, Necessary Data backup, and Installing Windows.

Article 3 covers editing the boot information on each hard drive to make the Dual Booting Process seamless.

This article is part 3 of a series of three articles that detail how I got Ubuntu 64 Bit Hardy and Windows XP dual booting on my box while being installed on different hard drives. It differs from some dual booting methods in that I already had linux installed prior to installing windows. Just like anything else though, this method may not work for everyone as it is specific to my system configuration, and that plays a big part in setting up any software. But you may be able to find some bits that help you in your own undertaking.

Quick Overview Of the Boot Process

When the computer powers up, it performs a POST(Power-On Self Test) to initialize hardware and assess the system state. If everything goes well, the bios can look for an operating system to boot. It does this by searching the CMOS list of devices until it finds one that is bootable. When it finds that device, usually a hard drive that contains a boot loader located in the MBR(Master Boot Record), it enters stage1 of the boot loader. The stage1 boot loader(primary bootloader) then tries to find the secondary boot loader(stage2) by scanning the drive’s partition table looking for an active partition. If all goes well, after this the operating system is loaded.

Linux Boot Process

In linux, the first and second stage boot loading is performed by grub or similar boot loaders. Grub accesses the menu.lst file in the boot directory to see a list of the available kernels to load. You can then select your kernel preference or let the default load.

Windows Boot Process

Just like with linux, the bios finds the MBR on the hard disk. Then the NTLDR(Windows Bootloader) takes over. Windows needs to be on the root of an active partition — this is why it usually tries to overwrite other operating systems. Once the NTLDR takes over, it reads a series of files including boot.ini while preparing to load the kernel for the windows system. NTLDR can’t load the linux operating system, but it can provide a menu listing so that you can list other operating systems. This list can then be used to call a boot loader that understands how to boot the other operating systems.

Preparing Windows XP To Dual Boot

Boot into windows and edit the boot.ini file located on the root of the C: drive. If the boot.ini file is not viewable on C:, unhide protected operating system files
Open boot.ini and add the the line C:GRLDR=”Start Grub” to the end of the file. After WinGrub is installed, this line will call the GRLDR (grub loader) to initiate the linux boot process.

timeout=30
default=multi(0)disk(0)rdisk(0)partition(1)WINDOWS
[operating systems]
multi(0)disk(0)rdisk(0)partition(1)WINDOWS=”Microsoft Windows XP Professional” /noexecute=optin /fastdetect
C:GRLDR=”Start Grub”
Download and Install WinGrub. Installing with the default options is fine. Launch WinGrub. You might get some errors the first time you run it, but the errors didn’t adversely affect my system in any way and didn’t appear again after the first run.
Now under general options, click the profiles drop down box and select default. After selecting this, the Menu Table should display something like Windows at (hd0,0). This will be important when we go to edit the menu.lst file in ubuntu, so write this down. It basically says that windows is located on the first hard drive, and on that hard drive it is located on the first partition.
Now go back to the Profiles drop down box under general options and select dskboot. You should see some more information appear in the Menu Table.
In the Menu Table box, click on the first line which should read something like: . Some boot information should appear in the Title boxes. This information basically tells the system where to find grub and where to read grub’s menu.lst file. Now click the save button next to the profile box.
Next, click the Tools Menu and select Install Grub. The following screen should appear:
Now in the second box under Boot From, choose Boot.INI and click install.
Grub should now be installed on your windows system and the windows menu list should be configured. Reboot the computer.

You should see the following menu after the bios has transferred control to the NTLDR. If you choose windows, the windows operating system should load. If you choose START GRUB, the ubuntu menu should appear but the operating system probably won’t load quite right yet.

Preparing Ubuntu Linux To Dual Boot

Select Start Grub from the windows boot menu.
You should get a screen similar to the following. At this screen you want press c to go to the grub command line.
At the grub command line, you need to find out the hard drive mapping for your linux system by typing find /boot/grub/stage1 at the prompt. You should get something like (hd1,0). Write this down. Press Esc to get back to the ubuntu boot menu.
From the ubuntu boot menu, highlight the kernel you want to load (usually the first one), then press e to edit the selection. If the line that reads something like root (hd1,0) matches the mapping you found in the previous step, you are done with this step and can press the Esc key. If it doesn’t match, you must highlight the line and press e to edit it to match the mapping in step 3. Then press Esc until you get back to the ubuntu menu listing. If all goes well, you should be able to select the first kernel and Ubuntu should load.

Now if you had to change the drive mapping for ubuntu to load, you are going to have to edit the menu.lst file located in /boot/grub/menu.lst. The menu.lst file will also have to be edited if you want to have the option to boot into windows from the ubuntu menu. So once ubuntu loads, open the terminal and type the following: sudo gedit /boot/grub/menu.lst — Now if you had to change the drive mapping to boot ubuntu, you need to change it in this file to make the change permanent so you won’t have to manually edit the listing every time. I just changed the mapping for all my kernels so I could boot into any one of them.

Output:

## ## End Default Options ##

title		Ubuntu 8.04.1, kernel 2.6.24-21-generic
root		(hd1,0)
makeactive
kernel		/boot/vmlinuz-2.6.24-21-generic root=UUID=1fa7e8db-b92e-4f30-900f-d2ca3cb60d10 ro quiet splash
initrd		/boot/initrd.img-2.6.24-21-generic
quiet

title		Ubuntu 8.04.1, kernel 2.6.24-21-generic (recovery mode)
root		(hd1,0)
kernel		/boot/vmlinuz-2.6.24-21-generic root=UUID=1fa7e8db-b92e-4f30-900f-d2ca3cb60d10 ro single
initrd		/boot/initrd.img-2.6.24-21-generic

title		Ubuntu 8.04.1, kernel 2.6.24-20-generic
root		(hd1,0)
kernel		/boot/vmlinuz-2.6.24-20-generic root=UUID=1fa7e8db-b92e-4f30-900f-d2ca3cb60d10 ro quiet splash
initrd		/boot/initrd.img-2.6.24-20-generic
quiet

title		Ubuntu 8.04.1, kernel 2.6.24-20-generic (recovery mode)
root		(hd1,0)
kernel		/boot/vmlinuz-2.6.24-20-generic root=UUID=1fa7e8db-b92e-4f30-900f-d2ca3cb60d10 ro single
initrd		/boot/initrd.img-2.6.24-20-generic

title		Ubuntu 8.04.1, kernel 2.6.24-19-generic
root		(hd1,0)
kernel		/boot/vmlinuz-2.6.24-19-generic root=UUID=1fa7e8db-b92e-4f30-900f-d2ca3cb60d10 ro quiet splash
initrd		/boot/initrd.img-2.6.24-19-generic
quiet

Now to have Windows listed as a boot option in the Ubuntu menu, add this to the bottom of the menu.lst file. These lines just basically tell grub to transfer control to the NTLDR so windows can be loaded. Once you add these line, save the file. Then reboot to check that everything works.

title WinXP
rootnoverify (hd0,0)
makeactive
chainloader +1

title WinXP is how the option will appear in the menu listing.

rootnoverify (hd0,0) tells grub not to attempt to mount the partition.

makeactive means to make the partition given in the rootnoverify line active.

chainloader + 1 just lets you specify mutiple systems and keeps track of enumeration so that you can always boot into your latest linux kernel.

Summary

This was all I had to do to get Ubuntu 64 Hardy and Windows XP to properly dual boot on my machine. For me this approach was definitely easier that others I’ve tried in the past. So now I have windows loaded for my art class and ubuntu linux loaded for my primary operating system. The only other thing I may use windows for is video editing. Kino and some of the other linux video editing options are great, but it would be nice to have an application that is a bit more stable.