However, my experience with HAProxy is nill, so I’ve been putting this off for as long as possible. After reading the ServerFault blog and what they did with HAProxy I decided it was finally time to get my hands dirty.

Our scenario involves multiple windows servers, but HAProxy is frankly pretty shit under Windows/Cygwin (feel free to interpret that link however you will). So, *nix it is. I’m OK with that – I’ve got just enough experience in FreeBSD and Ubuntu server to get by, and Google is your friend. But that begs the question:

HAProxy on Ubuntu, or HAProxy on FreeBSD? I couldn’t find any real benchmarks comparing the two out in the wild, so I decided to do my own. Warning: From hereon in my stats are, to be honest, not going to be hugely accurate because I don’t have enough *nix knowledge to test all the proper metrics for this kind of exercise. I’m just hoping that if one of them is a clear winner, it wouldn’t matter if I honed in my statistics gathering. We’ll see.

The setup
Everything we do at work is virtualised, so it makes sense to make the tests virtualised.

The VMWare Host: I’m doing this on a development machine, which is a Quad Core 2.8Ghz, 8Gb of PC3200, RAID10 drives, in VMWare Workstation. The live environment will be a VMWare ESXi cluster, but this is as close as I can get.
Web Servers: These are what we will be load balancing. I have two identical Windows Server 2008 R2 web servers running a simple, no-frills HTML webpage in IIS7, straight out of the box, no tuning whatsoever.
HAProxy VMs Two vCPUs, 1024Mb of RAM, 20Gb fixed-size HDD. VMWare tools installed and updated.
FreeBSD: FreeBSD 8.0-released-amd64. This isn’t the very latest version of FreeBSD but it’s all I had lying around. “Minimal” install
Ubuntu: Ubuntu Server 10.4 amd64. No packages installed.
Jmeter: I’ll be using Apache’s JMeter to run real-world tests on the system. I want to simulate real-world usage for our servers. I’m not looking to absolutally max out the servers, mainly because I don’t have the ability to do that. I’m looking to simulate between 6000 and 8000 requests per second.

The process

Both of the VM’s are identical in specs. I’ve kep the load on the host as small as possible, to eliminate as many variables as possible. I will only be running three VM’s at a time (two web servers and one HAProxy). The connections will all be initiated from the host machine.

The first thing that I had forgotten about FreeBSD (especially with a Minimal install) is that installing VMWare tools is a bitch. Install the Perl package, install its dependancies, install the compatibility packages, answer these questions, blah blah blah. Installing VMWare tools on Ubuntu is as easy as pie.

I will be building HAProxy from scratch to ensure that the installations are as close as possible. The version of HAProxy in question is haproxy-1.4.8. This is where Ubuntu and FreeBSD were reversed. FreeBSD had everything needed to compile HAProxy immediately, whereas Ubutnu required GCC and libpcre++-dev. Swings and roundabouts…

What was immediately obvious is that the compile times were fast on both operating systems. Less than 3 seconds in both cases. Nice.

Thankfully the haproxy config file was identical for both HAProxy installations, because they weren’t using any os-specific options. It was as follows:

global
pidfile /var/run/haproxy.pid
daemon

defaults
mode http
contimeout 500
clitimeout 500
srvtimeout 500

frontend http-in
bind *:80
default_backend servers

backend servers
server web1 192.168.0.210:80 check
server web2 192.168.0.211:80 check
balance roundrobin
option httpchk HEAD /index.html HTTP/1.0

So. On to the results.

The results

(if you’re interested you might want to click through to the full-sized versions. Multiple the scale by 100x to get the real values).

They look pretty similar. So, let’s go and put them ontop of eachother:

The result – almost identical. Which goes to show – in real-world conditions (and if you’re getting 8,000 hits a second – lucky you) there is no discernable difference between FreeBSD and Ubuntu when it comes to running HAProxy.

The idea came to light about 6 ~ 8 months ago. Speaking to ‘Insanity’, he suggested – nay, encouraged – to write up a full outline of the what, when and how.

• What does the program do.
• When does it do it.
• How does it do it.

Or, otherwise known as a Functional Specification and a Technical Specification.

In researching ‘guides’, Joel Spolsky – famous for his various books and his site Joel on Software, I found this guide on Functional Specifications.
There are four parts to it, and I highly encoursage reading. It’s definitely helped me kick start my write up.

One catch is the need to have a Windows OS enabled device.
The creators have nicely created a couple of different versions – ranging from PDA / phones with Windows Mobile, as well as the ability to have it on any GPS device with a Windows CE Core.

Having purchased a Windows CE based GPS unit from Aldi, I found a few guides which would allow me to modify it so as to bypass the default ‘interface’ and to interact with the core OS.
Here’s a concise guide how to do it – and should only take you about 20 minutes.
Credit: I discovered the specific chipset used on the device from here.
There was another guide dealing with a similar device that outlines how to do this ‘hack’ here.
Warning: I provide this guide ‘as is’ with no warranties or guarantees. This worked for me – but there can be things that change in models, or other glitches that can happen along the way. If you happen to kill your device in the process, that is your fault – and have only yourself to blame.

You’ll need:

Step One
Having installed ActiveSync / Mobile Device Center, connect the PDA/GPS to the computer
Next, browse to the files / folders on it.

The manual actually tells you to backup the GPS software (Found under ResidentFlash\NNG). Do this if you want, but it’s not necessary for this hack.

Step Two
Fire up CE Reg Editor. Connect to your device, and you should be able to see the registry of it, just like you would on your own PC.
Note: For Windows 7, I had to put it into ‘compatability mode: XP SP 3′, and “Run as Administrator”.

Expand the HKEY_LOCAL_MACHINE folder. You’ll have to do a few modifications.
Note: The guide tells us to remap the My Documents, Application Data, Desktop & Programs. I’m fairly confident this isn’t important – but remap to the SD card if you’d like (means you have more space to work with).
The real key is the first change. IF you’re going to change the others as well, might as well do it all now.
To Change the value of the registry key, simply select the key on the right hand panel (eg. Launch50), and then change the text in the “Value name” field:

List of values to change:

1- HKEY_Local_Machine\init\Launch50
Old value: launch.exe
New value: explorer.exe

2- HKey_Local_Machine\System\Explorer\ShellFolders\Desktop
Old value: \Windows\Desktop
New value: \SDMMC\Desktop

3- Hkey_Local_Machine\System\Explorer\Shell Folders\Programs
Old value: \windows\programs
New value: \SDMMC\Programs

4- Hkey_Local_Machine\System\Explorer\Shell\My Documents
Old value: \My Documents
New value: \SDMMC\My Documents

5- Hkey_Local_Machine\System\Explorer\Shell\Application Data
Old value: \Application Data
New value: \SDMMC\Application Data
Note: There doesn’t seem to be an “Application Data” folder created in the \ root directory (where as there is the “My Documents” / “Desktop” / Programs”). I created one on the SDMMC just for good measure (in case “Application Data” is a hidden directory).

Once you’ve modified the registry, copy the folders from where they were originally ( \ root directory) to your SD Card.
Also, whilst you’re at it: Create a folder on the SD Card called “Ozi Explorer”. Copy the contents of the ZIP File into there.

At this point, we’ve updated the registry, and copied over the the Ozi Explorer program. Next, restart the system. You can either goto “Tools | Reset Device” within CeReg Editor, or you can just turn the device off and on again.
If all has gone well, you’ll be greeted with the warm, familiar Windows Explorer Interface:

From now on, every time you start up the device, it’ll boot into Windows CE.
This only presents a small problem with accessing the navigation programs: So, we’ll create some shortcuts to these programs on the desktop.
Remembering, we have no mouse or keyboard, so we have to improvise with a touch screen. There’s an ‘onscreen keyboard’ we will be using – accessed via the little keyboard icon in the bottom right of the screen.

On the top left corner is an icon called “My Device” – this is basically the equivalent of “My Computer”, the file explorer for Windows.
Double tapping that opens up the main screen:

Browse to the ResidentFlash drive – which is the system reserved partition for the GPS software.
Open NNG, and you’ll see the files pertaining to the GPS software:

Select NNGNAVI (with the blue ‘iGo’ icon).
Go to EDIT – COPY. Then EDIT – PASTE AS SHORTCUT.
Select the “Shortcut to NNGNAVI” file and go EDIT – CUT.
Close that Window, and when on the desktop, go to the onscreen keyboard, tapping first CTL and then V (Ctrl-V, universal shortcut for ‘paste’):

Similarly for Ozi Explorer; Go to My Device, Browse to the SDMMC drive.

Open up the ‘Ozi Explorer’ directory you created, and copied the files into. Select “OziExplorerCE”, EDIT – COPY, EDIT – PASTE AS SHORTCUT.
Select “Shortcut to OziExplorerCE” and EDIT – CUT; Go to the desktop, and paste (CTL then V).

You can do this with any program that you install onto the SDMMC card.
Note on installation of files
This particular Windows CE installation wipes any changes you make to the system – ignoring shortcuts on the desktop (remember how we moved the links to the desktop, etc. at the very beginning to the SD card? That’s so that the system wouldn’t wipe our hard work every time we shut the system down). This is because Windows CE is the same OS used on ‘Kiosk’ Terminal machines: Any changes attempted to be made to the system will be wiped / undone as soon as the system reboots (all the core OS files are stored on a ROM – Read Only Memory). Unless you’re logged in as ‘administrator mode’ which normally requiring a combination of key strokes to be pressed on boot – you can’t modify the core system settings or save it back to the ROM.
So make sure you install additional software to your SD card.

When you press the power button, the system will only go into hibernation mode; luckily, the system designers added a shutdown command on the desktop:

Double tape the icon marked “EXIT” and you’ll get a prompt saying “Shutdown?” with a big YES and NO to choose from.

There you have it – a more functional GPS unit, for only AU$129.

]]> http://equima.pfpfree.net/2010/aldi-go-cruise-4300-hack/feed/ 17 http://equima.pfpfree.net/2010/going-with-the-flow/ http://equima.pfpfree.net/2010/going-with-the-flow/#comments Wed, 24 Feb 2010 03:54:21 +0000 farseeker http://equima.pfpfree.net/?p=158 Several months ago I took a 2nd, part-time, after-hours job doing sysadmin work for a company that I have a close relationship with. A large part of their business infrastructure is a 10-blade Dell Bladecentre system. It’s a mean beast. 20 CPUs, over 40Gb of RAM, and is set up as a VMWare ESXi cluster, hosting about 30 heavilly-used virtual machines.

Managing this cluster means using the VMWare vSphere Client, which is fine. It’s a brilliant tool! Worth every penny. There’s just one thing that really, really, really bugs me, and after almost 6 months of using it every day, I’m still falling in the same trap.

Take a look at the following login screen for the vSphere client. Don’t linger, just take a glance. After all, it’s just a login screen:

Now, for someone who logs into dozens and dozens of things (apps/websites) every single day, a login screen is very standard. So what I see when I log in, is actually this:

Time to play spot the difference. Can’t be bothered? OK, here’s the big difference:

This difference is HUGELY important. Especially considering the vSphere domain is different to my local domain, so this option actually destroys my login. And it’s incredibly annoying, because:

1. The “Save my credentials” button goes on every single other login dialogue box I’ve ever used has been replaced by something that’s quite the opposite

2. After you’ve already typed your (in my case, very complicated) username and password

3. It deletes your username and password that you just typed in

And by the time I remember, it’s too late, I’ve already hit ENTER, and I have to wait 15 seconds for the login to fail just to be dumped back at the login prompt without any username or password.

What it SHOULD look like is:

This way, you’re doing things in a flow and order that makes sense, and nobody will ever wonder why vSphere wiped out their session credentials instead of saving them.

(just to be clear, not providing any option for saving your vSphere credentials is quite sensible. That’s not my issue here).

Originally, I wanted to talk highly about the ‘shift’ in the developers goals of creating mindless, violent First-Person Shooters where you go around killing anything that moves – to that of more ‘thought-inducing’ strategy based learning teaching games.

Back when I first started playing games, I wasn’t allowed to play Doom or Wolfenstein 3D – which were the original FPS’s of the time. Instead, I grew up playing games like The island of Dr Brain, and The castle of Dr Brain, along with The Incredible Machine, Hugo’s House of Horrors, Kings Quest, Day of the Tentacle and even Indiana Jones and the Last Crusade, (Sam and Max, and The Adventures of Willy Beamish also were in my fav’s list – which goes on and on….)

I’m not sure how many of these games you’ll recognise or know, but they were ‘classics’ in my mind, and bring a nostalgic flair to my memory. What was different about these games is that you had ‘puzzles’ you had to solve. Be it having to go between 4 or 5 different characters to get the rope you needed to swing across the chasm, or working out that you had to mix 6 different items in your ‘bottomless backpack of inventory items’ to create the one item you need to get that shiny ring.
It was all in the name of good fun – and a lot of humour, subtle or otherwise.

Recently, Portal provided a similar experience, one that was different from the ‘norm’. Sure, we had the 3D engines, and everything looks really schmick on screen – but that’s only 10% of what makes a game. A good storyline is critical. I know I always get criticism from people when I harp on about this, but I stand by my ‘guns’. Movies and Books are the same – I don’t enjoy many modern ones because like many modern games, they lack this crucial element.

Mirrors Edge was a game I thought had a unique game-play style, backed by a positive and solid story. It was only when I reached the end that I was quite disappointed; primarily due to the lack of ‘fulfillment’ of where the story ended up, and also because I wanted to do some more tricks; Like so many ‘chick flicks’ – the main issues are discovered, and the dilemma of the protagonist is related to the viewer. However, the resolution seems to come to quickly and swiftly. Perhaps it’s because they feel the audience has been left on the edge of their seats for long enough – or they just ran out of ‘intermediary’ people and sub-plots to place between the ‘plateau’ of the issues and the final point of resolution.

Another game I’ve started in recent months (and yet to finish) is Assassins Creed. This one is an interesting mix between FPS and a more puzzle-based game, with goals and quests – not always requiring killing people, sometimes you have strange tasks such as picking pockets, or collecting a set number of ‘flags’ within a time limit. The more interesting elements is the main story plot, and then sub plots within the characters. Scientifically speaking, it’s mostly none sense: But when has anyone let that get in the way of a good story? You get ‘sound’ in space on shows like Star Wars and Star Trek – but not in Firefly.

In comparison, games on Wii are different again – primarily due to the controls. I know, there are FPS’s and various other games on there, but some of the ‘Wii Fitness’ games are just plain fun! I love the Tennis and Golf that comes with Wii Sport – partially because it is a ‘sporting’ game, but also because it’s something different.

The future of gaming is interesting. Microsoft’s latest attachment for the Xbox-360, an interactive camera nicknamed Project Natal, should be an interesting shift in how games are played – and may even prompt a new interface for computing (though, that one is being actively targeted by touch-screen systems).

Apart from the controls, the environment is the next part that developers are working on – though that is harder; allowing users to fully interact with their environment (read: destroy everything they can) provides a challenge and a freedom – which can be fun at times, but definitely not primary to the main story or purpose of the game.

Let’s watch and see what the next 10 years of gaming will bring us to our respective gaming ‘console’ of choice!

Unfortunately, in the other spirit of open-source, the simple documentation is lacking a bit, so this blog post will hopefully help pave the way for people who might follow in my footsteps.

Intended Use

Basically, within one of the applications, I desired to have a link available in a menu that simply says “Report a problem”, which then submits a prefilled report to a bug tracking system including important information such as which version is being run, OS details etc.  My experience with Ubuntu has taught me that Apport is an ideal system that already supports the intended use-case, and is integrated into some existing apps in the same way that I’m looking for.

UbuntuOne Apport Configuration

The existing documentation that I referenced most heavily was found at https://wiki.ubuntu.com/Apport and https://wiki.ubuntu.com/Apport/DeveloperHowTo.  The first thing that struck me was that there seems to be no simple uses cases for what I thought would have been a reasonably commonly required use case, i.e., how to integrate with an app being released from a launchpad-hosted ppa.  The existing apport configuration makes apport-bug prevent reports being generated for non-Ubuntu packages (i.e., packages that are not in the main Ubuntu repositories).  This creates a hassle if releasing from a private or team ppa.

Fortunately, through browsing /etc/apport/ I found that the ubuntuone-client already uses apport with a non-genuine Ubuntu package.  This is done by placing a config file in /etc/apport/crashdb.conf.d, and an apport hook in /usr/share/apport/package-hooks/source_ubuntuone-client.py.

Basic Configuration

So, from all this, the most simple example I can come up with is the following.  To configure Apport to behave with an application, the following steps are required.

1. Configure the CrashDb

Place a CrashDb config file in /etc/apport/crashdb.conf.d called <package-name>-crashdb.conf (note: you can actually called it whatever you want… but for sanity’s sake I’d keep it as package-name ).  Inside this file place something similar to the following:

<package-name> = {
 'impl' : 'launchpad',
 'project' : '<package-name>',
 'bug_pattern_base' : None,
}

2. Install an apport hook to specify the CrashDb

Place a file in /usr/share/apport/package-hooks called <package-name>.py or source_<package-name>.py.  This file should contain the following, which simply  sets the CrashDb to use the correct configuration:

import apport

def add_info(report):
 """add report info"""

 if not apport.packaging.is_distro_package(report['Package'].split()[0]):
 report['ThirdParty'] = 'True'
 # replace <package-name> with the crashdb config defined in the previous step
 report['CrashDB'] = '<package-name>'

Testing the settings…

Assuming all has been correctly set up, you should now be able to report bugs against the custom package by entering the following command in a terminal (or using alt+f2):

ubuntu-bug <package-name>

This will open up a window allowing you to send a report.  If you click submit, you should be taken to launchpad, with a new bug being logged against the specified application.

One of the things I really wanted to demonstrate was the compatability mode, showing how old programs can be forced to run in a environment that pre-dates Windows 7.
It’s quite useful for older programs that worked fine under Windows XP, but seem to struggle / have issues (for whatever reason) under newer Operating Systems.
So, I hunted around my office, and found a copy of Lotus Smart Suite 9.6.
I figured that since this program came about in 2000/2001 it should break enough to require Compatability Mode.

However, sadly (or goodly, as it may be) it didn’t need to. When installing, I did have to run the setup.exe as an Administrator (because otherwise it couldn’t write to the registry). It prompted that there was an incompatability with Lotus 1-2-3 printing certain formats, but you clicked through that, and it installed OK. And then – it ran fine. No problems.
I even tried an old version of Winamp. That worked fine too.

I’ve only ever had to force Compatability Mode for playing older games, and they work fine under that.
I’m keen to find an actual program that WONT work in Compatability Mode, and that requires the Windows 7 Professional ‘XP Mode’ Virtual Machine.

So far, all I can say is that Windows 7 is freakin’ awesome.

I must warn, the sound quality is VERY dependant on your network speeds and CPU loads at either end.  If you are running a lot of programs on either the server or the client, the sound will be erratic.  See the bugs on the pulseaudio website.  I was unable to get the sound to run very smoothly for any length of time on between two Ubuntu VM’s on pulseaudio version 0.9.19-0ubuntu4.  The specific bugs are:

• http://www.pulseaudio.org/ticket/670
• http://www.pulseaudio.org/ticket/477

Either way, it is a fun exercise in pulseaudio configuration.  In order to demonstrate this I have set up two Ubuntu VM’s using VirtualBox on a local network.

Graphical Setup

The first step is to install the tools we’re going to use to configure this. Install PulseAudio Device Chooserm PulseAudio Volume Control and PulseAudio Preferences. Select these items from the Ubuntu Software Centre. You can easily reduce the list to these items by searching for pulseaudio in the search bar as shown in the following screenshot:

Screenshot of Ubuntu Software Center installing Pulseaudio

Once this is installed, you can launch it by selecting it from Applications menu.

Open the Pulseaudio Device Chooser

This will put an audio plug icon up near the clock. Left-click (or primary click) . Select “Configure local Sound Server” and change to the Multicast/RTP tab in the Preferences window:

PulseAudio Preferences

Server Computer

The first computer set-up will be the server. This is the computer you will be controlling & playing the sound from. On this computer you will need to open up rhythmbox or your chosen audio program that supports pulseaudio output. Totem or VLC also fit this prescription.

On the server computer you will want to “Enable Mutlicast/RTP sender”. If you want to play the sound out this server computer’s speakers, select “Send audio from local speakers”. If you want to be more specific about what is sent, I would recommend you select “Create separate audio device for Multicast/RTP”. For my setup, I want to be able to send only music to the stream when I want it, so I am going to chose the second option.

Client Computer(s)

The client computers will all need to be set up with the “Enable Multicast/RTP receiver” option enabled. Assuming the networks are working properly, you should immediately start hearing music on all the client computers!

Manual Configuration

If you don’t have the benefits of a running Xserver, or want to configure these settings to work on a headless machine, you will need to be manually configuring the pulseaudio config files. The pulseaudio config files are stored in /etc/pulse/system.pa if you’re running in daemon mode, and /etc/pulse/default.pa if you’re running in a per user setup (this is the default method in Ubuntu). I’m going to write this assuming you are using daemon-mode, so we will be editing /etc/pulse/system.pa.

Server Computer

To setup the server computer, you need to edit the configuration. In a terminal or on a console with admin access type the following:

sudo nano /etc/pulse/system.pa

In this file, find and uncomment the lines that look like:

load-module module-null-sink sink_name=rtp format=s16be channels=2 rate=44100 description="RTP Multicast Sink"
load-module module-rtp-send source=rtp.monitor

(If these don’t exist, just add them to the end of the file)

Once this is done, restart the pulseaudio server by running:

sudo /etc/init.d/pulseaudio restart

Client Computer(s)

The client computer will be done by editing /etc/pulse/system.pa (alternatively, you can copy this file to ~/pulse/default.pa). Edit this file by typing:

sudo nano /etc/pulse/system.pa

In this file, find and uncomment the lines that look like:

### Load the RTP reciever module (also configured via paprefs, see above)
#load-module module-rtp-recv

Again, if this doesn’t exist, just add them to the end of the file.

Once you’ve saved this, apply the changes the same way you did previously… by restarting pulseaudio:

sudo /etc/init.d/pulseaudio restart

Once again, assuming the network is behaving correctly, you should be hearing sound out both ends.

Troubleshooting & More Information

Troubleshooting guides are well written on the pulseaudio site. If you have any issues I would suggest reading the wonderful article at http://www.pulseaudio.org/wiki/HowToListenToTheRtpStream. The key thing is to check that you can see the broadcast packets (see the tcpdump section in the article).

Another good tip is to start the pulseaudio server in non-daemon mode. This will allow you to see and interact with the pulseaudio from the command line. To do this, stop any existing pulseaudio servers with the following command:

sudo /etc/init.d/pulseaudio stop # in system or daemon mode or

pulseaudio -k # for single user mode

To start pulseaudio server, run the following:

sudo pulseaudio --system -Cv # in system or daemon mode or

pulseaudio -Cv # for single user mode

Take note of any errors you see when starting this. Once the server is started, you can list the modules loaded by typing the following in the pulseaudio console:

list-modules

For more information on how to load modules see http://www.pulseaudio.org/wiki/Modules, and specifically http://www.pulseaudio.org/wiki/Modules#RTPSDPSAPTransport.

More information about the commands that can be run in the pulseaudio console, see http://www.pulseaudio.org/wiki/CLI

Windows 7 is fresh out in the wild.
Following close behind is Ubuntu 9.10 Karmic Koala.

What better oppertunity to
1. Do a clean install of Windows 7 and
2. Test the dual-booting ‘ease of setup’ inbuilt into Ubuntu.

I started with Windows 7. It installed in 15 minutes quite happily. One Criticism: It didn’t ‘automatically’ find and install drivers for my Broadcom PCI-Gigabit Network Card; but it was ok – I simply plugged in the onboard network  and Windows downloaded the drivers for the Gigabit Network Card in 30 seconds.

Next, I tried the “Wubi Installer“. I tried the default “Download and fetch Ubuntu from the internet” option. It started to download the AMD-64 bit version. Unbeknownst to me at the time, this was the typical 64 Bit version of Ubuntu, but because I read ‘AMD’ in the name, and I’m running an Intel system, I thought this could potentially lead to a bad outcome. I now understand that the Ubuntu labeling of this distribution as AMD-64 is merely a reference to the 64 Bit Extension set that was developed by AMD – and fully compatible with Intel systems. If I’d known that at the time, I probably would’ve proceeded with it.

As it was, I cancelled the auto download, and booted the system from the 9.10 Karmic bootable disc which I’d burnt earlier.

When it booted, I was faced with the standard questions:

I was after the second. After selecting it, it went through the usual motions of asking where you lived, time zones, keyboard type: But then the big, hairy one – How do you want to install Ubuntu?

There are a few options:

Now, here was where it became tricky, and this is only from a User Interface perspective – At the bottom is the slider to decided what percentage of the drive you want to allocate to the first partition (in my case, dedicated to Windows 7 use), but so far as I can tell, that has to do with the FIRST point Install them side by side, choosing between them at startup.
Perhaps it’s just me, but I would’ve put the slider under this particular option, so as to show a link between them; particularly since neither of the other two options have any relevance or association with the slider.

Insanity encouraged me to report it as a ‘bug’ – so I did. It’ll be interesting to see the comments, and what comes of it.

Once it had finished the installation and the computer rebooted, I was presented with a newer GRUB bootloader to what I have seen in the past. It’s up to GRUB-2 version 1.97 Beta 4.

Ubuntu is set to the default first boot (though this can be changed in the GRUB settings), along with the other alternate boot options, with Windows coming in at the bottom. When you first boot Windows after this installation process, it runs a CHKDSK to ensure the consistency of the system files, and disk drive. This is because when Linux does it’s partitioning, it messes with Windows mind (basically, Windows remembers the original size of the drive, and suddenly things have changed – so it makes sure that everything is still intact).

Edit: I replicated the install process and the setup on my laptop: To my surprise, Windows did not require a CHKDSK upon it’s first boot post Ubuntu install / GRUB writing to the boot record. I thought that was most interesting.

I’ll be following up with some reviews and thoughts on how the two systems behave, and a comparison / contrast of the two in the coming days. Stay tune!

One of the common panelists is John C Dvorak – and I generally like what he writes, he makes sense, and often is right – admittedly I relate to his cynical undertones.

However, recently he wrote an article on what would make the perfect tablet. I’m not too sure I agree.

My main reason is it all depends on what you want the tablet to do, or achieve.

John’s ‘list’ follows:

1. The machine should be about the thickness and size of a standard clipboard. It’s a size we are used to and a device this size would fit wherever a clipboard fits.
2. It should weigh about a pound. The weight of today’s Kindle e-book reader should be the same weight.

Agreed. The device does have to be relatively thinner than a notebook, and perhaps not as bulky. But until they find a better power source to batteries, I think this will remain a slight pipe dream. The truth is most notebooks are quite light – even with the battery inserted.

3. It should have a digital ink or MEMS display-technology screen. There are many versions of this so-called digital ink technology and there must be one that would be ideal for this application. This low-power technology does not have to be powered up to display. The image is permanent until it is powered up and erased. Most e-book readers will use this sort of technology since it has a high contrast ratio, thus increasing readability.

This one makes sense. You want a technology that is vivid and crisp enough to replicate print on paper, whilst ensuring longevity in battery life.

4. Multitouch screen. I do not think I have to explain why this is necessary. I’m sure Apple will feature this for sure.

I’m not entirely sure how this would be useful – maybe for that gimmicky ‘zoom in / zoom out’ trick with four fingers. Though, you don’t realistically need ‘multi-touch’.

Perhaps in some primary school applications, or medical fields it could be useful – but you can’t enforce this as a ‘standard’. So, back to my key argument: Depends on the application for which the tablet is being designed.

5. Massive I/O. As a PC user I like my machines to have a lot of I/O and connectivity options. This means Wi-Fi, WiMAX as well as all sorts of USB connectivity. If the machine is to be thin it may have to be done with a dongle, although I can see the device having a lip on one side with room for connectors. I hate losing dongles.

This one is a big wrong in my books. Excessive I/O may seem ‘cool’ and ‘handy’ – but how many people realistically need to have firewire, usb, audio in/out, external video, etc. on a device that has limited applications?
It makes designing a case complex, and makes designing the board even more complex- along with the additional cost factor for components to make the I/O work.

An alternative would be to have a narrow connector at one side of the device which ‘docks’ into a docking station. This docking station would allow for:
Charging, Connectivity (USB, FireWire, E-Sata,etc), External Video (HDMI, VGA / DVI), Gigabit LAN.
Though, that said, you might be able to get away with a mini-HDMI connector on the unit itself for presentations, or even an SDHC slot for storage of files.
Besides, with the way things are going, we may end up having a wireless ‘charging station’.

The unit itself could have Wireless Networking (N-Standard should be more than enough for most people), as well as BlueTooth connectivity.

6. Built-in GPS so the whole thing can be used as a car navigation system.

Interesting thought: but maybe not. A 12 or 15 inch display is hard to mount adequately for using as a guidance system. Stick to the 3 or 5 inch navigators.

7. Built-in Apple TV and full screen video-processing capability. It could be used as a small hi-def television in the kitchen, perhaps.

That’s just silly. At some point we’re forgetting the purpose of tablets.

A tablet, I see, should be used much the same way as the ‘Data-Pads’ in Star Trek. Replacing pen & paper with a digital, more re-usable form.

Modern tablet computers are wasted on gimmicky solutions. Let’s start thinking of real useful applications where these devices can be used – such as hospital’s to replace patients charts, where the records are being stored on a local server, rather than fragile paper.