Ardent Meridian

As a professional software engineer/architect I make frequent use of all three of today’s primary PC operating systems – Windows, OS X and Linux on a regular basis.

To this point, each of them has fulfilled a specific role, or been required for a particular purpose.  And, while I do categorize and view them differently, I have no religious preference for any of them.

They are facilitators, environments, for running other software and tools which I use in both my personal and professional life.

For a good while now I have been wondering just how well each might fulfill the role of my primary desktop operating system.

I do not, necessarily, believe there is one single environment that can meet all of the needs I currently have of my primary desktop, but I am interested in finding out which might be the most generally applicable and, perhaps, most pleasurable to work with.

I spend a great deal of time working at my computer, so if there are things I can do that make the experience either more productive, or more enjoyable, then they are worth evaluating in a frank and honest manner.

That is the purpose of this experiment; to spend a week each with Windows, OS X and Linux as my primary desktop operating system and see how well each works and what is good and bad about my experiences with each.

General Approach:

This first posting merely sets the stage and terms for the experiment to come.  It will be followed by accounts of the actual setup and provisioning of each system.  Then it will progress to cover the salient parts, good and bad, of each of three weeks spent with a different primary operating system.  Finally I will summarize my thoughts on the experiment as a whole, and document any changes in my environment and day-to-day computing environment that result from this process.

Since I am a professional software developer, and have been for almost 30 years, I know my way around computers, and am comfortable dealing with whatever technical issues crop up.  I will document them, they may irritate me to a greater or lesser degree, but I will work around what comes up since I have the capacity to do so.

Obvious Concerns & Biases:

To begin with, it should be noted that, today, my standard environment is Windows, specifically Vista x64 SP2.  This is less out of any specific preference, and more as a result of being, over the years, the platform I have done the most professional development work on and for.

As a result the bulk of my software purchases are either Windows specific applications, or have been the Window’s versions of licenses of multi-platform software.  Some of these applications are not available on other platforms and/or there is no functional equivalent.  Since this would be true if coming from any of the other systems as well, the easiest way to eliminate this as a variable is to make use of VMware to allow me to continue to use those must-have applications in any of the systems under evaluation here.

I actually host my primary development environments, which tend to be fairly fluid, in VMware images anyway, so I am more than comfortable with the available performance that can be had running Visio or Visual Studio or SQL Server 20008 in a virtual machine as a day-to-day production-development environment.  I do not see this as a negative, and do it under Windows already (it is extremely convenient for isolating specific configurations and tool sets, as well as being easy to restore in the event of a catastrophic failure).

Specific Software Issues:

This experiment is not designed to compare anything but the operating systems, and the experiences they facilitate.  So, for example, this is not going to turn into a Microsoft Office vs. OpenOffice.org debate.

I will go on record right now as saying that the only circumstances I would ever consider running OpenOffice.org would be if I could not afford the licenses for Microsoft Office.  That is a personal preference, and provided I can run Microsoft Office either through API emulation, a VM environment or as a host-native application I have no issue with that.

I will detail the precise set of must-have and alternative software as I post my experiences with the system builds for each operating system.

Hardware Environment:

The goal here is for me to see what might now work better, or be more enjoyable (assuming at least equal productivity) to use, as primary operating system for me.

To be completely fair to each system here I will be running them directly on a nice, modern, well equipped and powerful PC.

That is how they would be used as my primary O.S. so relegating them to a VM and then complaining about performance is not useful to me in this experiment.  While running an O.S. in a VM is perfectly valid, such an environment is hard to look at as one’s primary desktop operating system.

So, the machine in question is an early-2008 model Apple Mac Pro.  The boot-drive for this machine will be replaced with specific drives for Windows, OS X and Linux installations.  It’s basic specification is as follows:

• Dual Quad-Core 2.8Ghz Xeon Processors (8 physical cores total)
• 16 GB RAM
• 4x 1.5 TB Seagate Barracuda 7200 RPM hard-drives (1x system volume, 3x storage).
• Dual ATI Radeon HD 2600 XT Video Cards
• Apple 30” Cinema Display Monitor (2560×1600)
• 2x Dell 2007FP Monitors (1600×1200, in portrait orientation).

In addition to the basic PC itself, I use two printers as part of my work:

• HP Color LaserJet 2500n
• Epson Artisan 800

Finally I use four cameras and two portable music players on a daily/bi-daily basis:

• Canon EOS 1D-Mk2, EOS 40D and G10 and a Panasonic DMC-LX35
• iPhone 3G and iPod Classic

All of these devices must work properly, though I have no particular issue with their support software, where required, being hosted in a VM if no native equivalent exists – unless such a hosting environment renders them less productive or reduces the available features.  But then improving overall productivity and enjoyment of my environment are the principal goals of this experiment.

Test Approach:

I am going to spend two days per system setting them up.  In my opinion, if it takes more than that for a professional with experience in all three operating systems to configure them for productive day-to-day use, then something is broken or missing or worse.

This two-day period is to include installations of the O.S., any required virtualization environment, and configuring all of my general hardware, drivers and software.

It will necessarily be a little different for each system.  There is no native build of Office for Linux for example, though I will be using Office 2008 for the OS X.

Depending on the host operating system different applications/operating systems will need to be available in a virtual environment.  I will not count the time required to install applications into the VMs themselves, just the time needed to setup the VM software itself.

And once the setup is complete, the real testing can begin.

As previously stated, my goal here is to see if switching my primary desktop operating system will bring about any productivity benefits, or if productivity is break-even, find an environment that suits the way I work better or is more enjoyable to work in.  And I will document my typical weekly usage as part of the individual week’s usage for each system.

From an objective standpoint, I am looking for increases in overall productivity.  If, for example, Linux is slower for Office tasks, but faster for development tasks, and the net result over a standard week’s work is that is a net positive, then that is a good thing.  The only caveat there being that if is something is slower to the point of being frustrating to interact with, that will be seen as a negative – though this is not something I feel is likely.

This is something I am doing for my own benefit.  Regardless of the outcome, the point is not to use it to bash one system over another, just to find what works best for me, on good hardware, based on my typical weekly usage and computing needs, both professional and personal. 

I am posting this here as I feel it might be interesting and/or useful reading for others, especially those in a similar position.  It is not intended to settle any religious debates, sway any preferences or proclaim general superiority of any one system over another:

I simply do not care what anyone else uses as their primary O.S. – variety is, as they say, the spice of life, and if you are using something other than which is either the most productive, the most enjoyable, or most inline with your personal philosophies, then you are not being rational.  Those will almost certainly differ from mine, though I will not be shy about stating what those are for myself.

Next Steps:

Next up will be my detailing the builds for each of the systems under test, followed by the experiences for each week using each system as my primary O.S., and then a general summary.

Post a Comment