All men's miseries derive from not being able to sit in a quiet room alone.
Blaise Pascal
In some interviews there comes a point where you realize that you don't want the job. It might be the moment you discover that the employer has conveniently omitted from the published job description the requirement for the incumbent tospend 50% of their time maintaining a one million line legacy application, written in Visual Basic. It may be shortly after you state your salary expectation, only to be greeted with a look of blank astonishment. For me, it is often the point at which the interviewer reaches into their bag of interview clichés and asks a question so trite that it betrays the total absence of advance preparation and original thought. Once the role has been safely relegated to the "no thanks" pile, it is difficult to resist adopting a certain playfulness while waiting out the duration of the interview, as courtesy demands.
For example, when asked "Where do you see yourself in five years time?" I like to borrow a witticism from comedian Steven Wright, and respond "I don't know - I don't have any special powers like that." If asked "Why are manhole covers round?" I might reply "Because God made them that way", simply to see if they will dare broach a topic traditionally considered taboo in interviews. And if they should enquire "What are your career goals?" I will almost certainly reply "I have only one - I want a door."
But in this last I'm only partially being facetious, for one of the most consistently difficult aspects of every software development effort I've been a part of has been the physical environment in which it is conducted. Having abandoned the lofty career goals of my youth (such as producing quality software) I have deliberately set my sights a little lower. These days, my sole ambition is to have an office with a door. My professional nirvana would then be to close that door, so I can get on with my work undisturbed.
As challenging as technical issues can be, they are at least considered approachable by most organizations. But environmental problems, particularly noise levels, seem to universally receive short shrift, and are often dismissedas an unfortunate but unavoidable part of office life and beyond anyone's ability to deal with.
Of course, the problem of office noise is far from intractable. Numerous approaches can be taken to relieve or at least ameliorate the problem, the most obvious of which involves the reintroduction of an antiquated and long neglected piece of spatial division technology - the door. The real reasons that environmental issues go unattended are somewhat different.
Software developers are knowledge workers. Our job is to produce intellectual property. You would think it self-evident that work of this nature requires sustained concentration, and that it is easier to concentrate when things are quiet.
Back in my school days, these facts seemed to be widely known and accepted. When you went to the library, the school librarian (who, in my school, was a particularly ferocious woman the students referred to as "Conan The Librarian") would do her best to see that the library was quiet. Why? Because people were trying to study, to think, to concentrate. When there was an exam to be done, the exam would be conducted in complete silence. Why? Because it's easier to concentrate on your exam when it's quiet. When the teacher gave the class time to work on an assignment, the class was expected to be silent. Why? Because it's easier to think about your assignment when it's quiet.
In university too, there was little dispute about the necessity for a quiet environment when doing intellectual work. The libraries and exam halls are silent, the lecture theaters and tutorial rooms are quiet so that the speaker may be heard and their message understood.
Prior to entering the workforce, I thought nothing of it. It all seemed to be just common sense. Imagine my surprise then to discover that the corporate world had decided that none of it was true. That, in fact, you don't need quiet in order to concentrate effectively - you can work just as well when immersed in an environment that is a noisy as a your local shopping center. Or so I infer is the reasoning, because the standards in both office accommodation and behavior seem to have been determined with such an assumption in mind.
Sitting at my desk at work, I am surrounded by distraction and diversion, which everyone just seems to accept will not impair my ability to work at all. But my own impression is very much to the contrary. I find myself constantly frustrated and annoyed by the ceaseless chatter around me and the incessant whir of printers and photocopiers. I have never known aworkplace to be any different.
How is it that the corporate and academic worlds seem to have completely different ideas about what characterizes an environment conducive to intellectual activity? Why is it that the academic community seems to have got it right, andthe corporate community ubiquitously has it wrong? Surely employers are not knowingly paying their staff to be only semi-productive, are they? Unless the corporate world is consistently behaving in a self-defeating and irrational way, I must simply be mistaken about the effect this office noise is having on me.
Perhaps I am actually quite unaffected by the conversations that my cubicle neighbors are having, on matters unrelated to my work ... all day. Perhaps the four foot high partition which separates me from them is actually enough to reduce their inane chatter and laughter to a distant whisper - I guess the sound dampening cloth on it must have some effect. Although the partition only covers two of the four sides of my "cubicle", perhaps adopting a "glass halffull" attitude would make the lack of privacy less disturbing. Perhaps the sound of the printers and copiers in the facilities area, just three feet away from my desk, really isn't that loud. Perhaps the guy in the next cubicle whoinsists on checking his voice mail through the speakerphone isn't the sociopath he appears to be, and I'm just not sufficiently tolerant of others. Perhaps it's not really all that visually distracting to have people walking throughthe corridor beside my cubicle every few minutes. Maybe some blinkers, like those given to cart horses, would lessen the effect of constant movement at the periphery of my vision. And perhaps the ten mobile phone calls that my surrounding cubies seem to get every day, each one heralded by a distinctive and piercing ring-tone sampled from some Top 10 dance hit, really isn't as wearing as what I think it is. And maybe having a pair programming partner leaning over your shoulder, barking in your ear and correcting your every typographic error isn't an obnoxious novelty that removes what little remaining chance there is of thoughtful consideration occurring in the modern workplace, but a mechanism for solving complex problems by having a chat over a nice cup of tea.
Or perhaps, just perhaps, the cubicle farm is a fundamentally unsuitable work environment for software developers. But how could that be, when the "open plan" office is the corporate norm? Could organizations really be so blind as to routinely give their staff an environment which is not conducive to the conduct of their work?
How could such a patently irrational trend develop and persist?
The modern cubicle had its genesis in 1968, when University of Colorado fine-arts professor Robert Propst came up with the "Action Office" - later commercialized by Herman Miller 1. At the time, offices usually contained rows of desks, without any separation between them. At least cubicles were an improvement. But once the facilities management people cottoned onto the idea of putting people in boxes, their focus became achieving maximum packingdensity and consideration of noise and interruption went out the window (if you could find one). That mentality persists today, largely because the costs associated with office accommodation and office space rental are concreteexpenditures that appear on a balance sheet somewhere. Somebody is accountable for those costs, and therefore seeks to minimize them. But the costs of lost productivity due to an unsuitable work environment aren't readily quantified, they just disappear "into the air", and so are easily forgotten or disregarded. There are also tax breaks in some localities, where legislation exists making it quicker to write off the depreciation of cubicles more quickly than traditional offices.2
The ostensible benefits of an open-plan office are its moderate cost, flexibility, facilitation of teamwork and efficient use of space. These are the attributes by which cubicle systems are marketed 7. Note that the ability to create an environment suitable for knowledge workers is not amongst those features.
Flexibility, although a possibility, is seldom realized in IT-centric environments where the need to re-route power and network cabling makes people reticent to re-arrange cubicles to any significant extent. Even individual variation and customization is discouraged in many workplaces, where such non-conformity is viewed as a threat to the establishment.
It is also commonly held that cubicles "promote communication" amongst staff. Unfortunately, one man's "communication" is another man's "distraction", the difference being whether the desire toparticipate is mutual. Alistair Cockburn, never one stuck for a metaphor, describes the wafting of conversation from one cube to the next as "convection currents of information" 3 and promotes the benefits that might arise from incidental communication. But when one is trying to concentrate, these currents of information become rip-tides of noise pollution that one cannot escape. The result is frustration and aggravation for the party on the listening end.
Unsurprisingly, companies that produce modular office furniture claim that cubicles are fabulous, and choose to selectively ignore their manifest disadvantages. In the advertising literature 7 for their "Resolve" furniture system, Herman Miller lauds the necessity of teamwork:
All the accepted research in this field says you have to have more visual and acoustic openness to get the benefits of a team-based organization.
... and downplays the need for individual work:
Although there will always be types of work that require intense concentration and protection from distraction, our research suggests that these needs can be effectively met outside assigned, enclosed workstations- through remote work locations or on-site, shared, "quiet rooms" for instance.
In other words, the workplace should be optimized for collaborative work, and those who want to concentrate can go elsewhere. Indeed, it seems to be a growing misconception amongst designers and managers that a high level of interactionand collaboration is a universal good, the more the better, and that the downsides don't matter.
For knowledge workers, who spend the vast majority of their time in isolated contemplation, this is decidedly bad news. Those who fit out offices seem to be either gullible to believe glib rhetoric such as the above, or more likely,choose to remain willfully ignorant of the fundamental requirements of their staff. Herman Miller would have you believe that the cubicle environment is good for your software development effort as well:
But the benefits of physical openness are gaining recognition even among the"gold-collar" engineers and programmers of Silicon Valley.
"The programming code we write has to work together seamlessly, so we should work together seamlessly as well", says a Netscape Communication programmer and open-plan advocate quoted recently in the New York Times.
Clearly, it is inane to suggest that software can be invested with desirableruntime behavior by adopting parallel behavior in the team that develops it.Does the code execute more quickly if we write it more quickly? Will it be moreuser friendly if the developers are more friendlytoward each other? No - it is just nonsensical wordplay. But the use of suchfaulty "proof by metaphor" techniques is illustrative of howdesperate the furniture industry is to ignore the workplace realities they areproducing, and the superficial level of thought that they employ in promotingtheir ostensible success.
Consider the following statement, again from Herman Miller:
Recent studies also indicate that people become habituated to background office noise after prolonged exposure. Over time, people get used to the sounds of a given environment, and noises that initially have a negative impact on performance eventually lose their disruptive effect.
Or perhaps, workers simply give up on the issue of office noise after their prolonged attempts to deal with it are continually met with stonewalling and denial. No references are given, so it is impossible to gauge the validity orrelevance of these studies. But it sounds so inconsistent with known research in this area that one cannot help but be suspicious.
Many studies have examined the effect of background speech on human performance. 8 One phenomena that consistently recurs is the "Irrelevant Speech Effect" (ISE). In ISE experiments, participants are given tasks to do while being subject to speech that is unrelated to the task at hand. Susceptibility to ISE varies between individuals, but in general ISE is found to be "detrimental to reading comprehension, short-term memory, proofreading and mathematical computations" 9 In general, work that requires focus and ongoing access to short-term memory will suffer in the presence of ISE and other distractions and interruptions.
Real estate has always been an indicator of status. Whether you're a feudal lord or a middle manager, the area in your command is usually proportional to your perceived status and importance. Those who suggest that the cubicle is an unavoidable part of the office landscape are often those whose status precludes them from ever having to occupy one, and who have a vested interest in the distribution of office space remaining exactly as it is - in their favor. Theunstated purpose of the cubicle is to serve as a container for the "have-nots", to more obviously distinguish them from the "haves." The preoccupation with offices (and the number of windows therein) and car parking spaces is often quite baffling to techies, who think first in terms of utility rather than perception. But for those more "image oriented", the true worth of corporate real estate has nothing to do with functionality and everything to do with positioning.
I would like to be able to say that companies are gradually realizing that knowledge workers such as software developers need support for both team interaction and distraction-free individual work, and are making changes to workplace accommodation accordingly. But I would be lying.
In truth, the workplace's suitability as a place to work is likely to sink below even its currently deplorable standard. The trend is towards ever smaller cubicles with fewer and lower dividing partitions 7. A 1990 study by Reder and Schwab found that the average duration of uninterrupted work for developers in a particular software development firm was 10 minutes. That's revealing, because it generally takes about 15 minutes to descend into that deep state of contemplative involvement in work called "flow". During the period in which one is transitioning to a state of flow, one is particularly sensitive to noise and interruption 5. If you're interrupted every 10 minutes or so, chances are you spend your day struggling to focus on what you're doing, being constantly prevented from thoughtful contemplation of the problem before you by visual and auditory distractions around you... and that's the typical working day of many software developers. As DeMarco and Lister comment "In most of the office space we encounter today, there is enough noise and interruption to make any serious thinking virtually impossible."5With the addition of some doors into the environment, developers could at least control their noise exposure.
Look around you now, and what do you see? Chances are there will be at least one and probably many of your colleagues wearing headphones. It's common practice for software developers to retreat into an isolated sonic world as the only way they have of overcoming the incessant distraction around them. Some companies pipe white noise into individual cubicles to try and mask the surrounding noise. I've found it helpful to run a few USB-powered fans from my computer - their quiet hum serves much the same purpose, as well as compensating for the often inadequate air conditioning.
Why don't developers revolt? Why is it so rare to hear them vocalize their complaints? Talk to them in private and they'll likely concede that their work environment is too noisy to enable them to work effectively. But they'reunlikely to make those concerns public, for fear of retribution or simply because they know that the noise level will be dismissed as being an inherently intractable problem.
So we will continue to grind our teeth and shake our heads in disbelief while listening to the dull roar of the combined efforts of the printers, fax machines, photocopiers, telephones, speakerphones, inconsiderate coworkers, slamming doors, hallway conversations immediately beside our desks and wonder how we can be expected to work effectively amidst such a furor. And as long as developers continue to tolerate unsatisfactory noise levels, and work longer hours to compensate for their negative effect on their productivity, organizations will continue to ignore their dissatisfaction.