On planes I take the aisle seat, in buildings the window. It's for the same reason in both cases - personal control.
From a users' perspective, perception of personal control is the single most important factor underlying comfort. This perception is also linked more indirectly with many other things like productivity, stress, ill- health, safety and security. This is so for people using planes, cars, computer software, most kinds of tools and technical devices and, especially, buildings1,2,3.
The obvious reason why people want more control is to improve things for themselves. It is less obvious that users often use coping strategies when they try to make things less uncomfortable or less dysfunctional. This is where designers and users part company because their strategies differ. Designers tend to optimise with respect to known constraints, notably to make things comfortable and functional. Users typically only try to avoid worsening consequences. They hardly ever optimise.
Users spring into action in response to random, external events like the sudden noise of a pneumatic drill in the street or a sudden change from sunlight to gloom, or in response to predictable events like twilight. They are likely to make the decision to use a switch or control only after the event has prompted them to do so rather than in advance of it. They will often wait quite a long time before taking action, only when they have reached a 'crisis of discomfort'.
Users also tend to over-compensate, completely closing all the windows in a railway carriage when it rains to keep out a few spots of rain, for example, so creating a hotter and more humid environment for everyone inside.
People will operate the element which is most conveniently to hand, not the one which is the most appropriate. They also often take the easiest and quickest option rather than the best for their immediate benefit.
The worst case for users is when they want to make changes but are denied them because there are no usable controls, or controls are unintelligible, or operating them makes conditions worse.
Having made the decision to switch, people will often as not leave the system in the switched state rather than altering it back again later, at least until another crisis of discomfort is reached. This tendency to take the 'on' decision but ignore the 'off' means that many buildings adopt inertial states in which their systems are left enabled or running unnecessarily. It is one of the reasons why many automatic control systems - especially for lighting - seem to work so badly in practice4. Not only is this inefficient for energy consumption but can also quickly degrade comfort conditions.
Inefficient and uncomfortable inertial states are much more common in open-plan environments. There, it is frequently difficult or tedious for people to arrive at 'off' decisions because of the complexity of the shared decision-making involved (abetted by poor design and layout of controls).
In specifying controls, the closer a control device is located to the general occupant the easier it should be made for them to understand, the more straightforward should be the technology involved and the more robust its construction and controls. In use, control devices must give direct and unambiguous feedback. They should almost plead with the occupant to be used. For example vertical sliding sashes, especially upper ones, must be easily reachable without standing on chairs and have no tendency to trap fingers, damage nails or leave dirt on the hands.
Setting a light
Users expect windows to be controllable, give views out, allow available daylight in for most of the time while avoiding glare, give fast - if not instantaneous - response to personal comfort requests for fresh air or cooling, look good and give weather-proofing, thermal and acoustic performance to a high standard. It is hard to think of other technologies of any type which have to undertake so many simultaneous functions. Many potentially clash with each other as the prevailing context changes, like the need to keep daytime traffic noise out but allow fresh air in.
The normal inertial state for windows in non-domestic buildings is closed. Windows closed at night by security staff or cleaners will usually remain closed, sometimes opened by occupants during the day. In many open-plan office spaces the opening of even a single window may cause conflict - either because of genuine differences in comfort needs between people or because of disagreements whose origins may be outside environmental controls. (These disagreements can spill over and affect use of controls, often irrationally so, and sometimes bewilderingly so to the outside observer.) Or windows may stay closed just because their operation is difficult - there may be a desk in front of the window - or the result of opening is too crude - for example the window opens too far or in the wrong direction, causing draughts.
The ideal window for users is openable on demand during the day as normal, but with a separate upper part which may also be automatically or manually opened during the evening or night to satisfy cooling demands. This upper element should adjust or shut again when the cooling requirement has been met so that the building is not left too cool (as can happen when windows are opened overnight in hot weather and conditions change for the worse during the night). The 'night-time' element could be an upper hopper and the day-time element a vertical sash. Or, perhaps better, a motorised upper sash for automatic night-time use with a local motorised over-ride control for day-time adjustment. And perhaps also an alternative manual option for day-time as well.
A window which approaches this specification has been operating at Marston Book Services, Milton Park, Abingdon. (Its performance is reported in reference 5 and its design features in reference 6).
Users like sash windows because:
they allow a wide range of adjustment, from a tiny crack to half open
people understand what they are for and how to use them
users can monitor their state and performance easily - you can see that they are open from a distance
they give instant, perceivable responses and feedback
they are relatively easy (but not the best) to clean and maintain
they can be combined straightforwardly with many other internal and external controls - blinds and other devices for solar and glare control
they fit the vertical plane of the building
they have many glazing options.
Drawbacks are often greater for the designer and specifier than the users, hence their increasing rarity. These include:
perceived higher maintenance costs, especially for wooden sashes
inefficient mechanisms, making windows difficult to open particularly aluminium systems with slender elements and most modern systems without intermediate glazing bars to provide useful hand-holds
difficulty in seeing when they are properly shut
Implications for design strategy
Providing people with more control is a major strategic issue for design because perceptions of control are linked to health and productivity3,7,8. The best way of doing this is usually through controllable windows which meet the kinds of criteria suggested earlier. The box sets out user-biased considerations for design of windows and window systems in naturally ventilated and/or mixed-mode buildings.
There are alternatives to increased personal control. In close-controlled, air-conditioned spaces with unopenable windows, an efficient and proactive help desk may be able to compensate for lack of personal control, provided that overall background conditions are stable, predictable and normally fall within reasonable comfort envelopes. Problems arise when occupants need to change things for the better for themselves. Sadly, this often occurs with help desk systems outside normal hours because they are frequently manned during core time only.
The worst circumstances for occupants are buildings which:
are unmanageably complex so that many of the functions provided may not be working properly or at all
have a high level of management and technology-dependency so that there is nothing the occupants can do about it for themselves, but the high levels of management required are not present. In these situations, people will be even more sensitive to control issues, which takes us back to where they prefer to sit in control-poor environments. Window or aisle?
Adrian Leaman is from Building Use Studies, which specialises in monitoring, briefing and user needs in buildings.