The recent Washington National Airport story certainly must have surprised a lot of Washington Post readers, with the possible exception of those of us who are architects and engineers.

After several years and millions of dollars invested in meticulous functional programming and design, airport planners created a design with a fundamental flaw: air traffic controllers atop the existing tower would have their mandatory view of the end of one of the runways obstructed by a wing of the new terminal building.

How could such a critical and potentially costly error be made?

Well, to be honest, such mistakes are embarrassingly simple to make and occur more frequently than designers like to admit. Fortunately, most can be rectified -- or concealed -- more easily than the one at National Airport.

Even while believing that checking sight lines at an airport is an obvious and indispensable thing to do during the early stages of planning, we architects know deep down that we too might have missed this one.

In reality, it is almost impossible not to overlook, omit or misjudge some subtle aspect or unforeseen detail of a building during the design process. Equally unavoidable are errors caused by not having complete or correct data on which to base a design.

The history of technology in general, and architecture in particular, offer countless examples of errors in research or planning that sometimes were difficult to avoid or anticipate. The surprise is that we don't make more and bigger mistakes more frequently.

Consider how complicated and risky the design process is, even when designing nothing more complex than a house. The possibilities for missteps seem infinite.

Unlike other manufactured and assembled devices, for which one or more design prototypes are built, tested and debugged, most works of architecture are designed to be built only once. Thus, the first version of a building is the last version.

This means that designers must rely solely on graphic representations -- models and drawings, generated by hand or computer, at scales much smaller than reality -- to simulate reality.

Further, drawings and most models don't operate or function, nor do they occupy or fairly reproduce the actual site where the building they represent is to be constructed and used.

When designing, architects can deal only with a small number of design issues and variables at any one time. Therefore, they must continually shift attention back and forth between the whole and its parts, element by element, gradually recording and developing their ideas into layers of design documents that attempt to show how everything fits harmoniously together.

Even before beginning design, architects and their clients must have in hand another simulation of reality, the building program. More verbal than graphic, the building program attempts to describe all of the processes and activities, spaces and functional relationships, movements, people, equipment and special requirements -- for lighting, acoustics, climate control, communications, safety, environmental protection -- associated with a project.

Program documents for a project like an airport can fill several volumes many inches thick. And they may include not only a quantitative functional agenda, but also qualitative criteria and objectives -- cultural, historic, social, aesthetic.

Finally, rigorous programs generally spell out economic goals and constraints, with specific budgets for design, construction, equipment and furnishings. You only have to look around your house, focusing on what its complete program might look like -- if you tried to write one -- to imagine what a program for an airport must entail.

Then there's the site, probably the principal source of design and construction slip-ups.

Site investigations normally are undertaken at the outset of a project to identify every conceivable natural and man-made site attribute. Detailed surveys show boundary lines, easements and rights-of-way, underground and overhead utilities, rock outcroppings and bedrock, water above and below grade, flood plains, wetlands, significant vegetation, wildlife habitats, topographic contours and all existing structures.

If any one of these is not located precisely or is inadequately surveyed, its eventual discovery can compromise an entire design concept. Worse, anomalies found during construction can substantially delay project completion and increase construction costs dramatically.

Even subsoil testing, performed by boring test holes around a site and evaluating the composition and strength of core samples extracted from each hole, can fail to reveal critical constraints. Normally, borings are spaced widely apart because engineers and architects assume -- with reasonable probability -- that subsoil conditions will be fairly consistent all across the site. Unfortunately, this isn't always the case.

During excavation, contractors may be surprised by encounters with boulders, underground springs, pockets of organic fill or abandoned pipes never revealed by any surveys or test borings. When such unforeseen events occur, the owner typically bears the financial responsibility unless surveys were done improperly by technical consultants.

Designers also must take into account surrounding properties, worrying about casting unwanted shadows, obstructing necessary views, adversely affecting traffic safety or generating unacceptable noise or pollution.

Further, site research must include identifying and interpreting all applicable regulatory codes that impinge on the design and development of a project.

At National Airport, one such code provision required unobstructed views of all portions of all runways from the control tower. Evidently, in establishing the geometry of the airport master plan and the airport's new terminal structures, designers simply didn't realize soon enough that, in three dimensions, this provision would be violated.

It's hard enough to cope with building on empty sites in the middle of nowhere. But an airport site, to which is added the complexity of thousands of feet of airspace above, is a nightmarish web of forms and systems.

Dozens of disparate buildings, miles of roads and ramps, parking garages and parking lots, rail transit lines, underground networks of tunnels and utilities, mobile equipment, aircraft aprons and runways have to find their proper shape and place.

Functionally and technically, it would be easier to design a town or city than an airport.

The good news is that the design flaw at National Airport was noticed before construction began and apparently can be remedied by building a new control tower, a solution much less expensive and time- consuming than relocating terminals, roads and parking garages.

With luck, this will be the only glitch on the way to a new and improved airport. But, with sympathy for the design team, I doubt it.

Roger K. Lewis is a practicing architect and a professor of architecture at the University of Maryland.