Based on my observations of hundreds of architectural staff working on literally thousands of projects over the decades, I think it's safe to say that of all the components of a building envelop, the roof is perhaps one of the "less popular" design elements to develop. The plans and elevations normally take the top spot in the interest category among the team, at least initially. After all, if the project has a "flat roof," especially with a parapet, it's not normally visible from adjacent grade. If it is a simple sloped roof, it is something that often fades away from street view at an angle, and not likely to draw much attention.
These factors are among several that contribute to the delay of developing the roof design until a bit later in the project, which can cause problems down the line. As an example, I recall one municipal building project that had a simple shed roof over the entry, forming a sort of canopy. Because this portion of the roof was visible from the interior of the building, as the entry lobby was open to the second-floor level and had windows overlooking the front of the building, the designer wanted to use a standing seam metal roof for a clean appearance.
The details were developed late in the project, and the project went out to bid. In the construction phase, the roofing subcontractor noted that the roof was "too flat" to comply with the warranty of the standing metal roof that was detailed and specified. With all the framing in place, the alternative was to use a standing seam membrane system that would look similar in appearance. The kicker was that it cost the client an additional $10,000! Since public projects have very tight-if not fixed-construction budgets, this situation resulted in much grief all around.
In the situation above, had the designer done their homework early enough in schematic design, they could have used the originally specified standing seam metal roof if they adjusted the roof pitch just slightly. Having been in practice for well over 40 years, I have seen many roof issues that could have been resolved with the proper attention given to it early in design.
In another example, I know of a flat-roof distribution center project that used perimeter roof scuppers, leader boxes, and downspouts to provide the proper roof drainage. This is actually a very common approach with industrial buildings, especially in wet climates. However, in this case, the roof design was developed a bit late in the project, and the downspout locations became a difficult challenge. Among the elements to coordinate for the positioning roof scuppers were rooftop equipment and overhead door locations along the building elevation.
This became so difficult that the team decided to reduce the number of scuppers and downspouts required by instead opting to provide the needed capacity by upsizing fewer drainage components. That approach worked out well until the required bend in the upsized downspouts (now 10-inches in diameter instead of the standard 6- or 8-inch diameter) to coordinate with structural footings and connect with the lateral storm line created a condition where portions up to 6-feet in pipe length were exposed above-finished grade! It turns out that the "standard detail" for downspout bends at footing locations was developed for pipe sizes of 8-inches or fewer.
This all could have been avoided had the roof drainage been thought about earlier in the project, and mechanical equipment, overhead doors, and footings could have all been coordinated to work with the number of 8-inch downspouts needed to provide the proper drainage.
These are just two reminders of how important it is to consider the roof design early in the project. While a roof may not be the most exciting of architectural features, it is one that requires attention early on.