3.4 Serial Variation
Within the limitations noted above, the digital model was the closest possible replica of the 1980s plaster
interpretation of the colonnade by Cardonner. The process of constructing it provided insights into the
composition of the parts and their possible relations. It also allowed for easy comparison between three twodimensional
images. The first of these was the original photograph image, the second, an orthographically
corrected photograph of the façade as built to date from the surveyors in the Spatial Sciences School in UPC and
the third, the equivalent view of the digitised model to the same scale. There were steps at this point to establish
whether the original photograph was a true orthographic projection using photogrammetric techniques.(Maher
and Burry, 2002) In Adobe Illustrator this could be extended to comparative vector information. While it is
difficult to derive detailed information from the photograph, there were obvious differences that had been
introduced in the plaster model. The whole gable is much taller and there are differences in the numbers of
instances in the frieze and gable for instance. On the other hand the digitised plaster model provided information
about the plan that could not be derived from the photograph. With the ultimate objective of achieving a very
close resemblance to the rendered drawing in the photograph, we commenced on the construction of a parametric
model of the whole assembly in which the parts are related by a clear system of relations and constraints but the
overall configuration can vary through varying parameter values. In doing this we sought to find a systemic
approach to the serial variation in the frieze and gable. Measurements from the photograph and plaster model for
instance, hexagon widths and heights or step depths in the gable were plotted as curves against the “number” of
the hexagon or step and the sequential change in these measurements were also plotted. The resulting curves
were then analysed for fit to various types of curve. The hexagonal prisms of the frieze clearly grew in a linear
fashion and, given that the number of prisms and their approximate relationship to the intercolumniation could
be read from the photograph, the only further question was to estimate the extent of growth. For the stepping
gable there was a good correlation with quadratics or parabolic curves in the growth of the steps. In the plaster
model interpretation, the three dimensional curves described by the edge of each layer of steps was linear when
projected in the front elevation and parabolic projected in the plan.
ITcon Vol. 11 (2006), Burry and Burry, pg. 443
FIG. 6: a) front elevation view of the digitised plaster model overlaid on the original