Shape Computation Lab

Mughal Gardens Redreamed

🞩

01. Automated production of a Mughal garden.

02. Flow charts of a) border formation function; b) subdivision function; border formation function.

03. A catalogue of six parametric Mughal gardens in the original grammar.

04. A catalogue of three parametric Mughal gardens in the original grammar. Second row: Different motifs for water reservoirs

05. Different motifs for canal ends (1-3) Different nesting of initial rules (4-6).

Nirvik Saha, Dennis Shelden and Athanassios Economou

2017

 

Keywords: Mughal garden; Shape grammars; Grammar Interpreter; Formal languages

The original Mughal garden grammar (Stiny and Mitchell, 1978) still remains one of the most enduring examples of the visual precision and generative power of shape grammars and since then it has been routinely referenced as an exemplary work in the shape grammar discourse and its applications in landscape design. The original Mughal garden grammar ended up with the image of an elder Shan Jahan, builder of the Taj Mahal, deposed and imprisoned in the Red Fort at Agra by his son Aurangzib, looking out from his pavilion in the fort, and dreaming of the char-baghs that might have been. The conviction of the authors was clearly that the grammar specified in their work enabled such dreams of char-baghs to be dreamed again by all the readers who were presented with the rules of the grammar; A poetic excursion indeed but enabling too because there is no better way to dream figuratively or literally but to play constructively with things and ideas of all sorts. This work picks up exactly where the original grammar ended and provides a computational application to implement and enable the play with the rules spelled out in the original shape grammar and the trial of possibilities that cannot be seen unless they are worked out – including the ones that are already dreamt and built by their original authors as well as others, potentially new, allowed within the rules of the game. The implementation of the rules does not follow one-to-one the rules as explicitly discussed in the grammar but instead it consolidates them in a parametric schema that consists of variables and combinations of sets of variables and is instantiated when values are assigned to these variables. This bundling of the rules in a parametric modeler is attempted here to take advantage of existing computational packages in Rhino /Python and shift the interest this in the automated production of designs and their possible extensions in other domains.