Before software, a façade was set by what lay within reach of the site. Joshi starts with the Taj Mahal, where the governing decisions were scale, the modulation of the surfaces and the four minarets that sharpen the identity of the mass they surround, each resolved through craftsmanship, local material and a reading of the climate. His second example carries no façade at all: at the Adalaj stepwell the building sits below ground, its surfaces face inward, and the architecture works through the experience of descending into it. That inheritance belongs to a hot country, and Joshi argues it produced an instinct different from the one façade engineering was built on.

India has a tendency of building with shade, because we have a lot of sun at our disposal. The west builds with the sun. They want the sun to come in.

Manoj Joshi, Hiren Patel & Associates

What changed was the medium rather than the ambition. As buildings went vertical, drawing gave way to two-dimensional CAD, then to modelling, and now to computation that tests a surface before anyone commits to it. Joshi is precise about the cost: the principles of proportion do not change, only their application, and he describes a profession fragmenting, with services going to one specialist and the envelope to another. Walinjkar reads the same period from the delivery side, where a shortage of skilled hands makes the tools less a convenience than a replacement. Their real contribution, in his account, is that decisions can now be made predictively instead of from experience alone.

The end of the template

The clearest measure of that is how a curved piece of glass used to be ordered. In the stick-glazing work of the early 2000s, geometry was resolved by measuring what had actually been built, and the inaccuracy of that transfer turned the process into a series of near misses. For semi-unitised work the answer was physical: a wooden template cut against the frame on site, sent to the glass processor and returned. Walinjkar delivered a job in Jaipur that way, and every panel needed its own template, its own spider holes and its own round trip.

The whole glass for that project was worked out from the 3D model. Not a single glass is a template, and each one is different.

Upendra Walinjkar, Aluvision Façade

He was describing the ATC tower in Mumbai, whose glass was cut entirely from the model and which has stood for twelve years. A 56-metre sky bridge at ITC Colombo pressed the case further, because each unit curved in both plan and section within a tolerance of two to three millimetres. The geometry was standardised row by row rather than treated as unique, so a breakage years later can be matched from the model instead of measured again. With aluminium prices climbing, optimising material, cost and heat gain is the part of the work that pays for the tools.

Change orders as the exception

Oza puts the benefit in the language owners understand: coordinated models move the discovery of conflict into the period when a drawing can still be altered cheaply. The workflow is unremarkable, Revit driven by Dynamo and Rhino by Grasshopper, but on a children's hospital in Australia the client wanted metal panels coloured at random, so the randomness was exposed as a slider the client could move and watch resolve until they chose a version to freeze.

Change orders used to be a norm. Right now change orders are becoming exceptions.

Chirayu Oza, Elysian Tech Studio

Hosting a project on a construction cloud makes each consultant accountable for their own model and lets an issue be raised against a named person with a date attached. The gap that matters most, though, opens later, at fabrication. Models are genuinely clash-checked on large projects and the tolerances agreed there set the building lines, but a consultant issues intent details, every contractor works to a proprietary system, and the fabricator appointed to build it frequently cannot bring the façade back into the model to verify it.

Where the model stops

Problems the model existed to prevent surface during execution. Walinjkar contrasts this with Dubai, where a fabricator's submission has to arrive with model support and architects check the relationships between the envelope and everything behind it. Capability rather than software is the constraint, and only a handful of Indian fabricators have it. Patel turns the argument back on the designers: the tools have made execution easier and visualisation honest, but ease is not a justification in itself, and the cultural reading of a building is not something a plugin supplies.

We also have to answer why we should use the software. Are we doing justice by using it or not? That is the individual designer's responsibility.

Bhaumik Patel, Bhaumik Patel Architects

He makes the point through Ahmedabad's own inheritance: Kahn at the Indian Institute of Management, working brick until it became the arch the material seemed to ask for, and Le Corbusier at the Mill Owners' Association, where the weight of the concrete is broken by sun-breakers, vertical on the river face and angled to the south. Neither decision came from a tool. Joshi presses further, arguing that software is a set of rules inside fixed parameters, and that impromptu judgement and tolerated imperfection are where a building's character tends to sit.

The question the software does not answer

He closes on maintenance, treating it as a design question rather than an operational one. Old buildings age gracefully and corten is admired for rusting, yet a contemporary façade in a hot, dusty city needs cleaning almost continuously. Joshi asks why the research money goes into glass rather than into surfaces that improve with age or shift with the season, citing Nouvel's light-responsive screen at the Institut du Monde Arabe as evidence that an envelope can generate an experience rather than enclose one. Walk a street in New York and the buildings resemble one another, and Ahmedabad is moving the same way. Oza answers that context modelling and shadow studies are how a kinetic idea gets tested, but Joshi's reservation stands: the industry is better at solving stated problems than at finding new ways to build.