Parametric modelling can produce complicated forms and streamline workflow with ease. How could your practice use this technology?
Parametric software has divided architectural practices into two factions: the haves and the havenots. Rather than being a reflection of size and turnover, the separation is symptomatic of differences in mindset.
One side is intrigued by the complicated forms that can be generated using parametrics, the automation of workflow and the ability to automatically update the entire model when a change is made. The other is turned off by the unfamiliar architecture that has become negatively associated with parametrics, the difficult interfaces, the shift in thought processes and the time investment.
Parametric software was first developed in the automotive and aerospace industries. The first architecture-orientated package to be marketed was Digital Project (DP) in 2002, developed by Gehry Technologies, the software company founded by Frank Gehry in 1992. DP is fundamentally related to Dassault Systemes’ CATIA - the program used by Boeing to design its 777 and 787 series aircraft and famously by Frank Gehry for the Guggenheim Museum, Bilbao. Autodesk and Bentley soon caught on, successfully marketing their own parametric softwares, Revit Architecture (2002) and Generative Components (GC (2004)) respectively. The suppliers and producers of these three
softwares all promise that parametrics will transform architectural practice and revolutionise design. It is tempting to believe them when watching an experienced user at work on a mature model, seemingly generating endless variations which report back changes. But what is the reality?
In order to reap the ‘down-stream’ benefits, users must exercise foresight when setting the initial parameters for a design. Regular Revit and DP user Tim Lucas of engineer Price & Myers says: ‘You have to make the right things adjustable from the beginning, as you can’t make changes beyond the scope you set out.’ It is easier to make small rather than big design changes. According to Stephen Griffin, architect at
Allies and Morrison (DP user): ‘The person modelling should be close to the decision-maker to best predict the parameters that might change.’
Thinking ahead about design possibilities and being disciplined about inputting data requires an ‘engineering’ approach to design says Andrew Watts, managing director of architectural practice Newtecnic, who is familiar with all three softwares. He says: ‘Some architects think it kills creativity but this is the way the profession is going.’
Parametric software is useful in the automotive and aerospace industries because outputs are repeated and time investment can be written off against quantity. But architecture does not allow for the same efficiency, as most projects are one-offs.
Data inputting is a huge time
investment during the early stages and although many design variations can be generated, this does require manually changing values and waiting for the model and associated data sets to regenerate. Depending on the complexity of the project, this can take days to set up. ‘It is easy to get carried away,’ says Griffin. ‘You can script an algorithm that will automatically generate iterations but the mentality of “model once, edit to infinity” can mean you are spoilt for choice. You have to assess usefulness against time cost.’
Consequently, Watts says Newtecnic often uses DP to model and coordinate only after initial design decisions have been made. The practice also sometimes redraws the model in traditional CAD packages (e.g. AutoCad) to check for unnecessary complexity and to present to contractors. ‘Once everyone accepts 3D models as the only description of a building these programs will come into their own,’ he says.
Parametric software enables the design of complex geometry, but this is not an inevitable outcome. Griffin bemoans those who blame the software for ‘the bad stuff. It is just another way of facilitating design intent. It does not have to impact on aesthetics.’
‘We can use up to 16 programs during design and presentation,’ says Watts. ‘All architects operate in mixed environments.’ Transfer between applications thus becomes a sticking point as the intelligent data built-up is lost between software. Asci Aybars, Design Director at SOM London, which has licences for all three softwares, explains this industry-wide problem: ‘Interoperability is critical when a model needs to be exported for analysis, rendering or coordination purposes. It is generally understood that a file protocol like IFC will enable data transfer between platforms but this has a way to go.’ So a model that has been exported into, for example, a rendering package, will not update if altered in the source parametric program and will have to be re-exported.
Lars Hesselgren of KPF, who is heavily involved in the development of GC, maintains that if architects are designing it, they have a duty to understand how their complex geometry works ‘because this is their chosen mode of expression. It shouldn’t be handed over to engineers.’
By creating precise and resolved models, architects are reducing risk of overspend but this in turn has implications. ‘The accuracy of the software is incompatible with the traditional working methods of other building professionals such as quantity surveyors. If I know exact quantities, I expect an exact estimate, but they are unwilling to expose themselves in this way,’ says Griffin.
The concept of a model that can be read and edited by all stakeholders provides huge open-source benefits but it also replicates the disadvantages, making it difficult for architects to monitor design changes after they have handed the model over to the contractor. Should the model then be presented to the owner the architect may risk being sued if the built version does not tally exactly with the model.
Currently relatively few practicesin the UK are using parametricsoftware to design on a day-todaybasis, and much of the outputis coming from academia.‘Never have students had suchpower in the architecturebusiness,’ says Griffin. ‘Theproblem is that architects at the
top don’t understand what thestudents can do. They see thebenefit of endless iterations, butthe students don’t want to use thesoftware in this way - they want to explore.’
Most practices see parametric software as a way of reducing team sizes by automating workflow. ‘This can work for small practices, but here at Foster + Partners we like to have large teams - to get results you need diverse skill sets,’ says architect Hugh Whitehead, a key figure behind the development of GC. ‘You need a computer-savvy team with a dual interest in programming and architecture,’ continues Griffin.
‘If you asked me if I thought parametric software was an inevitable future for architecture, I’d say yes,’ he adds. ‘But the real breakthrough will happen when a young firm takes on the challenge of redesigning the design methodology to match the new tools.