Towards a parametric universe
“In the future everything will be designed to be reactive.”
What does that mean and why should you care?
Our world is changing. It’s happening at an ever faster pace. It’s getting hard to keep up with the changes, to stay on top of the latest trends, to enable the latest technologies and satisfy the latest demands.
Enter parametric design and reactive models. These two concepts (which aren’t new) will harness the velocity of change and feed off of it, empowering you and the rest of us to take advantage of progress without getting run over by it.
First some definitions
Parametric design is a paradigm in design where the relationship between elements are used to manipulate and inform the design of complex geometries and structures.
source: wikipedia
In computing, reactive programming is a programming paradigm oriented around data flows and the propagation of change.
source: wikipedia
Now some simple examples
When you instrument your ideas, your models, your prototypes in such a way that changing one parameter automatically adjusts all other parameters to maintain some fixed set of criteria you do not want changed you have parametric design. Reactivity happens when at least one of the parameters is changed from outside the model and you receive immediate feedback from the design.
For instance, imagine you have a rectangle and the only thing you care about is that it is 200px by 100px, everything else is optional. The height and width are your fixed criteria and everything else is a flexible parameter.
This should be a pretty familiar example to anyone who has ever used a 2D design tool with a ‘rounded-rectangle’ object. A more general example that everyone should be able to identify with is when formatting a text document in an application like Microsoft Word. When you adjust the margins of the page, the size of the text does not change but the number of lines it takes up does. This is an example of a parametric design.
Now imagine that you have 100 of these same rounded rectangles and you want to adjust all of them at the same time. Here we’ve changed the corner radius from 20px to 40px for all of the rectangles at one time.
This is the simplest of examples. It may not seem to be a big deal but if you are trying to fit your rectangles into a fixed amount of space then constraining them by their height and width is just the thing to ensure you meet your requirements. You still get to adjust that corner-radius, color-fill, border, opacity, etc. and could do so individually or in groups while still meeting your requirements. That is the point of parametric design. It enables creativity while still solving for a specific problem space.
A Brief History
Let’s step back now that we have a simple example of what we’re talking about and look at where we are in terms of adoption of this paradigm.
Architecture, Engineering and 3D Modeling
The first industries to adopt parametric design were architecture, engineering and later 3D modeling tools. In fact they didn’t so much adopt it as define something they were already doing and give it a label. From the beginning real world structural engineering was about solving for physical constraints like material strength, weight/load limits, available space, materials, budget and certainly utility. Parametricism is the recognition of these constraints as a tool rather than an obstacle. We owe a lot to the early pioneers in these disciplines, so much so that there isn’t room enough in this article to detail it all. I highly recommend that you google parametric architecture, constraint modeling, inverse kinematics and generative art for examples.
Enter the Internet and HTML/CSS/DOM
HTML, CSS and the web browser rendering of the Document Object Model are likely the most widely used example of parametric design. Every element and every attribute of every element contributes to the overall final rendered layout of any given page. This is an amazing feat in and of itself (kudos to the browser developers out there) and is one of the primary reasons why it is such an accessible communication medium — it is very forgiving. CSS or cascading style sheets layer additional complexity on top of the browser’s html rendering and with the DOM an entire functional model is embedded in the whole thing. It is surprisingly complex and yet superficially simple — which is at the heart of it’s widespread adoption and the global transformation that resulted.
Machine learning as parametric design
A next generation example is theGrid.io — a service being offered wherein a website is made fully parametric, the content, features as well as the (responsive) design. Though they refer to it as AI, it can also be described as an elaborate setup of constraint modeling, with learning systems layered on top to help optimize to the local maximum. Put more simply, it tries to find the best possible design using the content and features you tell it to include. Features like built in face detection to optimize image crops are being included with the first release. It will be remarkable, some will call it magical. We’ll see if they’ve built something that is scalable in a way that satisfies a mass market. I’ve already signed up but that doesn’t guarantee success, I’m no oracle — just a guy who thinks a lot about the future.
Put more simply, it tries to find the best possible design
using the content and features you tell it to include.
Service as parametric design
Elon Musk’s Tesla vehicle platform is also parametric, it’s a parametric service design. They just announced a service to provide directions to the nearest charging station based on charge level and GPS coordinates and a list of charging locations. Presumably it will continuously monitor all of these to ensure that the vehicle never leaves a safe zone without alerting the driver. Well that and the fact that it’s entire drive train is one big constraint model.
As we move completely away from mechanical devices towards fully electronic devices, it makes more and more sense to simply design out failure states and design in the ability to continuously monitor for optimum performance. To do this effectively we will have to parameterize the important variables and establish dependencies between them to make our models not only accurate but highly predictable while still flexible enough to be constantly optimized for the current situation.
But What About Reactivity?
Reactivity is something that hasn’t yet been fully realized. We are just now exploring what it means to be able to have a product or service adjust and react to our needs in real time.
The technology needed to make it a mass market option is just now becoming available at an affordable price. Highly personalized service has always been available for those willing and able to pay for the best people to cater to their needs. However, access to just in time information such as High Frequency Trading systems, Turn by Turn directions with road construction avoidance and all of the various recommendation engines for shopping are only possible with the big data solutions we have just recently invented and learned how to operate at a scale large enough to work.
Reactive products
Now and in the future, the most successful products will be connected and able to be updated, they will be reactive. I believe the reason for the iPhones success has nothing to do with it’s physical design, it’s visual appeal, hardware performance or Steve Jobs notorious reality distortion field. It has everything to do with the fact that the it was the first mobile device that was purpose built to be updated.
The iPhone was a reactive product.
Not only could it’s core services be updated to fix problems, add new features and completely change it’s entire operating system — it allowed apps to be installed that extended it’s functionality way beyond anything that had come before. Yes, there were smartphones that had a few apps and could be flashed with security updates — but the scale that Apple implemented was orders of magnitude ahead of what had come before. The iPhone was a reactive product.
Elon Musk’s Tesla vehicles are another great example of reactive products. Through a software update Tesla will soon be enabling vehicles to drive themselves (in select situations).
Reactive Services
Tesla isn’t just making reactive products, they’re also making reactive services. In an earlier software update to the Model S (at least), the vehicles will now auto-discover charging stations near the vehicle and the route being taken and notify the driver if the combination of battery charge, route and charging station locations are any cause for concern. The driver can then make informed choices about when and how to coordinate their drive to avoid unforeseen low battery problems.
We’re looking at a whole new way to go through life.
These kinds of services are going to continue to roll out and evolve. The more of them that are out there, the more we’ll expect all services to be reactive. We’re looking at a whole new way to go through life. Smart recommendations that take in all available information as well as any individual requirements or task based requirements. Intelligent agents that let us know how to invest and when, based on our priorities and resources. Grocers with drones who deliver our produce based on information they get from our ovens and fridges. The internet of things isn’t a means unto itself, it’s a way to enable reactivity across our lives. Business models that currently are not economically viable will become viable with the addition of reactivity within the context of parametric design. I could go on and on but I think I’ve covered enough ground for one read.
Stay tuned. It’s just getting started.
