Software isn’t everything
Why the construction industry has a hard time tackling digitalization, who is doing it correctly and what we can learn from them. Systematic thoughts on digital processes in construction. Inspired by the digitalBAU, open source oftware and one grand vision.
Be it fully immersive VR experiences or surreal hologram displays of 3D models, walking across the digitalBAU in Cologne doesn’t exactly leave one with the impression, that the construction industry is one of the least digitized industries in the world. At least not at first glance. But that first appearance is deceptive. For to properly “use” digitalization for the greater good is not to convert paper into PDF, drawing board to CAD and 2D to 3D. It’s more, as we shall see.
It is no surprise that mainly the traditional sectors and tasks of construction find their space on the fair. Whether it is the App that replaces a paper-based timesheet, platforms for online tender or the photo documentation tool using your phone. It might sound harsh, but to digitize these tasks is obvious, if not to say mandatory (by law).
One often gets the impression that digitalization is understood in a very narrow way in the building industry. That is to say, the process is being translated “word-by-word” from the analog to the digital world. Rarely is there any new process thinking or serious new business models. A good example here are the CAD software companies, which by the way use up the single biggest slot on the fair. It’s no longer a secret that the vast and without a doubt impressive features of these applications are being build around an aging core. Besides the obvious performance issues in bigger projects, this is a software-engineering nightmare. Combine that with non-compatible data types and pricing models, which enforce the lock-in effect and the whole thing doesn’t look progressive anymore.
The next big thing… BIM
A hint of true progress — behind some nebulous fairy tales — is the buzzword BIM. While one still has to hear past questions like “What BIM-Software is there? (BIM isn’t software), the realization slowly makes its way through the ranks that the whole process of construction ought to be redesigned. It makes its way rather slowly and hesitantly though. Thus, many of the BIM related talks in the forums of the digitalBAU try to relate theory to practice and show that this ominous BIM can not only incur cost but also be useful.
With BIM, the silo thinking in the field slowly relaxes. Platforms for BIM data and respective facilitators like BIMsystems, BIMswarm or SDaC bring fresh air into a stale process, not only visually with modern UIs, but also in the way they consider collaboration between the various roles. But more on that later.
It’s not about Technology
Summing it up, the quote from Thorsten Dirks — initially intended to be cross-industry — neatly suits construction: “If you digitize a shitty process, you’re left with a shitty digital process” (Translated from german)
I came across this quote just recently at a hackathon in Hamburg, where it was our task to tackle the highly dynamic traffic problem in the harbour with tech. Eventually, our project succeeded, albeit not targeting traffic at all; simply because the underlying problem wasn’t a lack of technology, but an old-fashioned system. The harbour is operated and used by a multitude of parties that don’t trust each other, don’t open their processes and don’t collaborate beyond what is legally required.
A very similar phenomenon afflicts the construction industry. But there are first approaches to circumvent it, without throwing everything aboard.
Before I continue to outline the following thoughts and analyses, I would like to clarify two things. First, we (ModuGen) are pro small and medium sized enterprises (SME/ mid-tier) in construction. We don’t want capitalistic consolidation into mega-corporations like Katerra. This has economic, but primarily socio-ethical reasons. Second, we a pro creative individuality in construction. We don’t want standardization at the cost of freedom. Building is an essential component of every civilization. The artistic beauty that pleases human perception must not be compromised by the technical urge for optimization.
The open source concept
A term that is mentioned pleasingly often in the presentations of the pioneers in the industry is: openBIM. The idea is to work with open standards for communication and thus enable multiple specialized programs for every phase and trade in the planning processes. Given the complexity of a big construction project, I believe the big open BIM approach is the only realistic one. Monolithic software giants (a.k.a god software) will never be able to cover the variety of functionality required, even if one of the major software providers still seems to have that goal.
What goes along with openBIM for me is necessarily open source. That means, collaboration without a claim on ownership. That means, sharing and collaborating on common problems. That means, in the direct sense of the word, open sources.
But how does open source get along with traditional, profit-seeking corporations? If new to the concept, one might question critically, how that’s going to work. How can we give insight into our company details, our software and our processes, without losing our competitive advantage? In the answer to this question lies the vision for a new, more productive construction industry.
Open Source modules for engineers
All this is mainly taken from the world of software development. It is there that the open source idea was invented and it is there that is fostered and lived like nowhere else. And still, or maybe because of that, software corporations are the intellectually most productive and economically most profitable constructs on earth.
If we look, for example, at the mobile operating system iOS from Apple, we will see that most of the source code is in fact open source. For Apple, this doesn’t result in any drawback, apparently. Rather, it is a major opportunity: less work. Because open source usually also means that the community contributes to the published modules. According to the maxime “you don’t have to reinvent the wheel”, even the biggest competitors Apple, Microsoft and Google share common code.
To give a simplistic example, every operating system should be able to read files from a hard drive. Instead of rebuilding that feature from scratch, however, the code is written once and then published. True, the competitor can now use it, without putting work into it. But he can also contribute and improve it.
An essential perspective besides that of the corporation is that of the individual. For me as software engineer, open source mans that I can use the same powerful tools as the big players. A great example being TensorFlow, Googles framework for high-performance deep learning, which is completely open source.
In the last for month I developed an open source tool for structured scientific research together with Florian Wilhelm at inovex. While only the two of us wrote the code, we use functionality on which more than 30 000 people contributed voluntarily. That means, it is possible for us to write software with a huge feature set by using components in which decades of engineering intelligence have been invested. In turn, we also publish our work.
But what about us?
Now all that is nice and well, you might think. But what does it have to do with construction? I think, no I know, that in construction there are also reusable and reused modules. These modules don’t need to be physical components but can be processes and important steps. Instead of sharing these modules, however, every engineering office diligently develops their own version behind closed doors. Every one of them reinvents their own wheel. With the benefit of being able to say: “Our solutions are unique, crafted in-house”. With the drawback that none of these wheels uses its full potential — or to speak metaphorically, runs smoothly. Looking at the whole macroeconomic system, the industry is wasting an enormous potential. Because what is higher productivity, if not producing more output with the same amount of work. If we share major parts of recurring development, we increase productivity. Without digitalization.
One task that is still ahead of the industry for the open source idea to work is modularization of processes. In software engineering “loose coupling” is one of the most important principles. Loosely coupled software architecture ensures that single components can be changed and developed independently of each other. Every part has to know of all the other parts only what is strictly required. By doing so, the user of a component (for example a data-reading-component) doesn’t have to know how it works. Many more principles like this are bundled in the so called CleanCode maniesto. A detailed discussion of that with an attempt to translate it into “CleanConstruction”, may follow another time.
Lets consider the example of bending-resistant wood connection, to make the whole idea more tangible. There are many ways to implement such a connection, that differ depending on the requirements of the setting. However, for every abstract use-case — that is given loads, moments and measures — one implementation can be found and published. Thus, not every engineer facing a similar problem has to reimplement the whole solution. Instead, he can use his abstract requirements out of his CAD- or structural model to find a set of possible parametric open source solutions, use it and improve it. Samuel Blumer is currently finishing up his PhD thesis at the ETH Zürich about such an abstract interface based on IFC (with upgrades).
In fact, there are already similar concepts in production, implemented by the Open Source Wood Initiative. The idea: timber engineering solutions open source. MetsaWood has introduced this platform, says Tuukka Kyläkallio, after they noticed that they can be of more use if they focus on the production of their products and leave all following engineering work to local offices. Thus, they short-handedly published all their work and now encourage everyone to contribute.
Open source processes
Not only specific, physical modules can be developed in that way. Common processes in the planning can be open sourced as well. To give an example from IT: to write code is one thing, a whole other business is to get the software production ready and “ship it”. However, according to the open source ideology, there are powerful tools for that, which almost completely automate the process. Which means that whenever I make changes to my code, a whole chain of processes is kicked-off. Some test my changes for style and functionality, some build the software executable, and some deploy that executable to a live server. Thus it is possible for me as a one-man-army to deploy software updates to my application every single day.
Now such a reference wouldn’t be worth anything, without a respective counterpart for construction. “For us at SDaC it is important, to develop a platform and an ecosystem without an entry barriers, so that everyone can benefit from it”, describes Maximilian Schütz the idea of the AI-research project SDaC at the Karlsruhe Institute of Technology. “We develop something like an App-Store for AI applications in construction”, Svenja Oprach puts the goal for which the german ministry for economics (BMWi) has provided almost nine million euros of funding. The basis for the platform and the applications running thereon is structured data from thousands of building projects.
The idea of BIMswarm, on the other hand, is to build a plattform for collaboration in BIM projects. Everyone should be able to view and use applications, services and catalogues, which were provided publicly. The leading concept is also openness, but the whole thing isn’t completely open source. Even if it would be the most effective way for such a plattform to gain traction.
What has to go along with both of these ideas is a modular approach to the processes in construction, because only if recurring tasks can be isolated we can begin to put in effort to solve these.
Product- instead of process-individuality
Indeed, it seems to be exactly the modularization against which construction is so resistant. The paradigm that currently prevails in most projects is one of process-individuality. Rarely any project is done following the same process. Sometimes the architect is planning independently of the engineer, sometimes they do it together, sometimes they both work for the project developer and sometimes for the construction company. From the first idea to the finished building there are as many paths as there as projects.
Why the construction industry has this peculiar structure probably remains speculation. For one, building is something very personal and individual by nature, not like a phone, for example. Hence there is a myriad of local providers and not only one or two global construction companies. Furthermore, the industry has grown over centuries with its roots in the guilds of the middle ages. And lastly, construction has enjoyed an economic boom over the last years. “Such a surplus in demand doesn’t exactly foster a systemic reinvention of a whole industry for the sake of productivity and sustainability”, says Philipp Hollberg from CAALA.
The focus on process standardization while also ensuring product individuality was shaped primarily in the automobile industry. Under the term mass customization, however, the concept is increasingly introduced in other sectors as well. What the car manufacturers achieve with their configurators is similar to the goal of the construction industry: His own individual product for every customer. At the same time, the benefits of mass production like scaling effects and automation should be used.
To be sure, the analogy to BMW, Daimler and others has to be taken with a grain of salt. Not last is the final product they strive for much more standardized by definition. Every car has four wheels, an engine and a steering wheel. The same amount of standardization cannot be found in construction. That means in practice the standard configurator approach will only find marginal use. “Only 10 to 15 percent of all construction projects can be covered with standardized buildings like modular construction”, clarifies Hans Jakob Wagner from the ICD Stuttgart.
More important than the specific solution is the underlying though. As I’ve mentioned before, there are recurring processes and elements in a building based on which a myriad of creative designs can be developed. Process-standardization doesn’t equal construction kits at the cost of artistic freedom. Much more, as Dominik Steuer of Steuer Tiefbau puts it, it is about getting from project to product thinking. That is to say, more standards in the process for more capacity in individuality in the product.
Combine the two for progressive constru-tech
The fact that the digitalBAU took place shows, that the sector is starting to change. Now the responsibility is with the digital facilitators and visionaries to take the right steps. That means, making digitalization tangible and perceptible, without barriers. That doesn’t mean to digitize tasks for the sake of digitizing them, but to implement smart solutions where the fruits are hanging low.
Construction has the potential to foster purpose-driven innvoation, because our goal is as clear as no other: to create living space worth living in.
What might sound a little romantic can function as a simple guideline. With this guideline in mind it is our task to rethink which solutions truly generate benefits, which truly avoid waste and which truly enhance the experience of all stake-holders. Visionary projects like the BUGA pavillion in Heilbronn show, that exactly that is possible. Digital, solution-oriented thinking and working through and through. Thus making building an experience for the customer, the workers and not last — thanks to the resource wood — for nature.
Recently, many new processes start to take off and it remains to be seen, who will win the race. My forecast is clear: innovative process thinking and a community based open source approach ought not be missing.
What do you think?