From Theory to Practice via Tools
How can we improve knowledge transfer in higher education?
Surviving the 21st Century: Growing a protopia
If builders built houses the way programmers built programs, the first woodpecker to come along would destroy civilization.
Why can builders build houses than survive decades of use?
The answer is both simple and profound: they have good tools.
But what are tools?
The history of technology is the history of the invention of tools and techniques and is one of the categories of world history.
— History of technology, Wikipedia
Tools are what enabled humanity to get out of prehistory and plant the seeds that grew into civilization and history.
Tools have been studied by scholars and laypeople for centuries, one of the primary aspects we have been focusing is how they increase productivity. Tools multiply our capacity and reduce our efforts, in a way that can be taught and learned by many.
As long as we see tools as primarily devices and gadgets we fail to grasp how deeply intertwined are tools in our culture and way of thinking.
Tools are ways we use to encapsulate knowledge and insights. The tool maker, as they gain expertise in their craft, is able to make better and more effective tools. One example of such tools are best practices: experienced practitioners are able to deeply understand what does work and doesn’t work, capturing this knowledge in form of principles and practices.
The knife wielder doesn’t need to understand how knifes are crafted in order to use them. All the knowledge about sharpness and precision and safety is embedded in the knife’s design: from the handle to the tip, the way it balances in the hand and transfers force to the edges, all of it is incorporated in the knife.
Writing is also a tool we use to better organize and share our thoughts, through space and time. Books and the printing press are improvements of writing in certain aspects (e.g. sharing thoughts of their writers), while neglecting the others (e.g. organizing thoughts of their readers).
One aspect of tools that we usually neglect to study and understand is how we can use tools to build other tools. Sometimes people talk about this when they think of using software to build better software, but this still is a dream rather than a reality.
There’s a way we have been using tools as a means of building tools for centuries, a way we usually don’t think about: academic disciplines.
Mathematicians work on several kinds of problems. As these problems are partially or completely solved, some new elegant solutions are crafted by the mathematicians. These new solutions are found to be useful when working on other mathematical problems.
The solutions they create (or discover) are useful for more than just the original problem they were intended for. Furthermore its common to work on a problem and, even if that problem ends up being unsolved, these new solutions used to tackle the problem are useful by themselves.
All the knowledge and insights tied to that solution make it useful as a tool that can be used by other disciplines.
Physics Nobel Laureate Eugene Wigner wrote about “The Unreasonable Effectiveness of Mathematics in the Natural Sciences,” on how mathematical tools are particularly effective for solving problems in the natural sciences.
A similar titled article could be written about “The Unreasonable Effectiveness of Physics in the Engineering Fields”, for example. There’s no such article because we take it for granted, it’s obvious as engineering is “mere” use of the solutions physicists discover (or create).
Nothing could be further from the truth.
Tools from physics enable engineers to use them to invent (or discover?) other kinds of tools, much like physicists use tools from mathematics.
Different kinds of complexity and precision can be mastered at each step in this journey of amplification of human capabilities and encapsulation of knowledge and insight. Mathematics enhance physics and physics enhance engineering.
As the saying goes, we are all “standing on the shoulders of giants” and, although some egos might be hurt, knowing that their own contribution are not as big as they might want or believe, as a species we should be proud of our capacity to build up our tools based on the existing knowledge.
Another, criminally underrated, step in this journey is the step taken by technicians, craftspeople, trade workers, and tinkerers.
Who would like for a mechanical engineer to fix their car? Surely they could understand the engine given sufficient time, but even then their theoretical knowledge of the mechanisms may fail to incorporate all the practical considerations and/or critical details, necessary for the whole machinery to properly work.
Who would like for a physicist to design a circuit? The same reasoning applies, the skills physicists excel in aren’t those appropriate for making efficient and cost-effective designs that can be mass manufactured with high quality and low costs.
At each step different kinds of problems are being solved and each step is both relevant on its own and critical for amplifying the skills and talents of those tackling the problems at their own step.
From theory to practice, but at each step further in the journey the earlier practice becomes now theory: a solid foundation on top of which further improvements and new practices can be developed.
Amplification and encapsulation.
That has been working for centuries. Builders build innumerable houses every year, none of them need to understand the details of engineering in order to use them with expertise to build constructions that most engineers would be unable to replicate as skillfully.
That works and we desperately need more of that.
What is the journey for programming?
Computer science curriculum is a mix of mathematics and trade. Software engineers are expected to both build databases from scratch and use them, without properly learning what properties they hold and how we can compose those components in larger systems.
There’s no clear separation of concerns, no building of tools that amplify and encapsulate. The disciplines are divided without rhyme, rhythm, and reason. They divides are historical accidents rather than pedagogical intent.
It’s not a surprise to see the mess we’re in.
What is the journey for medicine?
Doctors are expected to heal, but also try untested treatments in the middle of a pandemic. Biologists are far too removed to offer much help. Biomedical sciences are understood to be a intermediary step in the right direction, but still we are missing a few steps in between.
Like in programming, the answer may lie in reorganization and adding intermediate disciplines, maybe more than one. The structure that goes from mathematics to trade builders is a description rather than a prescription, of what works.
How can we test this idea and see if it works?
How can we improve an existing journey so that the tools built at a earlier step in are more effective to solver problems in the next step?
As an exercise, let’s design a rough plan for a journey’s improvement and estimate how much would it cost.
Higher education costs can be estimated to be between 30% and 60% of a country’s GDP per capita. This includes not only tuition fees but other costs. NCES published an article about “Education Expenditures by Country” for an overview of actual surveyed costs.
Developing a new curriculum is tricky and estimating its costs with accuracy is beyond the scope of this exercise. Instead we are going to consider it to be a percent increase in the overall costs of the experiment.
Let’s say we are redesigning a group of disciplines and we’ll end with four “new” disciplines.
We’ll need to account for the education costs of four years of graduation and four different groups of students, one for each discipline. We’ll assume fifty students per group, so we have a total 800 education years. That’ll be our baseline cost.
The experiment may have the usual subsidies offered by local governments, but we’re going to ignore it for now.
As the cost of developing the new curriculum, adapting material, and so on, we’ll assume an additional 25% above the baseline costs, bringing it to 1,000 education years.
In the US that experiment would cost $31.6 million or $7.9 million per year.
Many government agencies’ marketing and public relations budgets are higher than that.
A one billion dollar trust fund earning 3% above inflation per year could pay for four such experiments and still have the inflation-adjusted principal.
There are many pension funds bigger than that. For a fraction of “lost” future earnings an union could make an investment that would return a thousand-fold benefits to society in a couple of decades. Who wouldn’t want better medicine for all during our retirement years in exchange of 12.6% less interest?
It’s possible to run multiple such experiments in parallel, for different discipline groups and in different countries.
The first organizations to successfully complete this experiment will have a huge advantage: even if others can try to copy the methods, it’ll require enough new tacit knowledge that the copying organizations won’t be able to effectively reproduce it for a while.
Who wouldn’t want a monopoly on highly effective professionals?
