Frequently Asked Questions about
Promise Theory
These questions are a work in progress, subject to continuous improvement.
The first part of the FAQ concerns Promise Theory as a tool in the area of business and technology. Additional FAQs, farther down the page, looks at other subject areas and offer targeted answers about the theory and formalism.
- Who needs to know Promise Theory?
- What's the relationship between Promise Theory and The Open Leadership Network's circle of 8 core patterns and practices?
- Where did Promise Theory come from?
- What kind of problems does Promise Theory help to solve?
- Why do I need Promise Theory? What use is it?
- How do I use Promise Theory?
- What is Promise drift mean?
- Surely the world works on commands, why does Promise Theory take away that certainty?
- What organisations have used Promise Theory successfully in business and tech?
- Why not guarantees? I don't just want a promise, I want a guarantee!
- What's the relationship between trust and promises?
- Can I take a course?
- Which book should I buy?
- Is there a video introduction?
- What do you mean by Thinking in Promises?
- Where can I hear a podcast about Promise Theory
- Are there any other competing theories?
- What's the relationship between Promise Theory and DevOps?
- What does Promise Theory predict? Is it "Not even wrong"?
General questions about the method and status of the theory:
- What is Promise Theory about?
- Is PT quantitative or only qualitative?
- Has PT got anything to do with Quantum Mechanics?
- Has PT got anything to do with Information Theory
- Has PT got anything to do with Game Theory?
- Has PT got anything to do with Category Theory?
- Can PT really help to formalize social sciences where other attempts have failed?
Who needs to know Promise Theory?
Technically minded leaders, who are the implementors of change will benefit the most from Promise Theory. Architecture, in a human-machine system, is not just about the technology, it's about how people and technology work together to offer a network of human-machine services, with mutual trust, to create repeatable processes that keep promises through tailored relationships.
Technical theory is not usually considered to be for C level execs, but this is a must-have playbook of patterns for solution architects and business process leaders.
What's the relationship between Promise Theory and The Open Leadership Network's circle of 8 core patterns and practices?
Promise Theory provides the underlying explanation for the eightfold core patterns of the Open Leadership Network, for successful organizational transformation. Where do permission and authority come from? When will directives fail, etc? The Promise Theory Basics course answers these questions to reveal potential problems before they occur.
Where did Promise Theory come from?
Promise Theory was discovered, by accident, by Mark Burgess, while working in technology. It emerged from the software CFEngine as a proving ground, and its generalization to Policy Based Network systems. The principles were discovered while designing system networks that would automatically converge to a "desired end state" or desirable planned outcome. A desired end state is like a picture of a cake in a cookbook, rather than just the step by step recipe. It guides a process into land, like a GPS finder. We can apply that idea to all processes.
In Software Engineering, Promise Theory was able to solve constraint architectures that Object Orientation techniques would not allow, indicating a shift back towards service oriented design (independent of technology). Imposed constraints cannot always be resolved, but promises always can.
What kind of problems does Promise Theory help to solve?
Unlike many planning models, Promise Theory is a tool for understanding cooperation from the bottom up, building on trust between different agents (whether people, machines, organisations, etc). It allows us to make clear statements of intended outcome, both in business and technology, and see where the responsibility for those outcomes actually lies. Promise Theory goes beyond Business Process Modelling and flowchart models to explain how trust and capability play a role, and also how faults can be embraced and neutralised.
The ability to identify consistencies and inconsistencies has led to Promise Theory being used to analyse and contribute to the public debate concerning the design of the fated Boeing 737 Max affair and the role of its sofware processes, and even being used to analyse political arguments in Brexit and Nuclear Weapons negotiations. Promise Theory has been used to analyze the rights of people and things in society. In a collective, when does a right for some become an imposition on others?
Promise Theory shows how we can get beyond quantitative business metrics to include interpretation, differentiation, and value assessment from brands. It shows how planning based on inherent capabilities can be balanced sustainably within an organization. Promise Theory also offers some very simple thumb rules for the resilience of systems, and where proactive measures need to be taken.
Why do I need Promise Theory? What use is it?
The ability to get things done (at scale) depends on enabling networks of cooperation from a start to an end, not just on team spirit, or on logistical chains. We need to plan for all contingencies. Promise Theory proposed the patterns behind DevOps and Continuous Delivery in IT, even before these branded movements were named. It's also compatible with analyses like Goldratt's The Goal too, which came before.
When you know Promise Theory, you'll be able to spot socio-technical patterns---from the highly technical to the socio-economic. Once you know the patterns, you will also know how they can go wrong, and how they can be strengthened. Promise Theory is a kind of secret weapon you can apply without even talking about it out loud. By following its methods, you might be surprised at how quickly it acts as a lens through which to view all kinds of interactions, uncovering strengths and weaknesses as it goes.
When are you undermining your authority? Who is responsible for successful and unsuccessful outcomes?
How do I use Promise Theory?
You begin by identifying all the agents involved in a collaboration, their promises, impositions, and assessments---and sketch the relationships, looking for common patterns. Each promise is a potential failure point. Dependency on promises from others (power, weather, supply chains, etc) may be vulnerabilities or stress points that can be addressed with backups and redundancy. Discover when assumptions might be exceeded, and when a collaboration is at risk of collapsing. The key to a successful and sustainable system is to maintain relationships between agents.
What is Promise drift mean?
Sometimes our intentions drift off course---promises get forgotten, changed, and even deprecated over time. This is okay, if everyone moves in step. But if some agents change, while others stay the same, the nature of a system is changed too. Promises might be updated slowly (like boiling a frog) so that no one notices until it's too late. If the agent counterparts making use of the promises is not aware of the changes, it might lead to sudden failures when they try to rely on the promise being kept.
Surely the world works on commands, why does Promise Theory take away that certainty?
Promise Theory doesn't deny the existence of commands, rather it explains how they work, how they are different from promises, and why they can be far less certain and even ineffective, unless supported by a sufficient underlay of promises that enable them to succeed. Commands are impositions in Promise Theory language---they cannot work without the right environment of promises to support them. Impositions tend to work only within a network of existing promises, so the focus on command or imposition is something of a red herring.
When leaders view systems in terms of command and attempted control, they often lose sight of the necessary cooperation between agents in a collaboration. Impositions, threats, requirements, obligations, etc, are not often effective strategies, but they sometimes appear as a valid way of looking at what happened afterwards, if one doesn't look at what really happened carefully.
What organisations have used Promise Theory successfully in business and tech?
It's public knowledge that companies like Facebook and LinkedIn built their data centres using the software CFEngine and its Promise Theoretic approach to management at scale. Thousands of companies around the world, from Amazon to the military, have used CFEngine and its descendants for managing the configurations and relationships between their computers. Cisco Systems built software networking products based on promise theoretic principles. Most companies don't want to share information about their technology however. Amazon's pizza team concept and the microservice delegation model can be explained using Promise Theory.
Why not guarantees? I don't just want a promise, I want a guarantee!
No one can claim outcomes with absolute certainty. A guarantee cannot be the absolute certainty that a particular outcome will be delivered or will come into being. A guarantee is, itself, only a promise to deliver some outcome in the event that another outcome fails, e.g. to replace a defective item, or to refund some money in the event of failure to deliver.
What's the relationship between trust and promises?
It's a bit circular, but central to understanding the value of thinking in promises! Trust is a human judgement, informed by experience of reliability, and how well agents keep their promises. It might be a qualitative judgement (such as an emotional response or a contractual specification) or quantitative (as a measure of reliability over repeated encounters). Promise Theory is the only framework that can include both.
We humans typically start with a guess about the trustworthiness of machines or people, perhaps based on related experience, on reputation, kinship, tribal affiliation, or on other genetic affinities like having shifty eyes). Aggregate rumours and public discussion may adjust our assessments. Trustworthiness is an assessment. It can also be promised about oneself or another agent. Promises may be trusted or not trusted. Rumours are a form of promises. Do we trust the rumours?
Even non-human relationships may use effective indicators of trust (quantitative and qualitative), e.g. in how often they choose to check out and monitor other parties. Do you trust your car to start in the morning? The story of trust is a large topic.
(Note, the concept of trust in computer security is not the common meaning of trust. It has a technical meaning.)
Can I take a course?
Online courses are available through the Open Leadership Network. You may wish to purchase 1 or 2 copies of the course notebook, which contains slides and blank pages -- one for taking rough notes alongside the videos and one for finalizing your notes into a neat reference-- here.
Which book should I buy?
The two main books are Thinking in Promises (O'Reilly media) and Promise Theory: Principles and Applications (Xt-axis press). Thinking In Promises is a more popular introduction using only text and pictures. The more advanced and thorough text Promise Theory: Principles and Applications (Xt-axis press) contains a more symbolic or mathematical introduction, with a lot more details, for advanced readers. There is also a deep example guide, known as the Treatise on Systems (Xt-axis press)---a monster book, which goes on the apply Promise Theory to human-machine systems in much more systematic detail.
Is there a video introduction?
You'll find some friendly video introductions to the main concepts of Promise Theory on the Promise Theory YouTube channel.
What do you mean by Thinking in Promises?
We tend to think of the world the way we interact with it---from brain to hands, by manipulating it, hands on, commanding and forcing. This doesn't work well unless we can overwhelm the thing we are trying to manipulate by force. The more general approach is to understand how to make use of what things offer (or promise) to do, without overwhelming force.
Where can I hear a podcast about Promise Theory
- The Jim Rutt Show Podcast: Mark Burgess on Promise Theory, AI & Spacetime
- Agile Uprising podcast -- Promise Theory
- Q&A with Mark Burgess about Smart Spacetime
- DevOps Special: Banks, Brains, and Factories: Thinking in Promises for a Future World
- Promise Theory: Understanding how to scale and create positive team cooperation
- Interview on Promise Theory and Software Wind Tunnel: Velocity SC 2014, October 2013
- DevOps Leadership profile and interview and video.
Are there any other competing theories?
There are plenty of heuristic approaches to planning, but few theories to choose from, and none exactly analogous to Promise Theory. Most theories are specialized; Promise Theory is quite generic. One of its promises is to be an umbrella for many different but related ideas. In Computer Science, the Actor model is a kind of implementation that looks a lot like the model of agents and promises, but it's essentially a programming model not a theory of cooperation. In physics, network theories and Quantum Mechanics fall within the scope of forms of Promise Theory, but QM is much more specialized than PT.
In management theory, one tends to build conceptual models that leave the details of interaction to the reader's imagination:
- Cake layer architecture (or markitecture) diagrams
- Boxes with arrows
- Hierarchy charts
What's the relationship between Promise Theory and DevOps?
Promise Theory is a very generic method that can be (and has been) used to describe the interactions in DevOps, Continuous Delivery, Business Process modelling, ITIL, Pipelines, Petri nets, and all kinds of frameworks for modelling business.
What does Promise Theory predict? Is it "Not even wrong"?
There are plenty of predictions from Promise Theory. The first is that systems only work when there is both offer and acceptance, so we need to attend to both---i.e. commands do not work without an existing "invitation" to give them. This is true in all systems, human or machine, because it's a basic property of communication. Promise Theory also implies that the agents in any system must maintain wakeful interior processes, not just be passive locations. They must be capable of sampling and reacting what becomes available to them. No agent can be passive if it's involved in keeping a workflow going. Promise Theory predicts that the chance of an imposition being honoured within its expected time is at best 50/50, but that may increase up to 100% for promises.
Promise Theory as a Theory
This part of the FAQ concerns Promise Theory as a theoretical formalism. Readers should look at the Treatise on Systems for many examples.What is Promise Theory about?
Promise Theory is a way of understanding cooperative networks of all kinds. It models networks of `agents' (which represent any active part of a process, human, machine, or otherwise) and the promises they make to one another. It consists of,
- Principles, including the assumption of a priori independence of agents.
- Necessary but not sufficient rules for constructing interactions between agents.
- Set theoretic representations for semantics and dynamics.
- Implicit rules for reasoning about networks.
- An independent theory of measurement or assessment.
Is PT quantitative or only qualitative?
It can be both. It's not always possible to make a meaningful quantitative representation, in terms of numbers, but we can always use sets. Assessments
Has PT got anything to do with Quantum Mechanics?
The structure of Promise Theory has some interesting similarities with Quantum Mechanics, and its intepretations, but it is not based on observational phenomena like Quantum Mechanics is. It might nevertheless have something to say about it. There are also connections and similarities to Quantum Circuits, Tensor Networks, entanglement, and other topics. You can learn more about that in the book Smart Spacetime.
Has PT got anything to do with Information Theory
Yes. Shannon information channels may be formed from two promises (+b) and (-b) between two agents. The mutual information (overlap of interior states) can be viewed as an assessment of one by the other. In general, Promise Theory both summarizes information theoretic assumptions and results in a graphical language (with assumptions built in to the algebra or rules), and it is more "elementary" as Promise Theory assumes less about the endpoints of an information channel than Shannon's Theory of Communication does. An information model can always be represented as Promises, but not vice versa. See this paper.
Has PT got anything to do with Game Theory?
Yes. A strategic form game may be formed from a collection of bi-directional (+) and (-) promises that are mutually dependent. An extensive form game may be elaborated as a graph of conditional promises. Promise Theory may be considered more elementary than Game Theory, since games can always be expressed in Promise language, but not vice versa.
Has PT got anything to do with Category Theory?
Promise Theory (PT) and Category Theory (CT) occupy different layers of process modelling. In principle, CT is about mappings and relationships, whereas Promise Theory is a description pr causal intent. CT has no `locations' or `agents' that cause things to happen, so all arrows in CT look like impositions from a PT perspective. However, PT does little to formalize the promise bodies it refers to, as its focus is often on causal relationships, or interactions rather than the specific semantics of the interactions. This is where CT goes into ultimate detail. So, one way to think about it is that PT defines the spacetime arena for processes, while CT is a layer of detail, which explicates and details the nature of the promises, on top of PT layer.
Can PT really help to formalize social sciences where other attempts have failed?
Social science is often criticized for being anecdotal and `soft', meaning that it has few ways to rely on statistical objectivity. This is not unique to social science. Biology also has difficulties with statistical studies, because many important phenomena are not obviously repeatable. This doesn't mean they can't be characterized. Anecdotes are data, with semantic context which can't be handled by regular statistical methods. Physicists also naively look for dynamical models such as gas models, thermodynamics models, Ising modes, cellular automata, etc, as a possible analogy. But the real distinction at the level of social science is the role of semantics, and the lack of a theoretical model for them. One is left trying to find patterns in data, without a map. This is where Promise Theory strikes a balance, e.g in being able to define concepts like Authority, where previously this relied on philosophical rhetoric. In short, the lack of success in formalizing social science is due to a rejection of methods other than statistics. PT allows us to formalize theoretical models for later testing by blind statistical methods.
UPDATED: 18th Feb 2022