Category: Christoper Alexander

Re: A Pattern is no Best Practice Yet!

Great article on patterns by Kris Meukens here

My slightly self-indulgent reply is below. I’ve always been fascinated by how our understanding of IT and organisational design in general seems to follow the same path of Christopher Alexander’s works on the design and architecture of the built environment. 


I think it’s interesting to see the parallel and delayed timeline between “patterns” as they evolve in built architecture theory, versus patterns in IT. 

I’m not an expert in either but I see the history of patterns in built architecture through the lens of Christopher Alexander:

  • Notes on the Synthesis of Form (1964, year 1). Starts to describe what later came to be known as “patterns”. 
  • Notes on the Synthesis of Form – Preface to the first paperback edition (1973, year 9). Already starting to rebel against those who focused on “design methods” as a meta study and assets “I reject the whole idea of design methods as a subject of study, as I think it is absurd to seperate the study of design from the practice of design”. 
  • A Pattern Language (1977, year 13). These are fully formed patterns with the notion that they can be combined to create designs. It’s not a simple mix and match – still leaves room for a design process. 
  • The Nature of Order (2003, year 39). Doesn’t exactly reject patterns but focuses on wholeness, a set of qualities, and a set of structure preserving transformation that help designs unfold. 
  • The Battle for the Life and Beauty of Earth (2012 – however focuses on 1985, year 21). This focuses on two world views that he saw as in battle – one of which was opposed to his style of building. 

Reading “Battle” in particular makes you feel our current understanding of patterns and our obsessions with Agile, Design Thinking, etc mean we are in the equivalent of year 25. 

Reading the above article feels like we’re heading towards the equivalent of year 30. I mean this as a compliment. 

Architecture Versus Management?

I’m increasingly seeing the management and architecture disciplines as being in a race to control organisations.  Both groups show behaviours that suggest they are trying to extend their own discipline to encompass more of the organisation.  Equally, both groups work hard to exclude areas they are not comfortable with from their responsibilities.  Architects want to control the organisation by controlling knowledge of the structure and value streams at all levels, while avoiding execution issues.  The management profession wants to control the organisation by controlling resources, while avoiding responsibility for technical issues.

Both the architecture and the management professions reveal their desires to control the organisation by the manner in which they grow the scope of their approach through ever increasing extensions of their disciplines.  The management discipline has grown from supervision, to general management, to strategic management, to change management, and to all of the business unit focused sub-disciplines that form the structure of a management degree (finance, HR, etc).  Conversely, architecture has grown from a technical discipline to include information architecture, solution architecture, business architecture, and information architecture.

Christopher Alexander popularised – if his ideas can be considered popular – the idea of generative sequences.  In essence, a generative sequence is the process of taking a structure and changing it through a series of structure preserving transformations. After each transformation the whole structure is then evaluated to determine if the transformation has – more or less subjectively – improved the structure.  This process is repeated.  Alexander also defines the so-called structure preserving transformations that are applicable at each step.

This is an interesting analogy to decision making in organisations.  Each time a decision is made the structure of the organisation changes.  Structure in this case may refer to anything: including the attitude of an individual team-member, the next task to focus on, or even quite literally a change in the organisation’s structure as we usually use that term.

What is interesting is that from both the manager’s point of view and the architect’s point of view the details of that structural transformation are only selectively considered.  Because managers are generally outcome focused, each transformation or decision is evaluated based on its perceived contribution to outcomes.  These outcomes may be long or short-term, or they might be project-focused, or they might relate to the entire organisation – but it’s the outcome that’s important.

While management is primarily concerned with outcomes, the architect is concerned with structure as a whole.  When decisions are made they not only impact the progress towards goals but they may also potentially impact other structural elements of the organisation.  Rather than a distinction between long or short-term time horizons, or between technical and business domains, the distinction between architecting and managing is generally about outcomes versus structure.

Currently, it’s difficult for architects to evaluate the impact of a transformation in terms of the progress towards desired outcomes because a comprehensive view of the desired outcomes is rarely shared, documented, or linked to the structural elements as defined by the architect.  Similarly, it is difficult for a manager to utilise the models created by the architect to make decisions because the models which describe the structure use technical language and contain much that is irrelevant to decision making.

This battle is not yet won, of course.  To be a successful architect you must manage carefully, and to be a successful manager you most certainly need to be an architect of sorts.  The MWT Model is driven from the theory that this battle will and is ultimately changing the practice of management itself.  This may be seen as victory going to the architects but is more likely to mean that successful architects will no longer be able to choose what issues they avoid.

While this might be interesting to professionals on both sides of the battle I’m just as interested in how important this is to the organisations that we work in.  As I’ve said before, I believe good IT is structural – when you implement an HR system that enables you to re-deploy some employees in the HR branch of the org chart, you should really hang that HR capability embedded in the IT system in their place.

As these structural IT changes are increasingly differentiating organisations and brokering their relationships with customers it is ever more important that organisations can effectively operate and enhance these technology-enabled capabilities.  Both managers and architects currently struggle to achieve this and in the organisation of the future (now?) it’s really the only game.


TRIZ as a pattern language for problem solving

After watching Merlin Mann talking about the idea of creativity patterns today, I started reading through volume 16 of Make magazine.  On page 57 there is a brief but interesting article about TRIZ.

TRIZ is an evolving set of patterns for problem solving which was first developed in 1946 by Russian Genrich Altshuller.  It doesn’t appear to be fully available on-line.  In fact, it doesn’t appear to be fully developed into a single definitive framework at all.  Nevertheless it appears to have some interesting things to say about problem solving.

From the Make magazine article:

…Altshuller observed that the same problem types appeared time and time again, and yielded to corresponding generic solutions…

This echos with what Merlin was saying about creativity patterns in his Macworld talk.  But how is this different to just knowledge?  It isn’t, I guess, but these sort of endeavors open up the scope of what we believe it is possible to have organised knowledge about. They also standardise how knowledge is organised.  Standardisation is important (in the sense that it allows efficient communication).

The essence of the TRIZ system appears to be utilising a number of patterns of interventions into a system in order to remove a constraint (or ‘Contradiction’) without compromising the system. To me this feels similar to the generative sequence approach of Christopher Alexander.

Alexander’s generative sequences unfold using a standard set of interventions into structure (which is system-like, I guess).  Also, when the intervention is made you effectively check for ‘compromise’ when you re-evaluate the degree of life / wholeness in the resulting structure before moving on.

This idea of removing ‘contradictions’ in TRIZ  also has echos of all that Ayn Rand ‘check your premises‘ talk.

Interestingly, I have always found the idea that contradictions don’t exist as very helpful in problem solving.  Basically, I see them as a signal to break something up into smaller and / or different parts.  I see disagreements between people as similar types of problems – and the breaking down of concepts into more elementary components usually means agreement can be found.

Of course, it doesn’t really matter if contradictions do or do not exist – the question is whether acting as-if they don’t exist is a useful problem solving tool.  And I think it is.

So, patterns can help problem solving.  And patterns are a ‘just‘ a way of organising knowledge.  And problem solving is about making system or structure interventions and then making sure you haven’t destroyed the integrity of the whole.

Management of knowledge work is a lot like problem solving.  So it’s likely to be pattern-based too.  MWT Collaboration Architectures are patterns…

Merlin Mann discovers Christopher Alexander, stops procrastinating, and fumbles a little

Merlin Mann stops talking about productivity and starts thinking about creativity. I’ve always had the nagging feeling that the GTD 43Folders clan were all really just procrastinating. He’s also discovered Christopher Alexander.

Powered by WordPress & Theme by Anders Norén