Trigpoint
Collections › Software & Systems Architecture › Modularity and thin crossing points
“Cut along function and the boundary is a good one”: it is not that simple.

Modularity and thin crossing points

Modularity & Thin Crossing Points · Baldwin & Clark (2000) / Baldwin (2008)
Bibliography checked ✓ Software architecture · siting boundaries Read first: Don't repeat yourself (DRY) Tell the editor privately ▸
In three sentences Where a boundary should open is settled by the cost of the traffic across it, not by which category the work falls into. The test is one sentence: can the traffic across this cut be defined, counted and priced. If it cannot, do not open an interface here.

↳ Seen a counter-example?

Holds when
When deciding where to cut a piece of work or a system, and how hard each exchange is can be observed in advance.
Evidence
Conceptual framework; Bibliography checked
Breaks down when
Assuming finer is better and never revisiting a cut: shared rules become hard to change, and boundaries drift out of place as traffic changes.
Modularity and thin crossing points Baldwin & Clark (2000) · Baldwin (2008) Cuts cost less where traffic is sparse and priceable Vertical: how frequent and iterative the traffic Horizontal: how easily it is defined and priced Dense territory many round trips, re-explained each time and no clear definition of done an interface here institutionalises friction Frequent but definable many trips, but inputs and outputs are clear an interface works, but must be maintained Sparse but hard to price rare traffic, but murky when it happens define the handover first, then cut A thin crossing sparse, definable, countable, priceable mundane transaction costs generally low open the interface here High Low Hard Easy the cells are continuous; judge by observing actual traffic, not by impression Figure: trigpoint.pub
Figure 1 · Two axes decide where a boundary should open. The horizontal axis is how easily the traffic across that cut can be defined and priced; the vertical axis is how frequent and iterative that traffic is. The bottom-right cell is a thin crossing point, where an interface is cheapest to open. The top-left cell is dense territory, where forcing an interface institutionalises the friction. What lies between is continuous, not two cells. The quadrants are conceptual: no costs, counts or measured numbers. View full size ▸

1. What happened

The intuition behind modularity is old: cut a complex system into pieces, let each piece govern its own inside, and have the pieces deal with one another through agreements.

What Baldwin and Clark's 2000 book Design Rules did was something else: it raised modularity from an engineering intuition to an economic theory[1]Baldwin, C. Y. & Clark, K. B. (2000). Design Rules, Volume 1: The Power of Modularity. MIT Press. ISBN 9780262024662. Bibliographic details confirmed by two sources. The split between design rules and hidden modules and its option value follows secondary accounts that agree with one another. Raises modularity from an engineering intuition to an economic theory: visible design rules are separated from hidden modules, so the modules form a set of real options. The book's content was checked here only through several agreeing secondary accounts.. The core structure divides design information in two: visible design rules that everyone has to obey, and implementation hidden inside a module. After that division, each module becomes a real option that can be replaced independently. You can swap one piece to try something without touching the whole. So the value of modularity is not only that it is easier to maintain. It is that you can run many local experiments cheaply.

What pins the criterion down, though, is Baldwin's 2008 paper. It asks a question that comes earlier: why does a boundary fall here rather than there. The answer is the cost of traffic. In a network of tasks, the places where exchanges are easy to define, easy to count and easy to pay for are where those mundane transaction costs are generally low, and module boundaries (along with firm boundaries) correspond to those thin crossing points[2]Baldwin, C. Y. (2008). Where Do Transactions Come from? Modularity, Transactions, and the Boundaries of Firms. Industrial and Corporate Change, 17(1), 155–195. Bibliographic details checked against the official record. Consistent with the abstract. From the abstract, verbatim: “The costs of this work (labeled mundane transaction costs) are generally low at the thin crossing points of the task network, which correspond to module boundaries.”. In regions where the traffic is dense, iterative and hard to price, forcing an interface into the middle institutionalises the friction.

The criterion has one very usable property: you can apply it before you start. Classification by function only becomes visible once the system exists. The cost of traffic can be observed in advance. You do not have to cut wrongly once in order to learn where the cut belongs.

2. Why this matters to you

Most of the boundary decisions you make daily are cuts along which category does this belong to. This page asks you to cut along how expensive is this handover instead.

The difference shows up most clearly when handing work to a model. Suppose you want to cut a piece of work out. If the traffic across that cut is I give a definite input, it returns a definite output, and I can tell whether it is right, that is a thin crossing point, and handing it over is cheap, because the handover itself can be defined and accepted.

If the traffic is I have to explain a pile of background, we go back and forth several times, and I have no simple way to judge whether the result is right, that is dense territory. Open an interface there and you will spend most of your time maintaining the interface.

The same criterion explains a familiar frustration: why some work goes slower after it is contracted out. It is rarely that the other party is not up to it. It is that you opened the boundary in dense territory.

Three usable questions fall straight out of the criterion. First, can I state clearly what counts as finished for the output of this handover. Second, how long does accepting it take me, and does that beat doing it myself. Third, how many round trips does this line carry in a week: if the answer is many, and each one requires explaining things again, the line does not belong there.

More lines cut out is not better, either: each one is another handover to maintain, and lines already cut are worth walking through these three questions again from time to time.

Verification layer · Members
The guided layer of this node is open to all readers.
Verification details and failure boundaries are for members.
US$8.25 a month on average · or subscribe monthly ▸

Lab cases Written only where there is evidence

No field record for this node yet.

Common misreadings

Page verification: Bibliography checked ✓ · Content last revised 2026-11-14

  • A common account says modularity was proposed by these two authors. The concept is at least thirty to forty years older (for example, near-decomposability in 1962 and information hiding in 1972). What the two contributed is the economic formalisation; the theory of where boundaries go comes from a 2008 paper by one of them, Baldwin. Do not cite them as the originators of the concept. #
  • A common account says thin and thick are two classes, so judging is classifying. The original's criteria are how dense and complex the traffic is, and how much work it takes to define, count and pay for what crosses. These are matters of degree, and a verdict needs actual observation of the traffic. Treating it as a binary leads people to sort by impression. #
References

How to read the source notes “Bibliographic details checked” means author, year, title and venue were confirmed. It does not mean the source's claims were checked. “Consistent with the abstract” means the claim agrees with the paper's abstract. It does not include sentence-by-sentence checking against the full text. “Follows secondary accounts” means the claim rests on sources that report it, not on the original document. Definitions as of 2026-09-20.

  1. [1] Baldwin, C. Y. & Clark, K. B. (2000). Design Rules, Volume 1: The Power of Modularity. MIT Press. ISBN 978-0-262-02466-2. ↩
  2. [2] Baldwin, C. Y. (2008). Where Do Transactions Come from? Modularity, Transactions, and the Boundaries of Firms. Industrial and Corporate Change, 17(1), 155–195. ↩

↳ Is a citation wrong?

Tell me when a new shelf opens

Tell me when a new shelf opens

Shelf notices are sent only when a new shelf opens or when the reading rules change.

Trial

Tell the editor privately: a correction, a counter-example or where it applies

Submissions are read only by the editor, and nothing is published before they are handled. Once a note is handled, a summary and response written by the editor are published. No sign-up is needed, and the nickname may be left blank.

What are you submitting?
Set a public name to be credited by name ▸
Submit anonymously, without signing in

Signing in lets you follow the status on your page. What you have written is kept.

You are signed in: this note is linked to your account, visible only to you and the editor, so you can follow its status on your page. The text is never published. Full privacy policy ▸

Submit anonymously instead, not linked to your account ▸

No email is collected. Your original text and nickname are read only by the editor and are never published; the site publishes only a summary the editor writes. The human check is provided by Cloudflare Turnstile. To limit abuse, the system keeps a hash of your IP address, not the address itself. Your original text is stored with our cloud provider and can be deleted on request. Full privacy policy ▸