Ivar Jacobson

Ivar Jacobson: The Engineer Who Drew Software

In 1967, working on telephone exchanges at Ericsson, a young Swedish engineer needed a way to show how messages travelled between the parts of a switch. He drew the participants across the top of a page and let time run downward, with each message an arrow crossing from one vertical line to another. That sketch is now called a sequence diagram, and some version of it hangs on a whiteboard in nearly every software company on earth. Ivar Jacobson spent the next six decades inventing ways to make the invisible structure of software visible.

Ystad, Chalmers, Ericsson

Ivar Hjalmar Jacobson was born on 2 September 1939 in Ystad, on the southern tip of Sweden. He took a master's degree in electrical engineering from the Chalmers Institute of Technology in Gothenburg in 1962 and joined Ericsson's research and development department, where he stayed for the formative twenty-five years of his career. Much later, in 1985, he completed a doctorate at the Royal Institute of Technology in Stockholm with a thesis titled *Language Constructs for Large Real Time Systems* — a title that describes his life's problem exactly.

Ericsson in that era was building AKE and then AXE, the computerised telephone exchange that became the company's defining product, programmed in an in-house language called PLEX. A telephone exchange is a brutal engineering environment: it must run continuously for years, handle thousands of simultaneous conversations, and be modified without being switched off. The software problems Jacobson met there were not academic exercises.

Components, Before the Word Existed

His first structural idea, in 1967, was to build the exchange out of *components* — separately specified blocks of software with defined interfaces, which could be developed, tested and replaced independently rather than woven into a single monolithic program. This is now ordinary practice. At the time it was unusual enough to require argument. His own firm describes the resulting architecture at Ericsson as the greatest commercial success story in the history of Sweden, which is a partisan claim from an interested source, but the AXE system's global reach makes the underlying point hard to dismiss.

Alongside components came notation: sequence diagrams, collaboration diagrams and state transition diagrams to describe how messages actually flowed. Jacobson also helped develop SDL, the Specification and Design Language, which became a telecommunications industry standard in 1976.

The Use Case

In 1986 Jacobson invented the idea he is best known for. Before use cases, software requirements were typically written as lists — the system shall do this, the system shall also do that — with no obvious relationship between the items and no way to check that the list was complete or that anything had been built to satisfy it.

A use case inverts the perspective. It describes the system from the outside, as a complete interaction between an *actor* — a person, or another system — and the software, running from a trigger to a result of value to that actor. *Withdraw cash.* *Reserve a seat.* *Reconnect a dropped call.* Each is a story with a beginning, a middle, alternative paths when things go wrong, and a definite end. The consequence is practical rather than philosophical: because a use case is a thread of behaviour, it can be traced. The same thread runs through analysis, design, code and test, so a team can say what a piece of code is for and whether the thing the customer asked for actually works.

He founded Objective Systems in April 1987 to commercialise the method; Ericsson took a majority stake in 1991 and the firm became Objectory AB. In 1992 he published *Object-Oriented Software Engineering: A Use Case Driven Approach*, the book that carried use cases into the wider industry.

The Three Amigos

In October 1995 Ericsson sold Objectory to Rational Software, and Jacobson joined Grady Booch and James Rumbaugh — three rival methodologists with three competing notations, known collectively and slightly wearily as the Three Amigos. Their merger produced the Unified Modeling Language, the standard visual vocabulary for describing object-oriented systems, and Jacobson's Objectory process was developed under Philippe Kruchten into the Rational Unified Process, or RUP. When IBM bought Rational in 2003, Jacobson left.

Making It Lighter

Everything after 2004 has been an attempt to undo the weight of what he built. He founded Ivar Jacobson International in mid-2004. In November 2005 he announced the Essential Unified Process, which he described as a super light and agile version of RUP, blending unified process, agile and process-improvement ideas; from 2006 he worked with Ian Spence and Pan Wei Ng on EssWork, a framework letting teams assemble their own methods from a small kernel of universal elements plus chosen practices.

In November 2009 he launched SEMAT — Software Engineering Method and Theory — with Bertrand Meyer and Richard Soley, arguing that software engineering needed a genuine theoretical foundation instead of a procession of fashionable methods. The result, Essence, was adopted as an Object Management Group standard in November 2014. He has also worked on aspect-oriented development, publishing *Aspect-Oriented Software Development with Use Cases* with Pan-Wei Ng in 2004, and founded Jaczone AB with Agneta Jacobson in 2000, whose Waypointer tool won a JOLT award in 2004.

Why Ivar Is Called a Genius

The specific ability is notational: Jacobson repeatedly found the representation that made an intractable problem thinkable. Sequence diagrams, use cases and component interfaces are all the same move — take something that exists only as tangled behaviour inside a running system and give it a shape a human can hold in mind and a team can argue about. That is a rarer talent than inventing an algorithm, because a good notation has to be simultaneously precise enough for engineers and simple enough for the customer paying for the work. Use cases succeeded because they were both. He did it from inside industrial practice, not from a university, which is why the ideas survived contact with real projects.

The honest counter-case is substantial and he would probably accept most of it. He is one of three UML authors, and UML itself is widely criticised as a committee product that grew bloated and is now used far more selectively than its designers intended. RUP, derived from his Objectory, became a byword for heavyweight process — the very thing the agile movement organised itself against — and Jacobson has spent the two decades since 2004 building lighter frameworks in explicit reaction to it. Use cases have been partly displaced in agile teams by the shorter, cruder user story. SEMAT and Essence, for all their standing as an OMG standard, have not achieved the industry-wide uptake he hoped for. His genius is real but it is the genius of the toolmaker, and toolmakers are judged by whether the tools stay in use.

Legacy

Jacobson has written ten books, from OOSE in 1992 through *The Object Advantage*, *Software Reuse*, *The Unified Software Development Process* and *The Essence of Software Engineering* to *The Essentials of Modern Software Engineering* in 2019, and published more than a hundred papers. Chalmers gave him its Gustaf Dalén medal in 2003; San Martín de Porres University in Peru made him an honorary doctor in 2009. He remains founder, chairman and chief executive of Ivar Jacobson International. Whatever survives of UML and RUP, the habit of asking who is using this system and what are they trying to get done — the question the use case forced onto every project — is now simply how software is specified.

Achievements

• Educated at Chalmers University of Technology and Royal Institute of Technology
• Worked as computer scientist and inventor