Software ecosystem, greenwash or new understanding?

Posted on February 17, 2008


Last week Computerworld reported IDC’s annual top ten predictions for the NZ computing industry. Number one is the impact of the skills crisis in NZ computing slowing investment. Number five is:

Green IT – issues behind the hype will shine through

What interests me most is their repeated use of the term “ecosystem”:

3. Emergence of the software as a service ecosystem – proposition will convert into penetration


7. Will the future of the channel ecosystem be revolution or evolution?

The report itself is behind a paywall so I can’t really tell how IDC is using the term ecosystem – is this greenwash (intentional or accidental) or an indicator of a greater understanding through a biological metaphor? The term software ecosystem or anything-really ecosystem is suddenly everywhere (google news search). I’m suspicious that people are borrowing a flag and waving it in the hope of being seen to meet green standards.
It is five years since David Messerschmitt and Clemens Szypersk’s Software Ecosystem, a book that renamed the software industry the software ecosystem and examined the complexities of our business in terms of larger systems. Like Tate though, Messerschmitt and Szypersk don’t go far beyond the emphasis on the interconnectedness of a variety of issues surrounding software development.

In the preface Messerschmitt and Szypersk argue:

The software industry is itself very complex, with many complimentary products necessary to for a systems solution and complex alliances and standardisation processes needed to meet the needs of numerous stakeholders. Together, the software suppliers, standardisation bodies, content suppliers, service suppliers, and end-user organisations form a complex web of relationships. The “ecosystem” metaphor is truly descriptive.

The overarching theme of the book is that software is different.

We’ve looked earlier at the value of metaphors in computing (biomimicry, Tate). Primary among the benefits are a common vision, a shared vocabulary, architecture and generativity. All of these, especially the last, benefit most from actually using the metaphor – borrowing from the idea to create new understandings. It is pointless to have a metaphor and then ignore it. M&S recognise the value of cross disciplinary thought:

it is interesting to compare software to other industries, looking for parallels that might offer insights to the software industry


some insights and perspectives on software from the economics profession are described in chapter 9.

Nowhere do Messerschmitt and Szypersk actually use the ecosystem metaphor. And they are missing so many opportunities. They do take something like a systems approach – they look for simple laws to describe flows of information and then try to consider these in larger systems. On page 34, for example, they examine Moore’s law and how it might operate over different scales.

Since previous technologies (like transportation or electrification) did not advance according to Moore’s law, there must be something distinctive about the material information technologies

It is a shame that a whole field of environmental science is ignored: landscape ecology. It has much to say about scaling effects. Missed opportunities for new understandings continue:

  • there is discussion about heterogeneity without reference to biodiversity;
  • specialization and change without evolution;
  • networks with out food webs;
  • value chains without food chains;
  • productivity without photosynthesis;
  • satisfaction without vigour;
  • models of development and maturity without succession;
  • community without community;
  • competition without competition;
  • change without disturbance;
  • distribution without patches;
  • services without services;
  • components without demography;
  • cooperation without relationships;
  • interfaces without edge effects;
  • business structuring without community architecture;
  • convergence without convergence;
  • migration without invasion;

In addition to missed opportunities, things seem wrong from an ecosystem perspective. “Environments” are described that in the various chapters that stop at the boundaries of that chapter (users, hardware or network). A model of the internet is provided where the heterogeneous internet consists of separated (but communicating) homogeneous environments. Other factors crucial to ecosystems are missing entirely: stochastic events, feedback, limiting factors.

Messerschmitt and Szypersk perceptively state:

The daily operation of most software systems requires human intervention

…these functions are called system administration.

Again, we could be so much further ahead if they then went on to borrow from from the field of human ecosystems. Instead they have a chapter (9) on economics. There, the abstract discussions on supply, risk, complementarity, competitive advantage, time, and re-use are all crying out for some actual consideration of ecosystems.

David Messerschmitt and Clemens Szypersk have won several awards for Software Ecosystem. It does a great job of getting people to see the bigger picture. It does not, however, benefit from the insight the title might suggest. It entirely ignores ecological science, it does not even vaguely approach sustainability.

So, with the term software ecosystem suddenly gaining prominence along with the recognition of green computing, I suspect people are putting two and two together and getting five. Feel free to use the word, I’m not a word-use Luddite (despite ecosystem being explicitly coined for the ecological sciences), but please don’t think that by putting ecosystem behind everything means you are acting sustainably. The term greenwash comes to mind.