Wednesday, September 17, 2008

Standards and Innovation

The emergence of standard components for ubiquitous and common activities does not impair innovation but accelerates it. This is the principle of componentisation; the speed at which a system develops is directly related to the level of organisation of its subsystems.

Mashing up web services to create a new application depends upon layers of standard components, including standard protocols, operating systems, CPUs etc. If there were no standard components and you had to start with chip design before writing an application, the rate of innovation on the web would be snail like.

The only people who oppose standardisation are generally those who sell their product on the basis that it provides a differential. Of course, an activity which is ubiquitous is not a differential in any sense of the word. The arguments against standards are false.

"Cloud" computing is ripe for emerging standards and those standards won't hamper innovation, they will accelerate it.

2 comments:

Bjoern Negelmann said...

Nice thought - Simon - but there are ways to many standardisation approaches in the software/web industry - every second there are new initiatives because of the people in this industry. Everybody wants to be famous and be the father of somekind of standard.
IMHO this will keep us limited towards the potentials that we might get out of cloud computing.

Greetings from Cologne. Bjoern

swardley said...

Hi Bjoern,

Good to hear from you, I hope all is well.

"Every second there are new initiatives" - agreed, that's why I talk about emerging standards - some of those will become de facto, most will be disappear into the bin of history and in certain cases some de jure standards will need to be enforced on a failing market.

As for limiting potential, the standardisation of ubiquitous activities is a natural consequence of those activities becoming well defined and understood. This standardisation does not limit innovation but accelerates it by creating order for that which does not truly matter.

This is the basis of the theory of hierarchy and componentisation (Herbert Simon, circa 1960s).

The commoditisation of higher orders of the IT stack and its provision as well defined standard components will create unprecedented rates of innovation in the web world, in much the same way that standardisation in I/O, CPU and memory has.

As for people wanting to be famous, well, all of this stuff has been outlined a long time ago. Many of these ideas are simply a rehash of cybernetic management concepts.

Even Carr (commoditisation of IT), McAfee (emergent organistions) and Davenport (process engineering) are simply advancing aspects of the same transition from innovation to commoditisation.

However, you are right that the mad scramble to become famous and known creates unseemly noise. Eventually, that activity will die down and become boring, dull and standardised and the scrum will move onto the next new thing.