Technological evolution

I was recently reading a paper by Gaynor and Bradner about using real options to value modularity in standards. As a concept, it's interesting to think that modularity will increase value by allowing you to create small, simple pieces and then slowly build on them in a structured way. They also present the idea that as you're building things in stages, you can also apply models for pricing real options. That's all nice and well, but this leads me to wonder about the nature of using simple standards and evolving them over time and what sort of implications this might have.

The authors cite Dyson about how "we should not attempt to construct the Internet, but we should act like gardeners providing a conducive environment for growth". The problem here is that even an enlightened gardener may be faced with difficulties in steering a garden's growth. A fine example is HTTP, which lays the foundations of the World Wide Web we know and use today. The adoption was so explosive that right now I have a gut feeling that you can't really do anything about the protocol, since it's so widely adopted. In fact, a new HTTPbis working group in IETF was recently formed to revise the protocol and fix editorial problems. Fortunately the group doesn't apparently have a mandate to make any large changes. But that wouldn't in fact matter that much, since I guess making concrete changes to HTTP is just about impossible. Standardization blokes can standardize all they want, but getting the changes adopted and getting everyone to update their servers and browsers and proxies means that it won't really matter at all.

So in light of this, it might be that modularity is very nice, but using the baby-steps approach of starting with simple and slowly adding things as it grows may be difficult to control if an explosion in the adoption occurs. But then again, designing protocols in a very complex and exhaustive way or running and ruling the process with an iron fist will undoubtedly kill the innovation and interest from around the technology. So, again it seems that everything is about equilibriums. Again...