Machine-to-Machine Web
Semantic web (or, machine-to-machine web) is still a pie in the sky. The problem is, as we’ve already discussed, that most of the resource representations that the machines serve in response to requests are highly unstructured.
Machines lack common-sense capability, and are thus completely dysfunctional when it comes to extracting any instructions out of the unstructured content.
Consequently, it is still devilishly hard to automate the machine-to-machine communication on the web.
Why Do Machines Have to Communicate?
Why is it that machines have to communicate? What is gained by two or more machines communicating on the web while humans are on a lunch break, or playing a game of tennis?
I’m not talking machine communication in closed, proprietary circuits, such as a nuclear power plant, or an elevator maintenance system. There, it is indeed beneficial for the machines to communicate, and there are countless proprietary protocols that have been designed to ensure smooth and resilient processing.
I’m talking about machines on the web. Machines submerged into this primordial soup of chaotic, fickle and totally unpredictable content. Why is it that we feel that machines should be able to intercommunicate there without our assistance?
Cart Before the Horse
I really don’t know the answer to this question, and I’m not even sure that there is an answer. But even if there was a definitive convincing answer that yes, machines definitely need to communicate when on the web, it would still be way premature to act upon that need.
There’s a bigger fish to fry right now. I’m talking about building Smart Servants who will be designed to help humans fulfill their goals. These Smart Servants need not focus on talking with each other as much as they must focus on understanding what is it that their human masters are doing and would like to do.
So I suggest we stop putting the cart before the horse, and do the first thing first. Let’s start working on building Smart Servants first.