The smallest virtual machines will inherit the earth
What if we give each object its own virtual machine? Of course, it would be a teeny, tiny VM, so small that it wouldn’t need any real OS, except for a pretend kernel, and the networking protocols that would enable it to communicate with the outside world. It would contain also a bunch of pseudo service managers, whose only purpose would be to render its object’s functions (I would say “methods and properties”, but I’m sure you can go beyond that) accessible to other objects.
And how would these “canned” objects find each other? Well, isn’t that what the Web 3.0 people is talking about? Granted, these people are weird, but as long as their work is useful I don’t really care. Of course, direct (random) access and URIs that work are a precondition, so we need some sort of indexing service that maybe distributed over the net. The “Web of all data” that people are talking about (again, sometimes weirdly) would be definitely useful at this point.
At some point we could be even thinking about it in a purposeful way, which would enable us to come to an elegant, minimalistic reengineering of the infrastructure our objects need, but right now we can get by with TCP/IP and all. What we need kinda urgently is a free, public, open cloud, that is extremely easy to use (without loss of generality)!
The most important question for last: after all this is put in place, what would these free objects do? Oh, nothing really important, individually.
Powered by Facebook Comments