Nothing like the real thing, he thought. Through all the simulations they had a vague feeling they had left something out. While this was normally the case in practice runs, the consequences could have a dramatic effect on the plans for subsequent activities. Too many conditional branch options for comfort.
Easy to just constrict the pipes to simulate loss of bandwidth to the outside. Re-creation of likely intermittent interruptions or high packet loss, more challenging. The biggest challenge? Uncertain latency.
On the most abstract level, primitive influence could be very helpful. To the degree it repeated across classes of objects a layer below, it served to verify working assumptions or exclude them as appropriate.
The fluid nature of these abstractions, a large vulnerability, though. Consistency at these levels does not require an 'out there' corollary. It is rare that such a construction can occur. And when it does … hold on.
They had agreed to let a third party choose the timing and other key aspects of the simulation so as to make it 'realistic'. Now that even that third party was not in control their predictive ability was in for a real shakedown.
"Aren't you the one who can't tell the difference?"
"Wait a minute, I was confusing the names, but not the associated content." As quickly as he had finished saying it he knew he had dropped the communications ball and he would need to proceed accordingly.
No comments:
Post a Comment