Re: DIS: Re: BUS: Proto: Generalize security

2007-10-17 Thread Ed Murphy
Zefram wrote: Ed Murphy wrote: Amend Rule 1688 (Power) by appending this text: A secured change CANNOT be performed except as allowed by an instrument with power at least as great as that of the rule defining that change as secured. I'm dubious about doing it this way round.

Re: DIS: Re: BUS: Proto: Generalize security

2007-10-17 Thread Zefram
Ed Murphy wrote: I'm not sure that this actually does what you suggest. What's the flaw in it? Defining a change as secured is equivalent to making it IMPOSSIBLE except as allowed by an instrument etc. That seems pretty much equivalent to mine. The biggest difference that I see is

DIS: Re: BUS: Proto: Generalize security

2007-10-17 Thread Zefram
Ed Murphy wrote: threshold defaults to the securing rule's power, but CAN be altered as allowed by that rule, up to a maximum of 1.1 times the securing rule's power. I think it's stupid to allow a threshold higher than the power of the defining rule. Such a threshold is

Re: DIS: Re: BUS: Proto: Generalize security

2007-10-17 Thread Ed Murphy
Zefram wrote: Ed Murphy wrote: threshold defaults to the securing rule's power, but CAN be altered as allowed by that rule, up to a maximum of 1.1 times the securing rule's power. I think it's stupid to allow a threshold higher than the power of the defining rule. Such a

Re: DIS: Re: BUS: Proto: Generalize security

2007-10-17 Thread Zefram
Ed Murphy wrote: If the restriction comes from the securing rule, then circumventing it requires (SR's power + 0.1). If it comes from R1688, then circumventing it requires (SR's power * 1.1). Nope. Either way, circumventing it requires only the SR's power: at that power an instrument can amend

DIS: Re: BUS: Proto: Generalize security

2007-10-16 Thread Zefram
Ed Murphy wrote: Amend Rule 1688 (Power) by appending this text: A secured change CANNOT be performed except as allowed by an instrument with power at least as great as that of the rule defining that change as secured. I'm dubious about doing it this way round. You have R1688