assurances data: hostnew a_t

[sourabhniyogi]

The answer to the question is 1350. This can be derived from interpreting 9.8 as a declarative statement with a_i, a_o and a_t as dependent on c and l

So in new specification the b: a_t apparent circularity should be ignored:

in favor of recognizing that the functional dependency is like this:

• $a_c, a_l$ (from new) <=depends= $a_i$ = ..., $a_o$ = ... (see 9.8) <=depends= $a_t$ (see 9.8)

and no "order of operations" is implied whatsoever. Since GP systematically appeals to declarative programming languages, this type of reasoning should be taken a general principle ("No specific order is needed. Ever.") rather than specific to new.

Since most implementers are likely to be using imperative languages (to set up an object "a", then initialize its attributes in some particular order), it is extremely likely that future debates will be settled with reminders of this general principle, and exceptions to the principle should be noted as actually needed.

[sourabhniyogi]

Matched the 1350 but added 500000 "manually" here in hostNew to simulate a disabled hostTransfer, which we will add back after addressing "trap" issue.