In Sect. 6.2, we discuss extensions of Angluin’s MAT framework with new types of queries. A learner may for instance ask which previous values and operations have been used by the SUL to compute some output value. Or she may ask if some previous input value may subsequently be tested or output by the SUL. Such queries may dramatically simplify the task for the learner, but can often be simply answered by the teacher using off-the-shelf code analysis tools. An example would be a query about which registers are needed in a specific state or location.
In Sect. 3.2, we derive a multi-user security bound on the SiM mode and explain how it differs from the single-user analysis of [9].
In Sect. 4.3, we reflect on the differences between Elephant v2 and v1 [8], and explain how the security bound has improved.
In Sect. 3 we define what we consider a reduction and what we mean with optimal security. This section also includes a formalization of the generic standard-to- ideal reduction. We derive a lower bound on the strong related-key PRP security in Sect. 4. We revisit LRW2 and Men2 using these formalizations and results in Sect. 5. In Sect. 6 we present a generalized tweakable blockcipher design, and in Sect. 7 we derive our impossibility result on the optimal security of a general- ized tweakable blockcipher. We present an elaborate discussion of the results in Sect. 8.
In Sect. 2.1 we have seen that the order on a poset X forms a non-symmetric comparison relation : X X 2 in PoSets, where ( ) is order-reversal. Now assume that X is an orthomodular lattice (see [18] for details), with orthocomplement ( )⊥ : X X. It satisfies, among other things, x⊥⊥ = x and: x⊥ y iff y⊥ x. When x y⊥ one calls x, y orthogonal, which is also written as x y. We obtain a comparison relation cp : X X 2 in PoSets (with identity involution), via cp (x, y) = 1 iff x⊥ y. By using orthocom- plement in the first coordinate the contravariance disappears. This relation is the same as (x, y) x⊥ ⊥ y⊥, that is, as orthogonality of orthocomplements. It forms a symmetric comparison relation, since orthogonality is symmetric. The resulting category of tame relations is known from [6, 13]. ⊥symmetric comparison cluster X×X
In Sect. 3.1, the iocoF relation was defined as a relation between an implemen- tation, modelled as an IOTS, and a specification, given as an LTS. We lift this definition to the level of STSs by appealing to their semantics. Definition 12. Let S be an STS and P a physical system, modelled as an IOTS. Then P iocoF S iff P iocoF [S ]. Table 4. Deduction rule for transitions ' λ,ϕ,ρ l −−−→ l type(λ) = ⟨ν1, . . . , νn⟩ ς∈U type(λ) ϑ ∪ ς |= ϕ ϑ' = (ϑ ∪ ς)eval ◦ ρ (λ, (ς(ν1 ),...,ς(νn)⟩) ' ' (l, ϑ) −−−−−−−−−−−−−→ (l , ϑ ) 5 On-the-Fly Testing
In Sect. 4.1, we revisit the state of the art on Trunc and re-derive the best security bound;
In Sect. 2, we focus on the clauses about “Right of persons”. If we consider the Right of rectification or cancellation: The controller is obliged to implement the right of rectification/cancellation of the data subject within a period of ten days. [. . . ], this can be expressed in CNL4DSA as follows: if hasRole (User, DataController) and if hasDate (RectifyRequest, Date) and if timeLessThen (CurrentDate, Date+tenDays) then after DataSubject send RectifyRequest then User must rectify Data Quite obviously, if the data subject would like to cancel, the rule is similar.
