Brick. (a) Bricklaying masonry shall consist of the laying of bricks made from any material in, under or upon any structure or form of work where bricks are used, whether in the ground, or over its surface, or beneath water; in commercial buildings, rolling xxxxx, iron works, blast or smelter furnaces, lime or brick kilns, in mines or fortifications and in all underground work, such as sewers, telegraphs, electric and telephone conduits. All cutting of joints, pointing, cleaning and cutting of brick walls, fireproofing, block-arching, terra cotta cutting and setting, the laying and cutting of all tile plaster, mineral-wool, work blocks and glass masonry, or any substitute for above material, the laying of all pipe sewers or water mains and the filling of all joints on the same when such sewers or conduits are of any vitreous material, burnt clay or cement, or any substitute material used for the above purpose, the cutting, rubbing and grinding of all kinds of brick and the setting of all cut stone trimmings on brick buildings and the preparation and erection of plastic castables, or any refractory materials is Bricklayer's work.
Brick. The verification of our approach [OSA13] was improved using enriched com- ponents with metadata. We proposed an integrated correct-by-construction approach for component contracts using metadata, , which extended our approach for arbitrary components with improved and lightweight side conditions. Metadata are derived from component-contract elements and are used in substitution to heavier verifications in the version without metadata. Additionally, metadata of compositions can be easily derived from the meta- data of its constituting components. As a result, the effort of verification is significantly reduced (particularly for noncyclic networks). The approach proposed in [OSA13] had some limitations. First, the ben- efits of using metadata were limited to the application of composition and feedback composition rules. Although this corresponds to two of the four basic proposed composition rules, the application of the other composition rules is compatible with our strategy with metadata. Moreover, one of these composition rules, the interleave one, is already very simple, and does not need further improvements. In other words, although systematic, our ap- proach was not local for cyclic communicating systems, potentially present- ing a state explosion in the verification of such systems. This drawback can make BRICK inapplicable to complex cyclic systems. This and other new scientific contributions of this deliverable are summarised as follows. • A strategy for local deadlock anaysis of cyclic networks based on ad- herence to some formalised behavioural patterns. • A compositional strategy for livelock analysis. • A notion of service conformance that ensures the preservation of the constituent functional behaviour that does not involve interation with other constituents. • A notion of substitutability at the BRICK level, allowing replacing constituents with valid refinements. • An approach to handle analysis involving timed behaviour. More details on each of these topics are given in the reminder of the introduc- tion; each topic is the subject of a separate section of this deliverable. Our strategy to do local deadlock analysis of cyclic models through adherence to communication patterns is presented in Section 4. This work is based on Roscoe’s solution based on architectural patterns that reduce the verification effort, by allowing a local analysis of deadlock, even for cyclic communication topologies [Ros10]. However, some of the existing architectural ...
Brick. Systematic Development of Trustworthy Component Systems In this section, we discuss the theoretical background of the report. We present the basic definitions and the composition rules in Section 2.3.1. The extended counterparts of the definitions and the composition rules are pre- sented in Section 2.3.2. A full account on the theoretical background can be found in our previous deliverable Deliverable D24.1 [OSA13].
Brick. In this section, we also formalise in some of the existing architectural patterns in the literature. The client/server pattern, which is used for archi- tectures where components interact in a client/server fashion, i.e. the server provides services that are requested by clients, is described in Section 4.2. In Section 4.3, we present the resource allocation pattern, which can be used to model systems where constituents are competing for some shared resources. The version that we introduce here is a slight modification from the one presented in [ASW14]. This pattern allows local deadlock analysis of one of our case studies, the leadership election, discussed in Section 3.2. As a further contribution, we also present the formalisation of a novel pattern. In Section 4.4, we present the async dynamic pattern, which can be applied to networks with systems with two types of entities: the participants and the transport layer. In this architecture, the participants of the system do not interact directly with each other, but exchange messages via the transport layer. The benefits of using behavioural patterns for local deadlock analysis are demonstrated with the application of this strategy to the two case studies introduced in Section 3. Our experiments, whose description and results can be found in Section 4.5, considered the asymmetric dining philosophers and the leadership election examples and demonstrated that both specifications are deadlock free using our extended strategy. Finally, Section 4.6 presents a discussion on issues and ideas that we consider important in the design of a tool that would automate the extended strategy, which is in our research agenda.
Brick. In the sequel, we introduce three patterns that we integrate to : the resource allocation pattern, the client/server pattern and the async dynamic pattern. The first two have been introduced in [Mar96, Ros98] and detailed formalised and automated via refinement checking assertions in [ASW14], and the third one has been originally proposed, formalised and automatised in [AOS+14b]. Additionally to presenting the patterns, we instantiate for BRICK each of them the INIT predicate. This instantiation together with the frame- work presented, represents the integration of the patterns into . The presentation of the patterns also help the reader to get a full understanding of the pattern-based framework.
Brick. The pattern also restricts the behaviour of constituent systems concerning how requests and responses are performed. This restriction is not imposed on the complete behaviour of the constituents, but in a particular subset of it. As deadlocks can only happen in constituents due to some kind of ill-interaction the behaviour that needs to be restricted to avoid this B problem is the one related with interaction between constituents. Hence, we use an abstraction function in the behaviour of the constituents to conceal the events that are not linked with synchronisation and therefore cannot participate on a deadlock. This abstraction function is given by Abs(Ctr ), which is defined below as the projection of the behaviour of Ctr ( Ctr ) on the channels that are used for interaction between constituents, i.e. the channels in C. Definition 4.7 (Abstraction function) Let Ctr be an I/O contract. The abstraction of the behaviour of Ctr considering the behaviour related to in- teraction between consituents is given as follows. ^ Abs(Ctr ) = BCtr † (Sc:CCtr {| c |}) The abstraction of a constituent, conforming to the client/server pattern, must be initially offering request events. Once a request is performed, it can behave in several ways, according to some conditions. If the request performed demands no response, then the process must offer, again, some request event. If the request demands a response, then there are two cases to consider, when this performed request is a server one and when it is a client one. In the case of a server request, the process must answer this request with at least one of the possible responses. In the case of a client request, the process must be able to accept all expected responses. The specification of this behaviour is given by the following process, which also has to deal with the replicate internal choice undefinedness for empty sets. Note again the use of acknowledgements after events for buffer tolerance purposes. Definition 4.8 (Behavioural server responses specification) Let Ctr be a contract defining a constituent system, and Ctrs the set of constituent con- tracts initially available for composition. RequestsResponsesSpec(Ctr) = let cEvs = Union({clientRequests(Ctr,Ctr’) | Ctr’ <- Ctrs}) sEvs = Union({serverRequests(Ctr,Ctr’) | Ctr’ <- Ctrs}) ClientRequestsResponsesSpec =
Brick. The resource allocation pattern can be used to model systems where con- stituents are competing for some shared resources. The version that we introduce here is a slight modification from the one presented in [ASW14]. The reason for these modifications is that the strategy makes some assumptions that are not valid in the aforementioned work. For instance, communication between systems in the mentioned work is done by event sharing, nevertheless, in , event sharing between systems is not al- lowed. Hence, we perform a few minor changes to cope with some of the differences in the settings. To begin with, we present the elements of interest of the pattern, which must be identified by the user of our strategy. These are: • Users: the set of components of the systems that behave as users
Brick. The building exterior will be a combination of brick, stone cladding and architectural precast trim pieces and an aluminum framing and glass system. The brick will be Utility (3-5/2" x 3-5/8" x 11-5/8") and Norman (3-1/2" x 2-1/4" x 11-1/2") sized brick with plain uncolored mxxxxx on structural stud backup.
Brick. Wire cut or split faced.
