A New Approach for Multi-View Models’ Composition using Probes

Event

ABDELALI EL BDOURI*, CHAIMAE OUALI-ALAMI, YOUNES LAKHRISSI

SIGER Laboratory,

Sidi Mohamed Ben Abdellah University,

Fez,

MOROCCO

are introduced as a modeling concept for event detection and control, with basic categories and methods for

projection, derivation, and composition. The VUML Probe profile is presented, integrating probe stereotypes

into the VUML meta-model, with conformance rules to maintain semantic consistency. The application of

probes in the VUML process is demonstrated, in particular, to ensure the autonomous evolution of model-

views. Abstract probes are defined during the composition phase, and then used in view models. Finally, related

work and avenues for future research are discussed, including language enhancements, integration with aspect-

oriented modeling, and tool development. In summary, the paper offers a comprehensive framework for

integrating event observation mechanisms into the VUML approach, aimed at improving the modeling and

management of complex software systems.

1 Introduction

The critical aspect in dealing with the intricacy of

large software systems is the concept of "separation

of concerns". This practice is indispensable to

maintain manageability throughout the development

process, the resulting models, and the code.

Achieving separation of concerns can be executed

through various methods, but the ultimate objective

remains consistent: the identification of relatively

autonomous "components" that can be allocated to

based modeling, [1], which is a variant of the

object-oriented modeling approach tailored for the

examination and design of expansive and intricate

systems. This approach concentrates on the

individuals who interact with the system and

disassembles the specification by their specific

the behavioral aspects of views—how they react to

invocations and how they collaborate to accomplish

the behavior of multi-view objects—the VUML

methodology primarily addresses the structural

elements linked to view composition and data

sharing.

composition of models was addressed, [2].

Nonetheless, when delving into the composition of

behavioral models, the issue becomes more

composition than conventional interface-based

composition. We illustrate that the novel behavioral

specification architecture, grounded in event

observation, yields favorable outcomes concerning

slice connectivity and support for evolutionary

design, particularly the addition of slices.

on event observation as the primary method for

handling first-class object interactions. This was

primarily achieved through the formalization of

have developed new techniques, including the

projection, derivation, and composition of probes, to

enhance the precision of probe-related concepts.

Our proposal is structured as a profile named the

"VUML Probe profile."

we establish the backdrop for our research, focusing

on view-based modeling exemplified by the VUML

profile. Additionally, we provide an illustrative

example of its practical application. Section 3 delves

solution with current strategies, formulating

conclusions, and outlining potential directions for

future research.

diagram depicting the multiview class "Car" is

provided in Figure 2. This class includes a "base"

section, stereotyped as such, accessible to all actors.

Additionally, it contains "views," stereotyped as

"view," representing parts specific to each actor

(Client, Manager, Mechanic).

3 Probes

A probe is a modeling construct used to detect and

control events that are important for achieving a

specific objective. We utilize probes in particular to

simulate implicit interaction between objects. The

existing modeling elements, particularly in UML,

cannot be used to concretize the goals of the probe

event notion. We lay out the foundational ideas for

comprehending our proposal in this part.

with supplementary properties while consistently

observing the same event type as the parent class.

We offer a second, more difficult notion that makes

use of the composition mechanism to customize a

class of probes. Composite classes can

simultaneously observe various events.

After a thorough examination of various events

occurring during system usage (as illustrated in

option to employ them as preconfigured classes in

their models. By instantiating each library type,

concrete probes can be created. As mentioned

earlier, these probe types can be tailored to specific

contexts through principles such as projection,

derivation, and composition.

manage data either at the model level (M1 level) or

the meta-model level (M2 level). For example, the

class type designates its membership class, an

the "startRepair" signals. To implement this

observation, we proceed with the following steps:

the observed event type, specifically the

"SignalSendProbe" type. The resulting

instance is labeled as "startRepObs." This is

accomplished by defining an

"InstanceSpecification" UML element and

assigning the stereotype "probe" to it.

the class and subsequently annotating it as

«probeUse».

behavioral description of the class,

particularly within the state machine

compatible with the class.

describe the behaviors of "MyClass" can be found in

Figure 5. The definition of the "startRepObs" probe

an event of that particular type occurs within the

system. However, the use of the "filter" attribute is

not obligatory since there are instances where events

need to be filtered based on contextual information.

To apply additional constraints for such cases, these

limitations must be defined and met when the probe

is activated.

"SignalSendProbe_testFunction" class is presented

in Figure 6(c).

We have discussed two methods for modifying

basic probes thus far. The first method is projection,

which employs a filter to customize a probe

according to a specific situation using data and

metadata events (see Section III.D). The second

method is probe derivation, which offers the option

meta-model detailed in [9]. A summary of the

abstract syntax associated with the proposed profile

can be observed in Figure 8.

The syntax "wait(obs)" in a state-machine

transition signifies that the Wait trigger is

awaiting the activation of the probe "obs" to

traverse this transition.

used to describe an event.

Reception and the stereotype "probeUse".

interdependence between the behaviors of the

model-views. To address this challenge, we have

proposed an observation approach based on event

probes.

Fig. 9: Using probes in VUML process

need precise information about when the car repair

process begins while developing the state machine

class, after structural amalgamation with a solitary

actor (the client), is delineated in Figure 9(c)-top.

The Car is exhaustively delineated by the

foundational class "CarBase," shared uniformly

across all perspectives, along with an assemblage of

perspectives, each explicating the attributes relevant

to a particular actor.

"repairStartProbe" probe is illustrated after the

antecedent parameter concretization that was left

This research builds upon the findings presented in a

event probes, along with an exploration of their

integration into the VUML profile.

employed in diverse domains, including formal

verification, software testing, debugging, and

various aspect-driven modeling techniques. To our

knowledge, object-oriented modeling languages

of event observation was first introduced as a

method for formally specifying attributes in the

Véda verification tool, [10]. This idea was

subsequently extended to tools for verifying

[11], [12]. While our work draws inspiration from

integrating the concept within a robust modeling

language like UML. This enables the modeling of

event types, including meta-parameters, as well as

refinement and composition processes. Notably, our

of event types (primarily data changes), our

approach introduces a novel perspective within the

UML framework. The exploration of decentralized