Object Constraint Language (OCL) is part of the Unified Modeling Language (UML) specification and can be used to enforce constraints on the attributes or methods of a class. It would greatly help the software developers if such non-executable OCL constraints specified in a UML model could be enforced on the executable code generated from the model. This thesis discusses the concepts, ideas and the approach in transforming a model developed in the Rational Rose software with OCL constraints into Java code shells, complete with fragments of code to detect the run-time violations of the constraints. The implementation and testing of a prototype tool that incorporates these ideas is also discussed.

Object Constraint Language (OCL) is part of the Unified Modeling Language (UML) specification and can be used to enforce constraints on the attributes or methods of a class. It would greatly help the software developers if such non-executable OCL constraints specified in a UML model could be enforced on the executable code generated from the model. This thesis discusses the concepts, ideas and the approach in transforming a model developed in the Rational Rose software with OCL constraints into Java code shells, complete with fragments of code to detect the run-time violations of the constraints. The implementation and testing of a prototype tool that incorporates these ideas is also discussed.

The aim of this Master Thesis named Transformation from OCL into SQL is to explore possibilities of enhancing Enterprise Architect by the ability of generating SQL code that implements OCL constraints of a Platform Independent Model, and to create a working prototype that demonstrates this capability. The thesis contains a brief introduction to the problem of Model Driven Architecture, why it would be useful to implement the specified constraints directly in the database, a description of the interaction with Enterprise Architect along with analysing how to get the needed data, the way OCL can be mapped onto SQL, and finally it describes the implementation of the OCLtoSQL plugin from its analysis and chosen architecture to its documentation.

Testavimo procesą galima gerinti jį optimizuojant. Vienas iš galimų gerinimo būdų yra vienetų testai. Vienetų testai padeda sumažinti klaidų kiekį nuo 20 % iki 90 %, tačiau tuo pačiu programų kūrimo išlaidos gali padidėti apie 30 %. Vienas iš galimų būdų išlaikyti pasiektą vienetų testų efektyvumą ir palaikyti mažą kainą yra automatinis jų generavimas. Vienetų testų generatorius, generuoja vienetų testus remdamasis OCL apribojimais. Generatorius JAVA programiniame kode įterptus apribojimus generuoja į Junit vienetų testus. / The development of modern software is a difficult process, there is a high possibility to leave uncorrected mistakes in the software, it would be more precisely to say that it is impossible to make software without bugs. Software testing is maybe the biggest part of development process. The unit testing is very powerful testing. It prevents from defects 20% and more [1]. The quality increase came at a cost of approximately 30% more development time. The main goal is create generator for unit testing. Automated test generator can reduce development time. Unit test generation is based on the OCL (Object Constraint Language) and software static model. OCL eliminate a test oracle problem. Software static model can be UML a class diagrams, but it very complicated approach. We use a reflection technology, because it is more precise and better today. Also OCL constrains are inserted in code.

Made available in DSpace on 2016-08-29T15:33:21Z (GMT). No. of bitstreams: 1 tese_8885_JohnMScThesis[final]20150703-110844.pdf: 2078817 bytes, checksum: b7867cf01fac04ad68223c7eeca3fe68 (MD5) Previous issue date: 2015-05-28 / Conceptual models often capture the invariant aspects of the phenomena we perceive. These invariants may be considered static when they refer to structures we perceive in phenomena at a particular point in time or dynamic/temporal when they refer to regularities across different points in time. While static invariants have received significant attention, dynamics enjoy marginal support in widely-employed techniques such as UML and OCL. This thesis aims at addressing this gap by proposing a technique for the representation of dynamic invariants of subject domains in UML-based conceptual models. For that purpose, a temporal extension of OCL is proposed. It enriches the ontologically well-founded OntoUML profile and enables the expression of a variety of (arbitrary) temporal constraints. The extension is fully implemented in the tool for specification, verification and simulation of enriched OntoUML models.

In this work some progressive object–oriented requirement definition methods were analyzed (ICONIX, Reggio, Scores). Rules and principles of these methods were generalized and comprehensive content of requirements model was presented. Requirements definition must define state and behavior of target information system independently of future design. Requirements model was proposed where use cases are presented as interfaces together with associated classes of problem domain, operations are specified with signatures, pre and post conditions wich must be written in formal language OCL. Model was described using UML class, sequence diagrams, state charts and OCL constraints and is illustrated with examples from case study.

The main aim of the system’s analysis is to find the most important information about real world (problematic sphere); and, firstly, that will be data about processes, their structure and data shift. Business rules combines all these three things that were mentioned above, that is why they can be regarded as the central part in the system’s analysis, and the system’s analysis can be applied to business rules. Business rule is a proposition that defines and limits some aspects of the business. Certain business structure is defined by the business rule; additionally, certain business is controlled or influenced by the business rule too. Moreover, business rules do not depend on projecting model or technical platform. Objective application informational system creation methodology was chosen for the analysis; and it was based on objective modelling of the real world objects and the use of that model for the projecting that does not depend on certain programming language. UML language was chosen for business rules modelling because this language is universal of conception modelling and it is an objective programming language. Despite of all advantages of UML, it does not specify business rules as a separate model’s unit. In UML the rules are depicted as comment; in the beginning they are written by the help of natural speech then they are made into OCL (object constrain language). OCL is not a programming language that is why the logics and control of the program cannot be... [to full text]

Šiame darbe sprendžiamos grafinio vientisumo reikalavimų vaizdavimo loginiuose duomenų modeliuose bei jų kodo realizavimo problemos. Sprendžiant šiuos uždavinius, buvo analizuojami vientisumo reikalavimų tipai, UML plėtimo mechanizmai bei esamos vientisumo reikalavimų įgyvendinimo CASE įrankiuose galimybės.Vientisumo reikalavimus UML modeliuose galima užrašyti OCL kalba, tačiau OCL yra per sudėtinga daugeliui projektuotojų. Be to, SQL kodo generavimas iš OCL kol kas nėra efektyvus. Šio magistrinio darbo tikslas yra padidinti UML CASE įrankių duomenų bazių projektavimo galimybes, sukuriant vientisumo reikalavimų vaizdavimo ir kodo generavimo priemones. / Fulfillment of data integrity constraint requirements is one of main problems in information system development process. The constraints are implemented at physical system development level. Despite the fact that most of current UML CASE tools allow to specify constraints using OCL language, for the most of system analytics and architects use of OCL is too complicated. Currently SQL code generated from OCL constraint implementation is not effective. In this study we will present alternative approach for constraint requirement implementation using UML2 extension possibilities – profiles and stereotypes with tag values. Also we will present SQL code generation template language and code template based constraint requirement transformation to SQL code algorithm. It was used to extend UML Case tool MagicDraw.

Tese de doutoramento. Engenharia Informática. Faculdade de Engenharia. Universidade do Porto. 2010

Per tal de garantir la qualitat final d'un sistema d'informació, és imprescindible que l'esquema conceptual que representa el coneixement sobre el seu domini i les funcions que ha de realitzar sigui semànticament correcte.La correctesa d'un esquema conceptual es pot veure des de dues perspectives. Per una banda, des del punt de vista de la seva definició, determinar la correctesa d'un esquema conceptual consisteix en respondre la pregunta "És correcte l'esquema conceptual?". Aquesta pregunta es pot respondre determinant si l'esquema satisfà certes propietats, com satisfactibilitat, no redundància o executabilitat de les seves operacions.D'altra banda, des de la perspectiva dels requisits que el sistema d'informació ha de satisfer, l'esquema conceptual no només ha de ser correcte sinó que també ha de ser el correcte. Per tal d'assegurar-ho, el dissenyador necessita algun tipus de guia i ajut durant el procés de validació, de manera que pugui entendre què està representant l'esquema exactament i veure si es correspon amb els requisits que s'han de formalitzar.En aquesta tesi presentem una aproximació que millora els resultats de les propostes anteriors adreçades a validar un esquema conceptual en UML, amb les restriccions i operacions formalitzades en OCL. La nostra aproximació permet validar un esquema conceptual tant des del punt de vista de la seva definició com de la seva correspondència amb els requisits.La validació es porta a terme mitjançant un conjunt de proves que s'apliquen a l'esquema, algunes de les quals es generen automàticament mentre que d'altres són definides ad-hoc pel dissenyador. Totes les proves estan formalitzades de tal manera que es poden tractar d'una manera uniforme,independentment de la propietat específica que determinen.La nostra proposta es pot aplicar tant a un esquema conceptual complet com només a la seva part estructural. Quan es pretén validar només la part estructural d'un esquema, oferim un conjunt de condicions que permeten determinar si qualsevol prova de validació que es pugui fer sobrel'esquema acabarà en temps finit. Per aquells casos en els quals aquestes condicions de terminació se satisfan, també proposem un procediment de raonament sobre l'esquema que s'aprofita d'aquest fet i és més eficient que en el cas general. Aquesta aproximació permet validar esquemes conceptuals molt expressius, assegurant completesa i decidibilitat al mateix temps.Per provar la factibilitat de la nostra aproximació, hem implementat el procés de validació complet per a la part estructural d'un esquema. A més, per a la validació d'un esquema conceptual que inclou la definició del comportament, hem implementat el procediment de raonament estenent un mètode existent. / To ensure the quality of an information system, it is essential that the conceptual schema that represents the knowledge about its domain and the functions it has to perform is semantically correct.The correctness of a conceptual schema can be seen from two different perspectives. On the one hand, from the point of view of its definition, determining the correctness of a conceptual schema consists in answering to the question "Is the conceptual schema right?". This can be achieved by determining whether the schema fulfills certain properties, such as satisfiability, non-redundancy or operation executability.On the other hand, from the perspective of the requirements that the information system should satisfy, not only the conceptual schema must be right, but it also must be the right one. To ensure this, the designer must be provided with some kind of help and guidance during the validation process, so that he is able to understand the exact meaning of the schema and see whether it corresponds to the requirements to be formalized.In this thesis we provide an approach which improves the results of previous proposals that address the validation of a UML conceptual schema, with its constraints and operations formalized in OCL. Our approach allows to validate the conceptual schema both from the point of view of its definition and of its correspondence to the requirements.The validation is performed by means of a set of tests that are applied to the schema, including automatically generated tests and ad-hoc tests defined by the designer. All the validation tests are formalized in such a way that they can be treated uniformly, regardless the specific property they allow to test.Our approach can be either applied to a complete conceptual schema or only to its structural part. In case that only the structural part is validated, we provide a set of conditions to determine whether any validation test performed on the schema will terminate. For those cases in which these conditions of termination are satisfied, we also provide a reasoning procedure that takes advantage of this situation and works more efficiently than in the general case. This approach allows the validation of very expressive schemas and ensures completeness and decidability at the same time. To show the feasibility of our approach, we have implemented the complete validation process for the structural part of a conceptual schema.Additionally, for the validation of a conceptual schema with a behavioral part, the reasoning procedure has been implemented as an extension of an existing method.