The languages we choose to design solutions influence the way we think about the problem, the words we use in discussing it with colleagues, the processes we adopt in developing the software which should solve that problem. Therefore we should strive to use the best language possible for depicting each facet of the system. To do that we have to solve two challenges: i) first of all to understand merits and issues brought by the languages we could adopt and their long reaching effects on the organizations, ii) combine them wisely, trying to reduce the overhead due to their assembling. In the first part of this dissertation we study the adoption of modeling and domain specific languages. On the basis of an industrial survey we individuate a list of benefits attainable through these languages, how frequently they can be reached and which techniques permit to improve the chances to obtain a particular benefit. In the same way we study also the common problems which either prevent or hinder the adoption of these languages. We then analyze the processes through which these languages are employed, studying the relative frequency of the usage of the different techniques and the factors influencing it. Finally we present two case-studies performed in a small and in a very large company, with the intent of presenting the peculiarities of the adoption in different contexts. As consequence of adopting specialized languages, many of them have to be employed to represent the complete solution. Therefore in the second part of the thesis we focus on the integration of these languages. Being this topic really new we performed preliminary studies to first understand the phenomenon, studying the different ways through which languages interact and their effects on defectivity. Later we present some prototypal solutions for i) the automatic spotting of cross-language relations, ii) the design of language integration tool support in language workbenches through the exploitation of common meta-metamodeling. This thesis wants to offer a contribution towards the productive adoption of multiple, specific languages in the same software development project, hence polyglot software development. From this approach we should be able to reduce the complexity due to misrepresentation of solutions, offer a better facilities to think about problems and, finally to be able to solve more difficult problems with our limited brain resources. Our results consists in a better understanding of MDD and DSLs adoption in companies. From that we can derive guidelines for practitioners, lesson learned for deploying in companies, depending on the size of the company, and implications for other actors involved in the process: company management and universities. Regarding cross-language relations our contribution is an initial definition of the problem, supported by some empirical evidence to sustain its importance. The solutions we propose are not yet mature but we believe that from them future work can stem.

Polyglot software development / Tomassetti, FEDERICO CESARE ARGENTINO. - (2014). [10.6092/polito/porto/2537697]

Polyglot software development

TOMASSETTI, FEDERICO CESARE ARGENTINO
2014

Abstract

The languages we choose to design solutions influence the way we think about the problem, the words we use in discussing it with colleagues, the processes we adopt in developing the software which should solve that problem. Therefore we should strive to use the best language possible for depicting each facet of the system. To do that we have to solve two challenges: i) first of all to understand merits and issues brought by the languages we could adopt and their long reaching effects on the organizations, ii) combine them wisely, trying to reduce the overhead due to their assembling. In the first part of this dissertation we study the adoption of modeling and domain specific languages. On the basis of an industrial survey we individuate a list of benefits attainable through these languages, how frequently they can be reached and which techniques permit to improve the chances to obtain a particular benefit. In the same way we study also the common problems which either prevent or hinder the adoption of these languages. We then analyze the processes through which these languages are employed, studying the relative frequency of the usage of the different techniques and the factors influencing it. Finally we present two case-studies performed in a small and in a very large company, with the intent of presenting the peculiarities of the adoption in different contexts. As consequence of adopting specialized languages, many of them have to be employed to represent the complete solution. Therefore in the second part of the thesis we focus on the integration of these languages. Being this topic really new we performed preliminary studies to first understand the phenomenon, studying the different ways through which languages interact and their effects on defectivity. Later we present some prototypal solutions for i) the automatic spotting of cross-language relations, ii) the design of language integration tool support in language workbenches through the exploitation of common meta-metamodeling. This thesis wants to offer a contribution towards the productive adoption of multiple, specific languages in the same software development project, hence polyglot software development. From this approach we should be able to reduce the complexity due to misrepresentation of solutions, offer a better facilities to think about problems and, finally to be able to solve more difficult problems with our limited brain resources. Our results consists in a better understanding of MDD and DSLs adoption in companies. From that we can derive guidelines for practitioners, lesson learned for deploying in companies, depending on the size of the company, and implications for other actors involved in the process: company management and universities. Regarding cross-language relations our contribution is an initial definition of the problem, supported by some empirical evidence to sustain its importance. The solutions we propose are not yet mature but we believe that from them future work can stem.
2014
File in questo prodotto:
File Dimensione Formato  
TOMASSETTI.FEDERICO.THESIS.pdf

accesso aperto

Tipologia: Tesi di dottorato
Licenza: Creative commons
Dimensione 3.51 MB
Formato Adobe PDF
3.51 MB Adobe PDF Visualizza/Apri
Pubblicazioni consigliate

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11583/2537697
 Attenzione

Attenzione! I dati visualizzati non sono stati sottoposti a validazione da parte dell'ateneo