OCommWEBIST 2016 Abstracts

Short Papers
Paper Nr: 1



Maria Jose Escalona, Manuel Mejías Risoto, Laura García Borgoñon and Julián Alberto García García

Abstract: N owadays software systems have become increasingly complex. Developing quality software products, with an acceptable time-to-market and competitive cost is the main challenge which many software organizations constantly face to. Besides, we have to add to this challenge the fact that time when the product development was completely located in a specific and concrete place is over. The access to the most important and main international experts has become easier in this global situation where we live. In the software context, it is especially evident that the technology is continuously in change and evolution. Software organizations must manage the complexity of their products, due to the inherent properties of their context: the participation of multiple development teams in different sites in the world, distinct knowledge, and diverse culture and tools, are elements to evaluate when new ways of working, also called processes, are established. In other words, software processes should include the set of technologies, best practices and strategies that allow us to manage effectively the development made by great engineer teams along the products lifecycle. There are many standards and reference models which are used by organizations as guidelines to build their processes, like a guarantee to be able to survive successfully in this environment. Thus, processes have been considered as one of the most important assets in an organization. At the time of defining the processes, organizations want to use increasingly a specific modeling language in order to carry out this task. But the decision of what is the most adequate modeling language is not trivial. There are a lot of modeling languages created for this purpose, the link established between processes and language is highly strong, and an organization faces to a kind of "Babel Tower" since, they have not only multiple choices but the possibility of these languages can understand each other is faraway. This situation is especially evident when two or more different organizations (enterprises or department inside the same enterprise) want to develop products together. This paper is motivated to solve this problem: Why is it necessary to choose exclusively one process modeling language? Why can we not use the most adequate language in each moment? Our objective is to get something similar to what Esperanto pursue in the field of linguistics. In order to achieve it, we have developed a reference framework to facilitate the interoperability and maintainability of software process models by taking advantage of the Model Driven Engineering paradigm. We aim to define a framework where organizations could include the modeling language that they have used during the definition of their processes, and interoperate seamlessly with other ones. Our proposal is supported on three pillars: a base software process modeling language, named INROMA, which embodies the minimal and necessary concepts of processes; the method to include a process modeling language into the framework; and, finally, the transformations which formalize the use of the framework. The three pillars are based on standards like ISO/IEC TR 24744, MOF and QVT. This theoretical framework has been implemented in practice as MONETA, a CASE tool that provides support to use the reference framework in real projects. The reference framework is one of the parts of the theoretical foundation of the EMPOWER platform, which allows managing process from MDE point of view. Nowadays, EMPOWER is being developed by Servinform (a Spanish software company) and has a budget of 700.000€ for the 2015-2016 period. As future work, we are trying to enrich the reference framework and their pillars in order to support interorganizational collaboration processes and case management, including aspects not only of definition, but also of processes execution.

Paper Nr: 2

Towards an Automated Analysis of the Online Supply Chain of Novel Psychoactive Substances


Clara Bacciu, Fabio Del Vigna, Andrea Marchetti, Maurizio Tesconi and Paolo Deluca

Abstract: Novel Psychoactive Substances (NPSs), also known as legal highs or smart drugs, are legal alternatives to illegal drugs. Many drugs consumers are appealed by the opportunity of buying substances without any legal consequences. Online shops, virtual marketplaces and other trade channels thrive in this legal grey area. The health risks connected to this phenomenon are high: every year hundreds of people present symptoms deriving to the use or abuse of those unknown chemicals, and health professionals may struggle to provide the appropriate treatments. EU is taking some countermeasures, forbidding the sale of the NPS as soon as it is established their risk for the health, but natural or synthetic new substances are continuously discovered or manufactured, and it is difficult for legislators to keep up. To cope with the lack of regulation in this market sector, EU is funding the CASSANDRA project, to study and comprehend NPSs lifecycle and supply chain, through the automatic analysis of user generated content (forums and social media) and online markets. During the first year of activity, we combined data gathering, analysis, and visualization techniques to i) provide an insight over two large forums, Bluelight and Drugs-forum, that host discussions about drugs since more than a decade; ii) investigate how substances sold by online shops of the NPSs supply chain map inside the forums and iii) investigate how social networks like Facebook and Twitter are used to avertise and discuss drug consumption. In order to gather as much data as possible from forums, we developed an ad-hoc web scraper. The system keeps track of the forum hierarchy and structure, keeping all tags and other metadata associated to posts, threads and forums. All the content is anonymized and stored in a relational database with an associated text indicization. We got a snapshot of Drugs-forum and Bluelight, whose content spans respectively from 2003 and 1999 to today, with more than 1 million and more than 3 millions of posts, and about 200 thousands and 350 thousands of users respectively. A selected set of 10 online NPSs shops underwent a similar scraping and storage phase, while we crawled the social media pages connected to those shops through the provided APIs. We extracted a list of the advertised products in those shops and pages, finding more than 250. The forums have been the starting point and the core of the analyses so far. We developed some interfaces to investigate their structure, the number of posts per thread and the number of posts per user (which, as expected, follow a power low distribution). Moreover, we analysed the textual content of the posts, showing the number of occurrences of terms over time (Figure 1), in which sections a series of known NPSs are first mentioned, the terms co-occurring with other terms. In particular, this last analysis is leading to an automated system to show the most frequent symptoms mentioned together with the name of a substance. We also analyzed the hyperlinks appearing in the forums, and compared them with a comprehensive list of online NPSs shops and related social media accounts, finding that they don’t quite overlap, and which NPSs sold in those shops are mentioned in the forums, finding that almost every substance is mentioned. In the future we plan to extend the analysis to dark web marketplaces. Future work will also involve the development of an automatic system to detect the mention of unknown substances, in order to monitor the discussion about them from the start, to understand where substances are first mentioned and sold, and how the supply chain evolves.