Seminar by Dr. Dalila Tamzalit & Dr. Guillaume Fertin (Université de Nantes)

Speaker: Dr. Dalila Tamzalit (Université de Nantes)                                                                                                  

Title: Towards Sustainable Software: an architecting approach

&

Speaker: Dr. Guillaume Fertin (Université de Nantes)                                                                                                

Title: Theoretical Computer Science and Applications in
        Bioinformatics : Walk on your Two Legs

Date: Wednesday, October 9th, 2019

Time: 3:00pm-4:30pm

Room:  EV2.260

*Note: this is a joint seminar with two speaker presentations

ABSTRACTS

Towards Sustainable Software: an architecting approach (Dr. Dalila Tamzalit)

     Designing, developing and maintaining qualified software
     systems, with a satisfactory longevity and reasonable costs has
     always been a major challenge. This issue is even more crucial and
     riskier in today's fast-moving, complex and multi-user environments.
     The evolutions and intensive use of digital technologies, in
     particular the Internet, and the democratization of connected
     products (laptops, tablets, smartphones, watches, IoT...) have
     massively complicated information, which is increasingly distributed
     and decompartmentalized, and its exchange. In this context, ensuring
     sustainability for a complex software system is a high-risk activity
     and requires considerable efforts. The financial, personnel and time
     costs increase with the age and obsolescence of the systems. Their
     evolution most often causes an undesirable side effect: corruption
     and deterioration in the quality of these systems, making their
     sustainability even more difficult. In this presentation, I’ll give
     my proposition for enabling software systems to be easily and
     intrinsically responsive to the need for changes. My conviction is
     the importance of (1) architecting software systems to make them
     easily adaptable and (2) architecting software evolution as a
     separate concern from other architectural concerns. How? I am deeply
     convinced that the fundamental solution lies in the principle of the
     separation of concerns. To do this, I use an architectural and
     meta-modelling approaches, applied for more than 15 years, to
     different architectural styles (object-oriented, component-based
     software architectures, SOA, SaaS, micro-services). I illustrate my
     point of vue on the migration of legacy systems towards SaaS
     multitenant and micro-services architectural styles. In particular,
     I am interested in various scientific locks that I mention in order
     to encourage exchanges and potential collaborations.

Theoretical Computer Science and Applications in
Bioinformatics : Walk on your Two Legs (Dr. Guillaume Fertin)

     I will present a small sample of the activities of the
     ComBi (Combinatorics and Bioinformatics) group at LS2N, University
     of Nantes. Doing that, my goal is to show that methodology is highly
     important in modern bioinformatics, and therefore should not be
     underrated – or, even worse, ignored.  In particular, using clever
     (possibly ad hoc) data structures and designing fine-tuned
     algorithms (e.g. on sequences, on graphs) is essential in this
     domain. Doing so, one can provide fast, scalable and reliable
     software solutions. I will illustrate the above by different
     examples, namely problems arising in mass spectrometry, genome
     rearrangements and metagenomics.