Notes from ISEC-2018
Innovations in Software Engineering Conference 2018
Day 1
Workshop: Emerging Software Engineering Education (WESEE)
Few key-points from first session from Ashish Sureka:
- Case Based Learning should be used more pro-actively in schools and colleges.
- Activities like discussions should be held in a round-table sitting than a lecture-oriented sitting. It seems intuitive yet so less frequent.
- SEABED: Database for case studies for software engineering
Dr. Bimlesh Wadhwa’s session talked about her best practices in software engineering courses. Some key notes are:
- Software Engineering Courses can be classified as UG/PG, reference curriculum and specialized courses.
- Software engineering is pervasive.
- Different tools used by her students are: JMS/JPA/Redis, D3, SNAP/TRAVIS, Telegram, GIT/GitHub, AWS-Lambda, GitHub Classroom etc.
- A typical development cycle includes:
Architecture -> Design principles -> Design patterns -> Design Quality -> Software Development - Students work with real databases like Semantic Scholar Database(which is of the size of TBs). This forces students to tackle big data issues like storage and do visualizations for e.g. using deep trees etc.
- Students also participate in SCORE competition.
- She mentioned about a distributed SE database at fer.unizeg.hr and D-Graphs.
- 3 pillars of designing a SE course are —
Concepts/curriculum
Delivery
Assessment
Next session in the workshop was by Paramvir Singh from NIT Jalandhar titled, ‘What’s Trending with University Courses on SE?’
- The key question asked was if CS graduates are SE employable?
- The classification lines adopted were creative. Short term and specific courses were excluded from the study.
- Key observation was that in the majority of courses focused on the design stage of SDLC.
Few remarks from audience after the session,
- Github lists universities and ranks them through various parameters which can be used in such surveys.
- MOOCs can also be included along with formal education to get a more holistic picture of affairs.
Next up was a session by Dr. Jane Huang-Cleland on Safety Critical Thinking.
- Drones were used to make the topic interesting for students.
- Project ideas related to application of UAVs/drones like
medical fleet
e-health
environment
mission control
aerial reconnaissance; were proposed - Goal for the course was to infuse safety critical thinking and not to really come up with market-ready products.
Tech Briefing: DLide - Deep Learning IDE
Shreya Khare, developer at IBM Research Lab India demonstrated the IBM Visual IDE for Deep Learning.
- Supports convolution, pooling, softmax and various other operations used in deep neural networks.
- Helps use best practices and design patterns
- Makes reproducing research results in DL easier
- Interoperability: use different libraries in one project under DL-IDE
Tech Briefing: MC/DC Testing by Prof. D P Mohapatra from NIT Rourkela
MCC or Multiple Condition Coverage has exponential increase in test cases and the order of execution of test set affects the time taken for testing significantly.
MC/DC ie Modified Condition/Decision Coverage on the other hand as only a complexity of O(n) for number of test cases and hence better.
Safety critical systems require MC/DC testing as per certification DO-178 B/C
Day 2
Keynote by Tao Xie on Intelligent Software Engineering
The talk was about synergy between artificial intelligence (AI) and software engineering (SE).
He mentioned few products like StackMine, XIAO, AnnaTalk, diffBlue, codata.
One key point to note was about metamorphic testing for ML based systems.
Analysis of requirements for Safety by Allenoush
Text processing was done in Java followed by a neural network written in python for making conclusions.
Validating Requirements Reviews by Maninder Singh
The faults were classified as ambiguous, inconsistent, omission, incorrect fact, misc and extraneous.
[will be updated soon]
