Thinking about Agile-Approach to PhD - Research

Size: px

Start display at page:

Download "Thinking about Agile-Approach to PhD - Research"

Jeffry Webb
5 years ago
Views:

National Technical University Kharkiv Polytechnic

ua) Department of Software Engineering and Management

Karazin National University of Kharkiv (http://www.

ua/en) Department of Systems and Technologies Modeling

1 National Technical University Kharkiv Polytechnic Institute ( Department of Software Engineering and Management Information Technologies V.N. Karazin National University of Kharkiv ( Department of Systems and Technologies Modeling Thinking about Agile-Approach to PhD - Research Prof. Dr. Mykola Tkachuk tka.mobile@gmail.com, tka@kpi.kharkov.ua ICTERI-2017 / Kiyv / May 15-18, /22/2017 1

2 Promising "Agile-Domains" for PhD Research Agile - Research Domain : What does it mean? From my (maybe subjective) point of view for such a domain the following characteristic features are typical - really new and modern scientific problems and technological platforms - complex and weak-formalized requirements / tasks to be resolved. - a lot of changes and rapid development trends are in progress.. - an important role of human-centered factors in system design and operating -.. (to be continued). ICTERI-2017 / Kiyv / May 15-18, /22/2017 2

3 Advanced Requirements Engineering in Agile Software Development (on Scrum method example) Best practices - Ok! But research challenges are: How to build a Product Backlog in effective way (w.r.t. domain specific features)? What is a way to prioritize Requirements correctly for any Sprint Backlog? What about some methods and metrics to organize control loops : Daily meeting and Sprint meeting? How to provide Finding&Feedback in Scrum process w.r.t. such criteria as Software Quality attributes? to be continued 5/22/2017 ICTERI-2017 / Kiyv / May 15-18,

4 A possible approach: Cybernetic-centered scheme of Scrum-method Several process control loops can are organized Feedback-control using target quality metrics have to be used 5/22/2017 ICTERI-2017 / Kiyv / May 15-18,

Domain-specific Modeling (DSM) and Domain-Driven Development (DDD) in Software Engineering Main (and well-known) idea of DSM & DDD Traditional development: Any specific problem from

Disadvantages: lack of the reusable components; high project costs.

5 Domain-specific Modeling (DSM) and Domain-Driven Development (DDD) in Software Engineering Main (and well-known) idea of DSM & DDD Traditional development: Any specific problem from the given Problem Space is solved by creating different applications, each of them has to be developed in a unique manner. Disadvantages: lack of the reusable components; high project costs.. DSM&DDD: Problem Space is divided into a set of domains, for each domain a specific domain model has to be constructed, that later is transformed into appropriate software solutions. Advantages: Reuse of domain models (knowledge) when developing new applications; decrease of total costs in software development.. 5/22/2017 ICTERI-2017 / Kiyv / May 15-18,

6 Open research challenges in DSM & DDD Domain How to formalize initial domain knowledge in order to build an appropriate domain model (DM)? How to classify domain specific methods to DM realization, w.r.t. reuse different project assets: domain knowledge, requirement specifications, architectural solutions, source code, etc.? М What are suitable expert methods and quantitative metrics to assessment of different practice-oriented reusable assets, e.g. code reuse extend? How to estimate a domain model complexity / code reuse extend, e.g. how to estimate finally an efficient ratio of DSM & DDD usage?.. to be continued.. 5/22/2017 ICTERI-2017 / Kiyv / May 15-18,

7 Classification of DSM Methods The most suitable way to classify DA and DSM methods is consider them by type of phases / artifacts to be reused in a software development process: [*]K. Czarnecki and U. Eisenecker. Generative Programming: Methods, Techniques, and Applications. Addison-Wesley, Informatik-2016 / Alpen-Adria Universität Klagenfurt / Sept /22/2017 7

essential higher extent of code reusability (CR) than by using of JODA method and

8 Prototype testing: Evaluation of Code Reusability Extent Report in PDF-format Result: we can conclude that domain modeling by usage of ODM / EMF tools provides the essential higher extent of code reusability (CR) than by using of JODA method and Actifsource tools. 5/22/2017 Informatik-2016 / Alpen-Adria Universität Klagenfurt / Sept

http://eclipse.org/aspectj/ ; http://www.

9 Post Object-oriented Technologies for Software Development and Maintenance AOD ; FOD COD 5/22/

10 Advent of new Software Development Approaches Approaches to software structural modification (adaptation) Without changing of a paradigm (Object-Oriented Programming / OOP) Changing of software components architecture Changing of a Data structure Aspect-oriented development (AOD) Changing of development paradigm (Post - OOP) Context-oriented development (COD) Feature-oriented development (FOD) Research challenges: To build classification (taxonomy) of new Post- OOP (POOT) approaches w.r.t. their features and purpose of usage: e.g. to eliminate a cross-cutting functionality, to decrease a number of programs bugs, etc.? How to define an appropriate type of legacy software system to be adopted with a given POOT? How to estimate an effort needed to modify a legacy software system with a target POOT? What criteria and their metrics can be used in order to evaluate an effectiveness of POOT usage? to be continued 5/22/2017 ICTERI-2017 / Kiyv / May 15-18,

11 A possible multi - dimensional approach to comprehensive estimation of POOT effectiveness Effectiveness Estimation? Effect Criteria SystemType, POOT efforts POOT average efforts COP FOP AOP I II IX System types ICTERI-2017 / Kiyv / May 15-18, / 43

12 Models, Methods and Tools for Adaptive Mobile Systems Development Mobile information system Navigation system Task management system Augmented reality system Social network client 1) intensive GPS usage; 2) intensive usage of data transfer network. 1) low computational load on mobile device; 2) usage data transfer network depends on features system e.g. data synchronization, media content storage, etc. 1) high computational load on mobile device; 2) intensive GPS and build-in gyroscope usage; 3) built-in camera usage. 1) low computational load on mobile device; 2) intensive usage of data transfer network; 3) displaying of huge volume and different types of multimedia data 5/22/2017 ICTERI-2017 / Kiyv / May 15-18,

Some R&D problems on example of Augmented Reality (AR) Geolocation System AR- geolocation: uses GPS-module to detect coordinates; digital compass and gyroscope to detect mobile device position and

mobile AR systems Restricted CPU performance, RAM and battery capacity Small screen size on mobile devices No possibility to extend hardware performance of mobile device in run-time mode.

13 Some R&D problems on example of Augmented Reality (AR) Geolocation System AR- geolocation: uses GPS-module to detect coordinates; digital compass and gyroscope to detect mobile device position and camera to analyze physical environment to provide additional information about objects from the real environment which are shown on mobile device Typical Problems / Constraints in Development of mobile AR systems Restricted CPU performance, RAM and battery capacity Small screen size on mobile devices No possibility to extend hardware performance of mobile device in run-time mode. needs to support applications on different types of mobile devices Research & Development challenges: How to elaborate models and metrics to quality estimation of AR-systems? How to manage restricted AR-resources effectively? What types of distributed software architectures and technological solutions can be used to implement adaptive AR-systems? to be continued 5/22/2017 ICTERI-2017 / Kiyv / May 15-18,

14 Traditional workflow of PhD research: similar to cascading life-cycle model. Information gathering Literature overview and analyzing Theoretical Issues Models & Algorithms Prototype running Implementation Results testing PhD thesis completed, but what about their actuality and quality..? Main disadvantages: sequential workflow, hard-coded business logic, some initial requirements become no more actual, a lot od changes are not taken into account, solution quality issues can be estimated too late... 5/22/2017 ICTERI-2017 / Kiyv / May 15-18,

15 Agile-Workflow for PhD research: Scrum oriented vision. Main scientific points to be resolved ( Product Backlog ) Prioritized list of current tasks ( Sprint Backlog ) Sprint iterations ->: Small iterations ( daily scrum meeting ): share current results intern, discuss problems, analyze risks / bugs, etc. PhD- thesis completed! Scientific Advisor ( Product Owner ) Scrum team : PhD - student + master students) : -> Check and revise initial scientific points, get critical remarks, improve research concepts via reporting on public seminars, participating on relevant conferences (e.g. like ICTERI.) etc.. Your profit: flexible workflow, agile business logic, any requirements / tasks can be considered in time, all changes are performed iteratively, solution quality issues can be estimated / improved step by step. 5/22/

16 Our selected international publications in agile research domains M. Tkachuk, I. Martinkus, R. Gamzayev et al. An Integrated Approach to Evaluation of Domain Modeling Methods and Tools for Improvement of Code Reusability in Software Development // Heinrich C. Mayr, Martin Pinzger (Eds.): INFORMATIK 2016, Lecture Notes in Informatics (LNI), Vol. P-259: Kollen Druck+Verlag GmbH, Bonn, pp M. Tkachuk, A. Vekshin, and R. Gamzayev. A Model-Based Framework for Adaptive Resource Management in Mobile Augmented Reality Systems // Proceedings of the ICTERI-2016: 12th International Conference on ICT in Education, Research and Industrial Applications: Integration, Harmonization and Knowledge Transfer, Kyiv, Ukraine, June 21-24, 2016, CEUR-WS.org/Vol-1614, pp M. Tkachuk, K. Nagorniy and R. Gamzayev. Models, Methods and Tools for Effectiveness Estimation of Post Object- Oriented Technologies in Software Maintenance // V. Yakovyna et al. (Eds.): ICTERI 2015: Revised Selected Papers, Series title: Communications in Computer and Information Science, Vol. 594: Springer-Verlag Berlin Heidelberg, pp Tkachuk M., Martinkus I. Models and Tools for Multi-dimensional Approach to Requirements Behavior Analysis // H.C. Mayr et al. (Eds.): UNISCON 2012, LNBIP 137: Springer-Verlag Berlin Heidelberg, pp Vekshyn O., Tkachuk M. Algorithmic Software Adaptation Approach in Mobile Augmented Reality Systems // 7-th International Conference on Software Engineering Advances, ICSEA 2012, Lisbon, Portugal, November, 2012, ISBN: , p M. V. Tkachuk, R. O. Gamzayev, H. C. Mayr. Models and Tools for Effectiveness Increasing of Requirements Traceability in Agile Software Development // Problems in Programming. (Scientific Journal). Кyev.: Ukrainian National Academy of Science (special issue). pp Mykola Tkachuk, Sergiy Polkovnikov, Sergey Bronin. Adaptive Control Framework for Software Components: Casebased Reasoning Approach // Proc. of 6 th International Workshop on Software Cybernetics (IWSC 2009), Seattle, USA, July 20-24, /22/2017 Thank you for your attention! 16

CMPT 275 Software Engineering

CMPT 275 Software Engineering Software life cycle 1 Software Life Cycle Sequence of processes completed as a software project moves from inception to retirement At beginning of project development, choose