4. Medical technologies and justifications
Various disciplines
Clinical Electronics Software Mechanics
Standards
Software Lifecycle IEC 62304 Risk Management ISO 14791 Usability IEC 62366
Internal Practices
Software Development Plan Risk Matrix Analysis Usability criteria
Justifications
Sprint Report Risk mitigation Report
Agility
16/11/18 - duffau@i3s.unice.fr PhD Defense 4
⇒ disciplinary practices but also transversal
⇒ interlaced
⇒ informal practical know how
⇒ differences between the expected and the produced
⇒ from ‘last day’ concern to justification tsunami
5. Some vocabulary
Justification requirement : a proof element that must be reached to be compliant
with a law. It can be set through a standard or a guide
Elicitation : action to help experts to formalize their knowledge to keep and share
them
Justification : an argumentation attesting that a justification requirement is fulfilled
Justification artefact : a data (e.g. document, result) to establish in order to
construct a justification
16/11/18 - duffau@i3s.unice.fr PhD Defense 5
6. Objectives
Standards
Internal practices
Regulatory
authority
Company’s
experts
3. produce
1. produces
4. reviews
2. contextualize
5. develop
Product
6. V&V
Deliverable product
7. audits
certified
16/11/18 - duffau@i3s.unice.fr PhD Defense 6
justification requirement
justification
conformance
conformance
O1 Elicitate
Help to produce
O2
O3
Check and analyze
Check and analyze
7. Outline
State of the Art
Contributions
Validation
Conclusion and Perspectives
16/11/18 - duffau@i3s.unice.fr PhD Defense 7
8. Outline
State of the Art
Contributions
Validation
Conclusion and Perspectives
16/11/18 - duffau@i3s.unice.fr PhD Defense 8
9. Derived from Toulmin argumentation schema ²
Comprehensive notation to explain why a result is trustable
Captures the rationale logical structure of all evidence that leads to the
acceptance of a high-level property
+ Proved to be useful ³
- Only a graphical notation
⇒ A good starting point but need to be formalized
1. Polacsek, T.: 2016, Validation, accreditation or certification: a new kind of diagram to provide
Confidence
2. Toulmin, S.E.: 2004, The Uses of Argument
3. Polacsek, T., S. Sharma, C. Cuiller et V. Tuloup. 2018, The need of diagrams based on toulmin schema application :
an aeronautical case study, EURO Journal on Decision Processes
Justification Diagrams¹
16/11/18 - duffau@i3s.unice.fr PhD Defense 9
10. Historically Goal Structuring Notation¹ (GSN) and Claim-Argument-Evidence² (CAE)
Used for safety case
New OMG standard : Structured Assurance Case Metamodel ³
SACM focus on 3 aspects :
● Argumentation
● Artifact
● Assurance Case
+ A formal model
- A minimal conformance relationship
⇒ A good semantics starting point
1. Kelly, T., Weaver, R.:2004, The goal structuring notation /- a safety argument notation
2. Emmet, L., Cleland, G.: 2002, Graphical notations, narratives and persuasion: a pliant systems approach to hypertext tool
design
3. OMG: 2018, Structured assurance case meta-model (sacm)
Assurance Case
16/11/18 - duffau@i3s.unice.fr PhD Defense 10
11. Justification production
● Traceability from requirements to production
● Specifics methodologies and tools for each stage
○ Requirement framework (ECHO ¹, SEMIOS ², ...)
○ Architecture framework (AREL ³, ...)
○ Development framework (Issue-based system, Continuous Integration ⁴, …)
+ Give ways to collect justifications
- No conformance with justification requirements
1. Lee , C., L. Guadagno et X. Jia . 2003, «An agile approach to capturing requirements and traceability», 2nd International
Workshop on Traceability in Emerging Forms of Software Engineering
2 Mezghani , M. , J. Kang , F. Sedes ; 2018, « Industrial requirements classification for redundancy and inconsistency detection in
SEMIOS
3. Tang, A., Y. Jin et J. Han . 2007, «A rationale-based architecture model for design traceability and reasoning», Journal of
Systems and Software
4. Fowler , M. et M. Foemmel . 2006, «Continuous integration»
16/11/18 - duffau@i3s.unice.fr PhD Defense 11
12. State of the Art matrix
16/11/18 - duffau@i3s.unice.fr PhD Defense 12
13. Outline
State of the Art
Contributions
Validation
Conclusion and Perspectives
16/11/18 - duffau@i3s.unice.fr PhD Defense 13
14. A multi-level conformance
16/11/18 - duffau@i3s.unice.fr PhD Defense 14
Justification Pattern
Diagram for
standards
Justification Pattern
Diagram for internal
practices
Justification Diagram
for a product
conformance conformance
15. Conformance at the baseline
16/11/18 - duffau@i3s.unice.fr PhD Defense 15
DUFFAU , C., T. POLACSEK M. BLAY -FORNARINO . 2018, Une sémantique pour les patrons de justification,
INFormatique des ORganisations et Systèmes d’Information et de Décision 2018 (INFORSID)
O3 Check and analyze conformance
16. A flexible conformance for justification step
O1
16/11/18 - duffau@i3s.unice.fr PhD Defense 16
O3
O1
DUFFAU , C., T. POLACSEK M. BLAY -FORNARINO . 2018, Une sémantique pour les patrons de justification,
INFormatique des ORganisations et Systèmes d’Information et de Décision 2018 (INFORSID)
Check and analyze conformance
Elicitate justification requirements
17. Evolution and conformance
16/11/18 - duffau@i3s.unice.fr PhD Defense 17
Justification Diagram
for a product
conformance conformance
feedbacks
from the field
Justification Pattern
Diagram for
standards
Justification Pattern
Diagram for internal
practices
18. Evolution characterizationO1
16/11/18 - duffau@i3s.unice.fr PhD Defense 18
e’34 R e3
e’34 R e4
O3
Elicitate justification requirements
Check and analyze conformance
19. Comparison and alignment operation
O1
16/11/18 - duffau@i3s.unice.fr PhD Defense 19
O3
Elicitate justification requirements
Check and analyze conformance
e’34 R e3
e’34 R e4
20. Bridge the gap with production
16/11/18 - duffau@i3s.unice.fr PhD Defense 20
Justification Diagram
for a product
conformance conformance
artifactsA lot of
Justification Pattern
Diagram for
standards
Justification Pattern
Diagram for internal
practices
22. Outline
State of the Art
Contributions
Validation
Conclusion and Perspectives
16/11/18 - duffau@i3s.unice.fr PhD Defense 22
23. Challenge
● New products
● Changing practices to Agility
● A small team
=> Capitalization on our practices while cost effective
quality
16/11/18 - duffau@i3s.unice.fr PhD Defense 23
24. study conduct for JPD
Stakeholders:
1 researcher/practitioner, 2 quality system managers, 3 technical leaders
Study guideline for a stage:
1. The researcher designs a JPD according to quality management team and technical
leader requirements
2. The technical leaders
a. Identify justifications items that must be produced
b. Develop tools to produce them
3. During the development step, the technical leaders
a. Possibly define new activities
b. Produce the necessary justifications
4. During the deployment step, the quality managers and the researcher analyze the
differences between the original JPD and the JDs
DUFFAU , C., T. POLACSEK M. BLAY -FORNARINO . 2018, Support of justification elicitation: Two industrial reports,
Advanced Information Systems Engineering - 30th International Conference, CAiSE 2018
JPD
JD
JPD/JD revision
16/11/18 - duffau@i3s.unice.fr PhD Defense 24
25. JPD for prototyping - IEC 62304 Software Lifecycle
16/11/18 - duffau@i3s.unice.fr PhD Defense 25
JPD design to elicitate
medical standards
26. JPD for system validation - Internal practices
16/11/18 - duffau@i3s.unice.fr PhD Defense 26
Interview on the
usefulness ⇒
Acceptable
Understanding between
experts ⇒ Acceptable
Feedbacks from auditor
⇒ Acceptable
O1
Elicitate justification
requirements
27. Justification Factory with
DUFFAU , C., B. GRABIEC M. BLAY-FORNARINO . 2017, Towards embedded system agile development challenging
verification, validation and accreditation: Application in a healthcare company, ISSRE 2017-IEEE International
Symposium on Software Reliability Engineering
16/11/18 - duffau@i3s.unice.fr PhD Defense 27
AXONIC
environment
Automate justification
construction based on
existing environment
Continuous Integration ⇒
79/82 automated
construction
Redmine ⇒ 184/184
automated validation
O2
Help to produce
justifications
28. ● Control lifecycle and automate output document
● Justifications Verification ( ~ 1 time /week /employee)
○ Error rate reduced from 33.9% to 16.3%
○ Timing reduced from 60 minutes to 1 minute
● Justifications Validation (~ 1 time / week)
○ Error rate reduced from 13.8% to 0%
○ Timing reduced from 30 minutes to 2 minutes
16/11/18 - duffau@i3s.unice.fr PhD Defense 28
Metrics, feedbacks on
O3
Check and analyze
conformance
29. Outline
State of the Art
Contributions
Validation
Conclusion and Perspectives
16/11/18 - duffau@i3s.unice.fr PhD Defense 29
31. Perspectives
● Co-evolution of justifications (hidden dependencies, hierarchy impacts, ...)
● Study the relation between business process and justification process (BPMN)
● Capture all the semantics in our meta-model (OCL)
● Extending Requirement Engineering with justification approaches (KAOS, i*)
● Validation on different contexts
16/11/18 - duffau@i3s.unice.fr PhD Defense 31
32. Justification and Experimental Systems
1. DUFFAU , C., C. CAMILLIERI M. BLAY-FORNARINO. 2017, Improving confidence in experimental systems through
automated construction of argumentation diagrams, ICEIS 2017
2. DUFFAU , C., C. CAMILLIERI M. BLAY-FORNARINO . 2017, Vers l’argumentation automatique d’expérimentations :
application à un portfolio de workflows
16/11/18 - duffau@i3s.unice.fr PhD Defense 32