requirements management - crystal -home€¦ · requirements management a knowledge-centric...
TRANSCRIPT
Requirements Management
A knowledge-centric approach
Dr. Jose María Alvarez-Rodríguez & Prof. Juan Llorens
Knowledge Reuse Group
Carlos III University of Madrid, Spain
CRYSTAL Final Dissemination Event
Madrid, 15-16 June, 2016
1
Sometimes...
Engineering fails...2
Final Product
3
and causing...
loss of human life
A major problem andconcern!
4
Requirementsto the rescue...
"A requirement is a need or a property that whatever human invention must be able to perfo
rm or fulfill."
5
How can you identify needsfrom humans?
Requirements Engineering
6
The context...It does not matter conditi
ons…
SAFETY is aMUST!
7
The need of a holistic viewof a system...
8
The process...
9
To reach...
10
What is a requirement?
RequirementsAttributes...
11
Typesof
requirements
12
Representation ofrequirements
13
How can we manage requirements?
14
Requirement Management System
(RMS)
15
and the quality?
16
Interoperability
Quality metrics
Representation
Challenges
17
InteroperabilityOpen Services for Life-
cycle Collaboration
Source: http://open-services.net/
18
The CRYSTAL use cases
RequirementsQuality Checking
KnowledgeManagement
19
Contributions
OSLC KPIs -> OSLC EMS*
OSLC KMKnowledge Management
20
K
RAT RQA
M
RequirementsAuthoring Tool
RequirementsQuality Analyzer
Knowledge MANAGER
Tools by TRC
21
OSLC OverviewSituating contributions...
22
OSLC Overview
Situated contributions...
23
KPIsIndicators forrequirements quality+56Compile: dimensions,units of measurement,etc.
ObservationsObservations linked to aKPIA value under a context
Time...
OSLC KPIs
Application to requirements quality management
24
OSLC Defined ResourcesBased on qb:Dataset and qb:Observation
W3C Recommendation RDF Data Cube & Ios_kpi:KPI Ios_kpi:Observation
25
Scenarios...
26
1-Checking the quality of atext-based requirement
27
2-Checking the quality of an OSLC-based requirement
28
3-Checking the quality of an OSLC-based requirement
and a Jazz Widget
29
30
SummaryOSLC KPIs-Key Performance Indicators
Use of the W3C RDF Data Cube VocabularyKPIs and Observations are the entities to be exchanged Exposing as OSLC-EMS provider.
KPIs + ObservationArtifact
31
Context for OSLC KMMultiple domains
Different types of artifactsNeed of intra-operability
Intra-domainNeed of interoperability
Inter-domain
32
Real SituationPoint to point connections-> Combinatorial explosi
on
33
OSLC Preliminary Evaluation
Great effort on interoperabilityCommunity ( industry ) effortDifferent types of artifacts ( resource shapes )OASIS standards (Core, Change Management, etc.)...Challenges (among others)
Increase of interoperability complexity when newdomains are definedProvision of services to all the resource shapes
which leads to the need of amore universal approach
34
KCSE
Knowledge -centricSystems Engineering
35
KCSE: notion of an OSLC bus
36
Representation
Services
COMMON
37
How many different types of artifactsare generated during the development
life-cycle ?
tools, formats, protocols, query languages , etc.
Representation
38
Needs and Challenges
A common representation modelInteroperability...
39
ServicesLanguage Uniformity Quality checkingVisualizationTraceabilityHuman machine interface...
40
Needs and Challenges
A common representation modelInteroperabilityKnowledge Management processesNatural language to express queries...
41
OSLC-KM
New domain
New resource shape
42
RDFResource DescriptionFramework
Common & shared DatamodelTriples
(subject, object, predicate)
Binary relationshipsUnderlying Directed GraphW3C Recommendation(2004)Query languages
RDFSRDF Schema
Data modeling for RDFdataClassesProperties (domain &ranges)RDF serializationW3C Recommendation (2004)
43
OWLOntology Web Language
Vocabulary for definingformal ontologiesLogic-orientedRDF serializationFlavours (2.0): EL, QL, RLW3C Recommendation (2012, v2.0)
RIFRule Interchange Format
Exchange of businessrulesRule-orientedFlavours: Core, PRD, BLD,etc.XML serializationW3C Recommendation (2013, v2.0)
44
RSHPRelationship "arship"
Property GraphAny kind of relationship
Arity and CardinalityIndustry-orientedNative tool supportQueries based on naturallanguageFirst publication 2004
Others
SBVRSemantics of BusinessVocabulary and Rules
ODMOntology DefinitionMetamodel
RASReusable AssetSpecification
45
Preliminary Evaluation
http://trc-research.github.io/spec/km/
RDF good option for exchange data on the web
Restrictions: arity and cardinality of relationships, lack ofnative tools, logics, etc.
RDFS and OWL
Oriented to define formal ontologies
RSHP
High level of ExpressivityNative tool support...
46
The approach
OSLC (Data Exchange)
RSHP*(Internal representation: metadata+contents
and services)
47
The Specification
Shape for KM resources
Metadata + Contents
48
Resource Shape
Based on the W3C SKOSRecommendation
Simple Knowledge Organization System
49
OSLC CoreProvenance* Access (W3C HTTP Access)
Metadata (Dublin Core, traces, etc.) Contents (resource shape)
Visualization (SVG)
Summary of Properties
50
OSLC Defined ResourcesBased on skos:Concept
W3C Recommendation SKOS (Simple KnowledgeOrganization System)
Ios_Km:Concept
51
Vocabulary
& Conceptual Model
NormalizationStandardizationSuggestionsX Breakdown Structures...
Patterns
RestrictionsExamplesSuggestions
Knowledge ManagementApplication to requirements authoring and requirements quality man
agement
52
Scenarios...
53
1- Listing the concepts of adomain vocabulary
54
2- Standardize a term ortext-based description
55
3-Get synonyms for a termor a text-based description
56
4-Get the list of patterns toguide the writing of
requirements
57
Real Example
SKOS Concepts through OSLC in CKEditor
58
SummaryOSLC KM-Knowledge management
Use of the W3C SKOS vocabularyConcepts and relationships are theentities to be exchanged Services for: search, trace, naming,visualize, etc.
Metadata + ContentsArtifact
oslc_asset:artifact
59
An Example...
Process any kind of OSLC Resource or RDF data ...E.g. Modelica
60
One implementation
61
OSLC KMKnowledge Centric Systems Engineering to gov
ern the development lifecyle
KCSE
Inputs Outputs
...
62
Common serv
ices
Representa
tion
OSLC KMNot a dream!
KCSE
Inputs Outputs
...
Common serv
ices
63
Represen
tation
AdvantagesStandard exchange ofdata: OSLCEnhance expressiveness( RSHP )Reuse of existingstandards andvocabulariesNative Tool support(Knowledge Manager)Cross-cutting services(semantic-based)Elastic approach...
Drawbacks
Scope of knowledgemanagement (needs )Potential overlappingwith other tools andspecs Need of spread theapproach
64
Evaluation of OSLC KMKnowledge Management: a key process
REUSE!Need of:
Reuse of standards and service-orientedfunctionalities (OSLC)Take the most for data exchange (OSLC) andrepresentation (RSHP)Bring technology to a human-orientedenvironment (NLP)
Technical issuesCompleteness of the specification?Implementation
65
Conclusions
Two main action linesRequirementsQuality MetricsKnowledgeManagement
Keypoints:Re-use of existingRDF vocabularies
Good experience withOSLC4Net
Change in the JSON SerializerJSON format->betterperformance for largeresponses
OSLC Domains Technical issues
66
Future workMerge and extend the spec with other knowledgestandards
E.g. STEP, Industrial Internet activitiesRefine of the resource shape Full implementation and support to all servicesIntegration patterns governed by knowledge...
Industrial Linked Data
67
68
69
Creditshttps://www.edx.org/course/software-architect-code-building-digital-uc3mx-inf-1xhttps://slides.com/josemariaalvarez/oslc-km-incose-2015-15
70
Speakers
Dr. Jose María Alvarez-Rodríguez
Carlos III University of Madrid, Spain
Member of INCOSE and the OSLC RM
working group
E-mail: [email protected]
WWW:
http://purl.org/krgroup/web
P ersonal site
Prof. Dr. Juan Llorens
Carlos III University of Madrid, Spain
Member of INCOSE
CTO of The Reuse Company Inc.
E-mail: [email protected]
WWW:
http://purl.org/krgroup/web
Personal site
71
Data modeling
Need of logic formalism ?Requirements qualitymetrics match OSLC EMSMultilingual data in OSLCAPIsOn-line vs Reference vsProxy resources...
Infrastructure & Dataintegration
Convergence OSLC andRDF Data Shapes , W3CLDP, etc.Joint efforts in OSLCspecs. and new domainsOSLC APIs in otherlanguages
Microservices...
Federated queries and stream processing
Trust and Privacy
72