Document 7454699

Download Report

Transcript Document 7454699

4/28/2020

Coping with Complexity in Design

Peter Johnson Department of Computer Science University of Bath 1

Aims for the workshop

4/28/2020  To develop and strengthen Interdisciplinary research in design.

– Complexity of  Environment (physical, social and psychological),  Design Process (including teams, methods etc.),  Designs (including problem/solution spaces, models/representations etc),  Systems.

 Evaluation.

2

4/28/2020

Questions for the workshop

What are our: – – – Common Problems Common Approaches Different and Interesting Problems – Different and Successful Approaches. Identifying: – – Research Challenges New Research Opportunities.

– Potential Interdisciplinary Collaborations.

Common Different Problem Approach 3

4/28/2020

Overview of Workshop

 Sessions Thursday: 1.

Design Processes 2.

Design Models  3.

Design Requirements and Contexts Sessions Friday: 1.

2.

Safety Critical Problems Awareness Problems 4

4/28/2020

Points arising

 Chris Mc Mahon – Increasing design knowledge – Mismatches between process models and activities – – – Reducing uncertainty/error Reducing cost Novel designs V things we already know how to design. 5

4/28/2020

Points arising

   Chris Earl et al – – Reuse - similarity v identity Making explicit what can be carried across Colin Johnson – SE is essentially craft based – – Automatic - genetic programming Shift to programming what rather than how Stuart Lockhart et al.

– crew demand met by systems HCI – – Interactions between crew as well Requires novel hci - increased modularity 6

4/28/2020

Points arising

  Dave Roberts – – Problem to Solution with UML through Multiple goals and roles - to Use Cases to designs – – – – TA Tools too constraining add safety aspects.

Public Objects & Private Objects Used to produce safety case Emergent Properties, knowing, responsibility 7

Points arising

4/28/2020    Colin Braine – System is an abstraction, conception and not real.

– – Complexity is invariably socio-technical.

Validation involves interactions that are arise in operation rather than by design.

– Better communication - clear shared concepts.

Michael Albers – Context, situation,user,goals,information needs,system. About user goals Not TA?

Carl Johan Orre & Ulf Hedestig – – Design philosohy and process for KM and e-learning Convergence of private and public dimensions – – Communities of practice Activity theory  re-contextualisation,developmental,social,semi-formal  Grounded Understanding 8

Points arising

4/28/2020    Susan Omadafe & Stephen Egarievwe – – HCI of - security, confidentiality, integrity, availability etc Not an emergent property. Identifies properties, approaches ,etc.

David England et al.

– – – Design process for breast cancer healthcare system.

Decision support system - NICE method National guidelines, governance etc; evolution, IT infrastructure, context of use issues.

Daniel Normark – Design approach for - co-presence in petrol stations – Models not possible, Context Associative design eg scenarios and prototypes).

– – Abstracting from examples observed in reality. Design trade-offs 9

4/28/2020

Points arising

 Chris Johnson – HCI issues are features of more deeply embedded issues   Wider engineering & “band-aid software” Management “systemic myths” – – – Re-use - of teams (error propagation - ties up resource).

HCI of/for Management Software and SE processes made more complex by wider engineering problems – – Lack of model of platform Poor project coordination no CSCW – HF is not HCI many HF personnel no HCI personnel.

10

4/28/2020

Points arising

 Anne Bruseberg –

New types of problems and errors (stemming from a new set of influences?)

Fundamental processes of collaboration

Computer system roles

Translator

– – – 

Co-worker

Delegation over time, with delegated process’ state changes (wait, active, mode) + changes in state of components they control (on, off, empty etc.) Remoteness ? (degree of control/intedependence; communication?) Types of change

  

Rapid -

instantaneous transformation of state

Dynamic

- continuous variation in a relevantly stable state Phased – gradual transformation

11

Points arising

4/28/2020        Penny Ross

Time – 40,000 patients pa killed by NHS blunders Staff see 180 people per day, physical records, chasing results, Multiple competing urgent priorities

changing on the fly

need to be able to down tools without damage

Noisy distracting environment - a rich situation ?

Large number of states affected by a “

compound of interrelated variables

” The key clinical information

12

4/28/2020

Points arising

                

Fabrice Drogoul

Eurocontrol “Integrated Risk Picture” Safety in air traffic management Can’t reduce complexity, only deal with it.

ATM is a contributing factor Accident reports

Measurement criteria for safety Design criteria for safe systems

Target level of safety – achieved level of safety Open system, dynamically evolving, … “Confluence of small variations in normal barriers” Eurocontrol is an independent data bank for accident reports

Anonymized wrt source authority

Cross-boundary hazard - Synchronic, Diachronic Transition (pace), migration

Representing uncertainty in display

Degree of confidence in the (assumptions of the) model Degree of precision in the model (given the assumptions) Wickens, Illinois, probabilistic reasoning false positives, and false rejections??? Billings, MIT, displays.

13

4/28/2020

Points arising

  Andy Howes et al.

– Reducing complexity by data fusion   Data v Information Using Simon’s notions of constraints on cognitive architecture.

 Strategy changes as constraints increase speaks against distributed cognition claim that people off-load cognition to the environment.

– Constraint modelling and resource streams.

Rachid Hourizi – Awareness of situation & environment – Selection of information - perceptual strength and task relevance. – – Notice v Action Both influenced by redesigns independently.

14

4/28/2020

Points arising

 Sine McDougall & Martin Curry – – – – Icon interpretation Semantic “pop-out” from perceptual content Concreteness, semantic distance, familiarity Increased workload- no effect for Accuracy but RT increased semantic distance better for novices, familiarity better for skilled people - add - verbal memory, conceptual links.

– – Chunking.

Cultural.

15

4/28/2020

Questions for the workshop

What are our: – Common Problems – Common Approaches – Different and Interesting Problems – Different and Successful Approaches. Identifying: – Research Challenges – New Research Opportunities.

– Potential Interdisciplinary Collaborations.

Problem Approach Common Different Design Emergence Reuse Cost Analyses Applications Applicability Granularity of problem Models Processes Formal, Semi-formal, Informal Granularity of explanation Engagement with design.

Forms of enquiry Forms of data 16

Research Issues

4/28/2020              Novel designs V things we already know how to design Reasoning about, transforming and restructuring complexity Characterise, understand, and cope with Complexity Abstraction, Sharing, Contexts Shifting Complexity eg from mech/elec engineering to software engineering “band-aid software” Management/responsibility issues increase complexity Collaboration issues increase complexity Integrating across levels of descriptions/explanation Integrating into design/analysis.

Including HCI (already HF) Involving - Psychological, Sociological, Engineering, Management perspectives.

17

Research Agenda 2

4/28/2020  Development of Research Networks Common Language, Prioritise Research Issues, Stimulate Collaboration  Large Integrated Research Projects  Smaller “Blue sky projects feeding in  Focus on education as well - integrated U/G, P/Gt and Research Degrees 18

Future Activities

4/28/2020  Maintain Interdisciplinary Perspective – Future workshops attendance special purpose Not necessarily associated with recognised conferences Professional Bodies IEE, Design Council, BCS, IMECH Report of this workshop distributed widely.

Book aimed at researchers and/or practitioners.

19