Rob Bracewell
A Tool for Capturing
Design Rationale
Project Team
Rob Bracewell Ken Wallace
Funding
EPSRC UTP Core BAE SYSTEMS
A Software Tool Called DRed
• What is DRed?
• What do designers think of it?
• Where did it come from?
• How was it researched?
– CAEDR Methodology
Rob Bracewell
What is DRed?
• Design Rationale (DR) editor
• An aid to designers that also captures DR for future use
• Used by aerospace designers from v0.1 onward
– just 3 weeks’ software development in v0.1
• Novel graph based derivative of the venerable IBIS
– DR captured in series of inter-linked planar graphs – like gIBIS, Questmap, Compendium, n-dim
Rob Bracewell
What do designers think of it?
• Questionnaire of trial users in Rolls-Royce:
• Found to be easy to use
– 14 out of 19 responses found easy or very easy
• Some immediate improvements to design process:
– 10 out of 10 found improved or slightly improved overall view of design process
– 7 out of 10 found improved or slightly improved keeping track of design process
– 6 out of 10 found improved or slightly improved evaluating and deciding between concepts
– No negative feedback, 0 out of 10 said hindered/slightly hindered design process
• Designers perceive future benefits. Main benefits:
– communicating to others 11 out of 14 designers – monitoring progress 9 out of 14 designers
A surprisingly positive response!
• DRed has some novel features, but is not unrecognisably
different from a number of earlier IBIS tools
• Consensus view, based on considerable research, has
been that this type of tool is not suitable for designers to
record their work
Rob Bracewell
How was the research conducted?
•
CAEDR
Methodology
• Fresh approach to researching software tools for
designers
• Proposed by Bracewell and Shea at ICED’01
• Emphasis on enabling early use and evaluation
of research prototype software in industry
CRITERIA
DESCRIPTION I
PRESCRIPTION
DESCRIPTION II
Measure
Influences
Methods
Applications Observation &
Analysis
Assumption & Experience
Observation & Analysis
Basic method Results Focus
Rob Bracewell Prescription Following Step Dependency Rapid Change Decide Visualisation, Interaction & Distribution Requirements Description II Criteria Description I
Define and Justify Measurable Criteria
Choose Knowledge Level Representations
Storyboard Tool in Context of New Design Process Specify Experiment and Data Collection Software Analyse Data Model Existing Design/ Process
Rapid Software Prototyping
S o c ia l S c ie n ce M e th o d o lo g y S u p p o r t E n g in ee ri n g S o ft w a r e D e v e lo p m e n t S u p p o r t Systems Analysis User Interface Design Computer Science/A I Data Analysis Method Selection Interview Techniques Observation Techniques Data Mining Experimental Techniques Interpretation CSCW Choose Knowledge Modelling Tools Choose Methods Knowledge Modelling
Prescription Criteria
Description I
2 Define and Justify Measurable Criteria (capture more rationale, improve its presentation)
3 Choose Knowledge Level Representations (IBIS-like graphs)
6 Storyboard Tool in Context of New Design Process (Graphs replace text for DDR Requirements, Alternatives & provide scheme folder map)
9 Decide Visualisation, Interaction & Distribution Requirements (Standard UI conventions, no info
lost in b/w print; paste into Word, cope with large projects,
run direct from CD)
13 Specify Experiment
and Data Collection Software
5 Model Existing Design/ Process (Structure design definition reports (DDRs) graphically)
Rapid Software Prototyping
4Choose Knowledge Modelling Tools
(Graphlet)
8 Choose Methods (Graphscript Tcl bindings
of GTL procedures)
11 Rapid Software Prototyping (8 releases in 8 months)
12 Formative evaluation (core team 6 designers in RR, 1 in BAE)
7 Choose Implementation
Level Representations
(GTL graphs, gml files)
1 Overall Success Criteria (Identify/solvecritical problems
faced by designers in KCSR)
10 Choose Development Tools
and Components
(Graphlet, TclPro, Ghostscript/GSView, BLT, Img, Schwartz)
Rob Bracewell
CAEDR
: Issues Crucial to DRed
Description I Stage
• Choice of knowledge level representation
(IBIS-like graphs), knowledge modelling tool
(Graphlet), and iterative use modelling rationale
from real designs (RR design reports)
– Found effective way of structuring/presenting rationale – Effective demonstration of potential of research to
Rob Bracewell
CAEDR
: Issues Crucial to DRed
Description I Stage
• In choosing Graphlet, looking forward to the
Prescription Stage, for
– ability to provide suitable implementation level representations and methods
Description I
3 Choose Knowledge Level Representations
(IBIS-like graphs)
6 Storyboard Tool in Context of New Design Process (Graphs replace text for DDR Requirements, Alternatives & provide scheme folder map)
5 Model Existing Design/ Process
(Structure design definition reports (DDRs) graphically)
Rapid Software Prototyping
4 Choose Knowledge Modelling Tools
(Graphlet)
8 Choose Methods (Graphscript
Tcl bindings of GTL procedures)
Rob Bracewell
CAEDR
: Issues Crucial to DRed
Prescription Stage
• Develop tool by incremental modification of an existing
general purpose tool (Graphlet)
– deliver usable software to industrial users, albeit of limited functionality, very early (3 weeks)
• Storyboarding the use of the tool
– must export to MS Office for inclusion as design report figures – designers not worried time spent using tool will be wasted
• Graphlet employs the “2 language” compiled+scripting
approach advocated by
CAEDR
CAEDR
: Issues Crucial to DRed
Prescription Stage
• Software development by the researcher using a very
high level interpreted scripting language (Tcl/Tk)
– no need for time consuming software specification
– research ideas, user suggestions and feedback on problems rapidly transformed into working software
Rob Bracewell
Prescription
6 Storyboard Tool in Context of New Design Process (Graphs replace text for DDR Requirements, Alternatives & provide scheme folder map)
9 Decide Visualisation, Interaction & Distribution Requirements (Standard UI conventions, no info
lost in b/w print; paste into Word, cope with large projects,
run direct from CD)
13 Specify Experiment
and Data Collection
Software
Rapid Software Prototyping
8 Choose Methods (Graphscript
Tcl bindings of GTL procedures)
11 Rapid Software Prototyping (8 releases in 8 months)
12 Formative evaluation (core team 6 designers in RR, 1 in BAE)
7 Choose Implementation Level Representations (GTL graphs, gml files) 10 Choose Development Tools and Components (Graphlet, TclPro, Ghostscript/GSView,
BLT, Img, Schwartz)
CAEDR
: Issues Crucial to DRed
Description II Stage
• Expand storyboard for use of tool
– No installation required - software to run direct from CD. Vital to run in companies with out-sourced IT. Eased by CAEDR
emphasis on producing platform-independent software – Appropriate training for new users
– Careful consideration of how existing procedures for design reports and folders should be modified when using DRed – Reduced effort writing design report at end of project
– Use of DRed to provide map to other computer files in an electronic design folder
Rob Bracewell
What future plans?
• Further evaluation - analysis of captured rationale
• Patent application submitted
• Aim to expand use in industry and education
• Use with Tablet PC or portable scanner for sketch input
as a designers’ notebook
• Link to Ascend or Modelica for representation and
solution of parameters, equations and constraints