Vous êtes sur la page 1sur 6

HEL SI N K I UN I VERSI TY O F T EC HN O LOG Y P M &R G - P R O DU C T M O DE LI NG & R E A L IZ A TI ON GR O UP

HEL SI N K I UN I VERSI TY O F T EC HN O LOG Y P M &R G - P R O DU C T M O DE LI NG & R E A L IZ A TI ON GR O UP

What Is a Design Rationale? The Elements of a Design Rationale


Henrik.Asplund@hut.fi, Mervi.Ranta@hut.fi Helsinki University of Technology Department of Computer Science & Engineering Laboratory of Information Processing Science (professor Eljas Soisalon-Soininen) PM&RG - Product Modeling & Realization Group
Copyright 2002 Henrik J. Asplund & Mervi Ranta

Tool for capturing and deploying design decisions (Lahti, Mntyl, Ranta) A well documented series of the design decisions and on what basis they were made. The semiformalistic approach for a design rationale presented here is only one of the possible alternatives. Still, we demand use of this model in the seminar.

Copyright 2002 Henrik J. Asplund & Mervi Ranta

HEL SI N K I UN I VERSI TY O F T EC HN O LOG Y P M &R G - P R O DU C T M O DE LI NG & R E A L IZ A TI ON GR O UP

HEL SI N K I UN I VERSI TY O F T EC HN O LOG Y P M &R G - P R O DU C T M O DE LI NG & R E A L IZ A TI ON GR O UP

Capturing and deploying design decisions


Earlier decisions Design Context Issue Criterion Criterion Product model Criterion

Elements of a DR
Issues (what to decide?) Criteria (what is the basis for the evaluation?) Alternatives (what are the possibilities?) Evaluation (how well the alternatives match the criteria?) Decision (which alternative do we select?) Respondent (whos responsible for the decision? who will be hanged?)
Copyright 2002 Henrik J. Asplund & Mervi Ranta

Capturing background information and rationale of design decisions Improving design quality
Maintaining criteria Alternatives considered Explicit evaluation

Alternative Evaluation Evaluation

Supporting teamwork and distributed development Allowing reuse

Respondent

Decision
Copyright 2002 Henrik J. Asplund & Mervi Ranta

HEL SI N K I UN I VERSI TY O F T EC HN O LOG Y P M &R G - P R O DU C T M O DE LI NG & R E A L IZ A TI ON GR O UP

HEL SI N K I UN I VERSI TY O F T EC HN O LOG Y P M &R G - P R O DU C T M O DE LI NG & R E A L IZ A TI ON GR O UP

Capturing and deploying design decisions

Criteria

ISSUE

Decision

HEL SI N K I UN I VERSI TY O F T EC HN O LOG Y P M &R G - P R O DU C T M O DE LI NG & R E A L IZ A TI ON GR O UP

HEL SI N K I UN I VERSI TY O F T EC HN O LOG Y P M &R G - P R O DU C T M O DE LI NG & R E A L IZ A TI ON GR O UP

Issue
The decision to be made
which data format? what kind of cover? Issue: Network technology

Be careful never mix issues and criteria


Example: Issue MUST NOT be How to make this thing light? you must have criterion lightness instead.

Copyright 2002 Henrik J. Asplund & Mervi Ranta

Copyright 2002 Henrik J. Asplund & Mervi Ranta

HEL SI N K I UN I VERSI TY O F T EC HN O LOG Y P M &R G - P R O DU C T M O DE LI NG & R E A L IZ A TI ON GR O UP

HEL SI N K I UN I VERSI TY O F T EC HN O LOG Y P M &R G - P R O DU C T M O DE LI NG & R E A L IZ A TI ON GR O UP

Criteria
Basis for choosing between alternatives
e.g. Must be suitable for innovation prototyping, Must be light, Reasonable cost

Criteria (cont.)
New criteria regularly added
Checking that ok for the current alternative Previous decisions have to be reviewed

Different priorities for criteria some may be considered important, some less important Subcriteria a criterion may be composed of several subcriteria
e.g. Mobility may be composed of lightness small size, area of usage etc.
Copyright 2002 Henrik J. Asplund & Mervi Ranta

Criteria definition is crucial


Write explicit definition for each criterion
Only naming the criteria is not adequate

Deriving subcriteria from a criterion while making specific decisions Understanding the meaning of criteria at later stages
1. e.g. mobility Light, size 2. Ability to move (cable) 3. Area of usage
Copyright 2002 Henrik J. Asplund & Mervi Ranta

HEL SI N K I UN I VERSI TY O F T EC HN O LOG Y P M &R G - P R O DU C T M O DE LI NG & R E A L IZ A TI ON GR O UP

HEL SI N K I UN I VERSI TY O F T EC HN O LOG Y P M &R G - P R O DU C T M O DE LI NG & R E A L IZ A TI ON GR O UP

Defining a Criterion Mobility (an Example)

Global vs. Local Criteria

HEL SI N K I UN I VERSI TY O F T EC HN O LOG Y P M &R G - P R O DU C T M O DE LI NG & R E A L IZ A TI ON GR O UP

HEL SI N K I UN I VERSI TY O F T EC HN O LOG Y P M &R G - P R O DU C T M O DE LI NG & R E A L IZ A TI ON GR O UP

Irrelevant Criteria
Example: When choosing operating system for a device, network bandwidth is irrelevant This is safer than using local criteria If a criteria is irrelevant in one alternative of an issue, it must be irrelevant for all the other alternatives in the same issue!

Criteria and Service Life Cycle


In different stages of product development and commercialization the criteria may be different
Innovation prototyping stage
Reasonable cost Lightness and size are not very important

Conventional prototyping
Mobility and usability as important as in commercialization phase

Commercialization
Low manufacturing cost Mobility and usability are very important

Copyright 2002 Henrik J. Asplund & Mervi Ranta

Copyright 2002 Henrik J. Asplund & Mervi Ranta

HEL SI N K I UN I VERSI TY O F T EC HN O LOG Y P M &R G - P R O DU C T M O DE LI NG & R E A L IZ A TI ON GR O UP

HEL SI N K I UN I VERSI TY O F T EC HN O LOG Y P M &R G - P R O DU C T M O DE LI NG & R E A L IZ A TI ON GR O UP

Criteria Innovation prototyping Mobility Usability Reasonable cost

Alternatives
Issue: Network technology

What kind of possibilities there are Include also nowadays technically unfeasible solutions as alternatives technology changes rapidly

Copyright 2002 Henrik J. Asplund & Mervi Ranta

Copyright 2002 Henrik J. Asplund & Mervi Ranta

HEL SI N K I UN I VERSI TY O F T EC HN O LOG Y P M &R G - P R O DU C T M O DE LI NG & R E A L IZ A TI ON GR O UP

HEL SI N K I UN I VERSI TY O F T EC HN O LOG Y P M &R G - P R O DU C T M O DE LI NG & R E A L IZ A TI ON GR O UP

Criteria Innovation prototyping

Evaluation

HEL SI N K I UN I VERSI TY O F T EC HN O LOG Y P M &R G - P R O DU C T M O DE LI NG & R E A L IZ A TI ON GR O UP

HEL SI N K I UN I VERSI TY O F T EC HN O LOG Y P M &R G - P R O DU C T M O DE LI NG & R E A L IZ A TI ON GR O UP

Evaluation Example I
Issue: Basic Technology Alternatives
Embedded system
Criterion: Reasonable cost Evaluation: 2000 mk 3000 mk (9.8.2000) Evaluation: 4000 mk 5000 mk (5.5.2001) (1)

Evaluation Example II
Issue: Choosing Motherboard Alternatives
Martnode
Criterion: Reliability Evaluation: Has some bugs (-1)

MAS-P5/3
Criterion: Reliability Evaluation: Generic and well-working solution (1)

Laptop
Criterion: Reasonable cost Evaluation: >8000 mk (9.8.2000) (0)

Copyright 2002 Henrik J. Asplund & Mervi Ranta

Copyright 2002 Henrik J. Asplund & Mervi Ranta

HEL SI N K I UN I VERSI TY O F T EC HN O LOG Y P M &R G - P R O DU C T M O DE LI NG & R E A L IZ A TI ON GR O UP

HEL SI N K I UN I VERSI TY O F T EC HN O LOG Y P M &R G - P R O DU C T M O DE LI NG & R E A L IZ A TI ON GR O UP

Criteria Innovation prototyping Issue: Network technology Mobility Usability Reasonable cost

Decision
Select the most suitable alternative
Use the results of evaluation to compare alternatives Remember different weights of the criteria

What if the decision feels strange or stupid?


Standard Ethernet !Innoprototyping + Mobility -Usability IR Reasonable cost + GSM / GPRS WLAN !Innoprototyping - !Innoprototyping + Mobility ++ Mobility + Usability IR Usability IR Reasonable cost Reasonable cost + It is important to analyze the intuition
Did you forget some criteria?
e.g. choosing cable to be the network connection for a mobile device, because ability to move and area of usage were not considered e.g. choosing CD player as a novel mobile music playing service, because innovativeness was not considered

Debate in your group thus you will find the missing criteria
Copyright 2002 Henrik J. Asplund & Mervi Ranta Copyright 2002 Henrik J. Asplund & Mervi Ranta

HEL SI N K I UN I VERSI TY O F T EC HN O LOG Y P M &R G - P R O DU C T M O DE LI NG & R E A L IZ A TI ON GR O UP

HEL SI N K I UN I VERSI TY O F T EC HN O LOG Y P M &R G - P R O DU C T M O DE LI NG & R E A L IZ A TI ON GR O UP

Criteria Innovation prototyping Issue: Network technology

Decision Select WLAN

Respondent

HEL SI N K I UN I VERSI TY O F T EC HN O LOG Y P M &R G - P R O DU C T M O DE LI NG & R E A L IZ A TI ON GR O UP

HEL SI N K I UN I VERSI TY O F T EC HN O LOG Y P M &R G - P R O DU C T M O DE LI NG & R E A L IZ A TI ON GR O UP

Criteria Innovation prototyping Mobility Usability Reasonable cost

Issue: Network technology

Decision Select WLAN Respondent Henrik

Chain of Issues
Formed by issues and selected alternatives Can be branched, if design ends in dead end and another alternative must be selected
Issue 1 Issue 2 Issue 3 Issue 4

Standard Ethernet !Innoprototyping + Mobility -Usability IR Reasonable cost +

GSM / GPRS !Innoprototyping Mobility ++ Usability IR Reasonable cost -

WLAN !Innoprototyping + Mobility + Usability IR Reasonable cost +


Copyright 2002 Henrik J. Asplund & Mervi Ranta

Copyright 2002 Henrik J. Asplund & Mervi Ranta

HEL SI N K I UN I VERSI TY O F T EC HN O LOG Y P M &R G - P R O DU C T M O DE LI NG & R E A L IZ A TI ON GR O UP

HEL SI N K I UN I VERSI TY O F T EC HN O LOG Y P M &R G - P R O DU C T M O DE LI NG & R E A L IZ A TI ON GR O UP

Chain of Issues (cont.)


Issue 1 Issue 2 Rejected decision chains are preserved in the model Preventing developers from making same mistakes over and over. Issue 5 Issue 6 Chain of Issues may become relevant again, when business and technology conditions change
Copyright 2002 Henrik J. Asplund & Mervi Ranta

Things to remember
Always check the whole decision tree after adding criteria There MUST NOT be implicitly decided issues. Everything must be EXPLICIT!
Use of your own brains is allowed and even recommended

Issue 3
Dead end

Issue 4

Reasons for evaluations must be documented simple + and - are not enough!

Copyright 2002 Henrik J. Asplund & Mervi Ranta

HEL SI N K I UN I VERSI TY O F T EC HN O LOG Y P M &R G - P R O DU C T M O DE LI NG & R E A L IZ A TI ON GR O UP

HEL SI N K I UN I VERSI TY O F T EC HN O LOG Y P M &R G - P R O DU C T M O DE LI NG & R E A L IZ A TI ON GR O UP

What Design Rationale is NOT?

DR in scenario generation and development

HEL SI N K I UN I VERSI TY O F T EC HN O LOG Y P M &R G - P R O DU C T M O DE LI NG & R E A L IZ A TI ON GR O UP

HEL SI N K I UN I VERSI TY O F T EC HN O LOG Y P M &R G - P R O DU C T M O DE LI NG & R E A L IZ A TI ON GR O UP

Criteria and scenario generation


The task usually is to find criteria that reveals the difference between the scenarios Local criteria are possibly more feasible than global criteria The criteria can be found by comparing the alternative scenarios and specifying the differences
Comparing two alternative scenarios can be found a criterion that can then be used to compare the rest ten scenarios

Decisions in scenario generation


The point is NOT to select only one alternative, but the most feasible or interesting ones In the previous slides only one alternative was chosen, but it does not have to be case, when generating scenarios When choosing many alternatives, remember, that they all must be considered in following steps of the scenario generation If an interesting alternative is rejected, then check your decisions and criteria
Copyright 2002 Henrik J. Asplund & Mervi Ranta

The criteria are a means to analyze scenarios utilize them fully


Copyright 2002 Henrik J. Asplund & Mervi Ranta

HEL SI N K I UN I VERSI TY O F T EC HN O LOG Y P M &R G - P R O DU C T M O DE LI NG & R E A L IZ A TI ON GR O UP

HEL SI N K I UN I VERSI TY O F T EC HN O LOG Y P M &R G - P R O DU C T M O DE LI NG & R E A L IZ A TI ON GR O UP

Viewpoint Integration
User centered
User studies Interaction models Use scenarios

Technology inspired
Technology studies Gadget and architecture building Performance analysis

Business driven

Market research Profit analysis Annual cash flows




80 70 60 50 40 30 20 10 0 2001 2002 2003 2004 2005

3D modeling device Student workgroup communication

Innovation prototype
Service scenario Design rationale Gadget architecture Assessment criteria

General remote controller gadget Jukebox interface

Profit analysis

MP3 proto

Copyright 2002 Henrik J. Asplund & Mervi Ranta

Copyright 2002 Henrik J. Asplund & Mervi Ranta

Vous aimerez peut-être aussi