1CS150 Spring 97CS150 Spring 98, Copyright © 1998 A. Richard Newton, Kris PisterCS150 Newton/Pister 2.1.1Outlinem Last time:Ü Design Example: Translating a word problem into a sequentialdesign languageÜ State Transition GraphÜ State Transition TableÜ Mealy and Moore Formsm This lecture:Ü Design versus ImplementationÜ Specification, Description, and DepictionÜ Conceptual Blocks & How to Overcome ThemÜ Good Versus Bad DesignÜ The Role of Language in DesignÜ The CS150 ProjectCS150 Newton/Pister 2.1.2From Ideas to Manufacturable ProductsConceptualizationSpecificationDescriptionImplementationThe process by whichone has ideas...CS150 Newton/Pister 2.1.3Specification vs. Descriptionm Specification: Saying what I want; presents behaviorin terms of results.e.g. ∀A { A[i,j] ← 0}m Description: Saying how to do it; describes behaviorin terms of procedure or process.e.g. for(i=0; i<N; i++)for(j=0; j<M; j++)A[i][j] = 0;m We do not have specification languages for general-purpose digital design. For some special-purposeapplications (e.g. DSP) we do.CS150 Newton/Pister 2.1.4But Can We Really Specify or Even Describe Anything of Real Interest... ?m It is not clear, in today's complex world, with complexinteractions, that descriptive languages are particularly usefulany more:"Any experience of reality is indescribable!" R. D. Laing"...conventional scientific language is descriptive, whereas alanguage used to share experience needs to be depictive. Itwould be a language more akin to poetry, or even music, whichwould depict an experience directly, conveying, somehow, itsqualitative character." Fritjof CapraCS150 Newton/Pister 2.1.5From Ideas to Manufacturable ProductsConceptualizationSpecificationDescriptionImplementationCS150 Newton/Pister 2.1.6"Pure-Science" Problemsm Usually characterized by a small number of well-defined,relatively independent, first-order effects.m Input and required output usually well defined.m Measure of success usually well defined.m Usually well-suited to algorithmic (or recipe-like)solution.m Description or specification languages are, almost bydefinition, adequate!(e.g. minimize the number of ANDs and ORs in acombinational logic function)2CS150 Spring 97CS150 Spring 98, Copyright © 1998 A. Richard Newton, Kris PisterCS150 Newton/Pister 2.1.7"Engineering-Science" Problemsm Characterized by a large number of interacting second-ordereffects.m Often difficult to define inputs and outputs precisely.m Measure of success often time-dependent and ill-defined.m Require the use of good experimental technique, modular design,and iteration.(e.g. Find the best multilevel combinational logic network thatminimizes the area in an IC implementation using a particularlibrary of logic gates while meeting the performance requirementsand not using too much power.)CS150 Newton/Pister 2.1.8"Real Design" Problemsm Characterized by the subtle, qualitative interaction of anuncountable number of effects, integrated with a very smallsubset of quantitative, "designable" variables.m Controllable and measurable variables are usually misleadingif used to manipulate and evaluate the overall quality of asystem.m The measure of success is more a "feeling," how much people"like" what you have done, than anything else - and that in itselfis very time and context-dependent.m The only viable approach is to use a depictive language (buildscenarios, build realistic models, build prototypes) and to "live"with it - to experience it directly in it's real-world context.CS150 Newton/Pister 2.1.9Classical Approaches to Designm Design as a Refinement of Representations (Katz)Ü Design SpecificationÜ Design ConstraintsÜ Design as Representationm Implementation and Assembly (Katz)Ü Top-Down DecompositionÜ Bottom-Up Assemblym Other Common Approaches:Ü Meet-in-the-Middle DesignÜ Waterfall Design MethodologiesÜ Cooperative Design Methodology (e.g. Unix, X-Windows, Common Lisp)Ü Object-Oriented DesignCS150 Newton/Pister 2.1.10Bottom-Up or Top-Down?m "In engineering one often finds the "Rube Goldberg"solution - the problem is solved by an inelegant andcomplicated collection of partial solutions."CS150 Newton/Pister 2.1.11The Paradoxof Engineering TechnologyComplexityTimeNew products oftencomplex & hard to useSimpler, more reliable,and more powerful productIncreased power and"capability" at expenseof added complexity(and decreased reliability).CS150 Newton/Pister 2.1.12The Paradoxof Engineering TechnologyComplexityTimeUser Documentation"Next Generation"Solution3CS150 Spring 97CS150 Spring 98, Copyright © 1998 A. Richard Newton, Kris PisterCS150 Newton/Pister 2.1.13Conceptual Blocksfrom "Conceptual Blockbusting," James L. Adamsm Perceptual Blocksm Emotional Blocksm Cultural & Environmental Blocksm Intellectual & Expressive BlocksCS150 Newton/Pister 2.1.14Perceptual Blocksm Seeing what you expect to see: Stereotypingm Difficulty in Isolating the Problemm Tendency to Delimit the Problem Area Too Closelym Inability to See the Problem from Various Viewpointsm Saturationm Failure to Utilize All Sensory InputsCS150 Newton/Pister 2.1.15Emotional Blocksm Barnyard...m Fear of Taking a RiskÜ Assess the possible negative consequences of an idea: Whatare your catastrophic expectations?m No Appetite for Chaosm Judging Rather than Generating Ideasm Inability to Incubate: Sleep on it...CS150 Newton/Pister 2.1.16Cultural and Environmental Blocksm Cultural BlocksÜ Taboos: Ping-pong & Pipe Example.Ü Fantasy & Reflection are a Waste of TimeÜ Playfulness is only for ChildrenÜ Design is too Serious for Humor!Ü Reason, Logic, Numbers, Utility, Practicality: GOOD; Feeling, intuition,qualitative judgmentsCS150 Newton/Pister 2.1.17A steel pipe is embedded in the concrete floor of a bare room, as shown below. Theinside diameter is 0.06" larger than the diameter of the ping-pong ball (1.50") that isresting at the bottom of the pipe. You are one of a group of six people in the room,along with the following objects:100 feet of clothesline A light bulbA carpenter`s hammer A chiselA box of Muesli A fileA wire coat hanger A monkey wrenchList as many ways as you can think of (in five minutes!) to get the ball out of thepipe without damaging the ball, the tube, or the floor.CS150 Newton/Pister 2.1.18Cultural and Environmental Blocksm Environmental BlocksÜ Lack of Cooperation & Trust among ColleaguesÜ Autocratic Boss who Values only His/Her Own Ideas; Does not RewardOthers.Ü Distractions: Telephone, Easy IntrusionsÜ
View Full Document
Unlocking...