SLIDE 1 ESEVO Quality and Simplicity Frömel
ESEVO Quality and Simplicity
Bernhard Frömel
based on slides by Hermann Kopetz.
- Institute of Computer Engineering
Vienna University of Technology
- 182.722 Embedded Systems Engineering LU
October, 2014
1/48
SLIDE 2 ESEVO Quality and Simplicity Frömel
Part I
Quality
2/48
SLIDE 3 ESEVO Quality and Simplicity Frömel
What is Quality?
”Quality ...you know what it is, yet you don’t know what it is. But that’s self-contradictory. But some things are better than others, that is, they have more
- quality. But when you try to say what the quality is,
apart from the things that have it, it all goes poof!” [Pirsig, 1974] ”Quality cannot be defined. If we do define it we are defining something less than Quality itself.” [Pirsig, 1974] ???
3/48
SLIDE 4 ESEVO Quality and Simplicity Frömel
Some Attempts to define Quality
Quality is ...
◮ ”Conformance to requirements.” [Philip B. Crosby] ◮ ”[the] Degree to which a set of inherent characteristics
fulfills requirements.” [ISO9000]
⇒ Seems to shift the problem to the definition of
’requirement’, ’expectation’ ... More attempts: Quality is ...
◮ ”[the] Number of defects per million opportunities.” [Six
Sigma]
4/48
SLIDE 5
SLIDE 6 ESEVO Quality and Simplicity Frömel
Robert M. Pirsig (born in 1928)
◮ American writer and philosopher ◮ IQ of 170 at age of 9 ◮ Diagnosed with paranoid schizophrenia and clinical
depression
◮ Author of two remarkably successful philosophical novels:
◮ Zen and the Art of Motorcycle Maintenance (published in
1974, after more than 120 rejections)
◮ Lila: An Inquiry into Morals (published in 1991)
◮ In these novels Prisig introduces his Metaphysics of
Quality (MOQ) which aims to give philosophical answer of what quality is.
6/48
SLIDE 7 ESEVO Quality and Simplicity Frömel
Pirsig’s Metaphysics of Quality – A Philosophical Model of Reality
Quality
reality classic
rational painting brushes imitation romantic subjective intuition creativity analytic knife cut engineering creating “discovering” ⇒ Subordinate everything to quality.
7/48
SLIDE 8 ESEVO Quality and Simplicity Frömel
Can we bring quality into being with a method?
◮ Standards? ◮ Coding guidelines? ◮ Testing? ◮ Hard work? ◮ Copying high-quality work? ◮ Do these things guarantee high quality products? ◮ Are they sufficient? 8/48
SLIDE 9 ESEVO Quality and Simplicity Frömel
How is the engineer related to the product?
”Die Ordnung und Regelmäßigkeit an den Erscheinungen, die wir Natur nennen, bringen wir selbst hinein, und würden sie auch nicht darin finden können, hätten wir sie nicht, oder die Natur unseres Gemüts ursprünglich hineingelegt.”, Immanuel Kant Kritik der reinen Vernunft
9/48
SLIDE 10 ESEVO Quality and Simplicity Frömel
Imitation versus Creation
◮ It seems that most of our professional (and personal) life
is based on imitation
◮ School / University ◮ Guidelines in companies ◮ International research lines ◮ Existing literature
◮ Typical thoughts of a student when given an exercise
◮ ”What do they want me to do?” ◮ ”How does one do this?” ◮ ”How many pages should I write?”
10/48
SLIDE 11 ESEVO Quality and Simplicity Frömel
...and it continues in the professional life
◮ ”I am a researcher, so I have to find a problem to work on
...I will look what others did, and see how to extend it”
◮ ”I will look at other embedded systems papers to find out
how to write an introduction/evaluation/conclusion/...for my work”
◮ ”They want experiments at the conference, so I include
some numbers”
◮ ”I look at the reviewer guidelines and bunker my paper”
This way, being creative seems to be very hard work!
11/48
SLIDE 12 ESEVO Quality and Simplicity Frömel
Requirement: Genuine interest
◮ Interest and excitement cannot be forced ◮ It has to arise naturally ◮ Seeking something that might be intersting to me
prevents true interest Interest will arise naturally if one gives up seeking! (side note: Concept of wu wei, i.e., passive creativity, in Taoism)
12/48
SLIDE 13 ESEVO Quality and Simplicity Frömel
Requirement: Freedom of the known
”The birth of a new fact is always a wonderful thing to
- experience. It’s dualistically called a ”discovery”
because of the presumption that it has an existence independent of anyone’s awareness of it. When it comes along, it always has, at first, a low value. Then, depending on the value-looseness of the observer and the potential quality of the fact, its value increases, either slowly or rapidly, or the value wanes and the fact disappears.” [Pirsig, 1974]
13/48
SLIDE 14 ESEVO Quality and Simplicity Frömel
Requirement: Attitude
◮ The ego can be one of the biggest traps regarding quality ◮ An ego that is too strong will never have humility to learn
◮ Very often the case with people that have to defend their
high position
◮ An ego that is too weak will suppress all the excitement
and kills motivation
◮ Very often the case with people in a lower position
True creativity requires ending the attachment to the
We are not talking about indifference!
14/48
SLIDE 15 ESEVO Quality and Simplicity Frömel
Requirement: Ask questions
◮ Don’t take exisiting practices as ultimate truth ◮ Question prevailing opinions ◮ Question your own opinions ◮ Never be shy to ask questions
True, there are stupid questions that you shouldn’t ask
⇒ Exactly those questions are stupid that you only ask in
- rder to demonstrate how clever you are
15/48
SLIDE 16 ESEVO Quality and Simplicity Frömel
Requirement: All the other classical skills
◮ Commitment ◮ Experience ◮ Willingness to spend time and energy ◮ Knowledge ◮ Methodologies and tools
◮ Select them wisely
16/48
SLIDE 17 ESEVO Quality and Simplicity Frömel
Quality tends to fan out like waves
”The Quality job he didn’t think anyone was going to see is seen, and the person who sees it feels a little better because of it, and is likely to pass that feeling
- n to others, and in that way the Quality tends to
keep on going.” [Pirsig, 1974]
17/48
SLIDE 18 ESEVO Quality and Simplicity Frömel
Part II
Simplicity
18/48
SLIDE 19 ESEVO Quality and Simplicity Frömel
Software for Dependable Systems
A report on Software for Dependable Systems: Sufficient Evidence? [Millett et al., 2007] from the US National Academies contains as one of its central recommendations: One key to achieving dependability at reasonable cost is a serious and sustained commitment to simplicity, including simplicity of critical functions and simplicity in system interactions.This commitment is often the mark of true expertise.
19/48
SLIDE 20 ESEVO Quality and Simplicity Frömel
Abstraction
”[Abstraction is] the process by which the particular is subordinated to the general, so that what is represented is applicable to many particulars.”, [Zeki, 2002] Abstraction is a fundamental innate task of the brain.
20/48
SLIDE 21
The Key to Success: Finding Proper Abstraction
In celestial mechanics, when we are interested in the interactions between heavenly bodies, we build an abstraction where we put aside the diversity of our world and consider it be a single mass point – the ultimate simplicity.
SLIDE 22 ESEVO Quality and Simplicity Frömel
Abstraction forms Categories
◮ A category is a set of elements that shares a set of
common characteristic features
◮ Often a name (string, sound) is assigned to a category.
This name is also used to denote an element of the
- category. This double-use of a name – for the category as a
whole and for an element of the category – can be sometimes misleading.
◮ The notion of category is recursive: the elements of a
category can themselves be categories (hierarchical composition is a basic abstraction mechanism).
◮ The recursion stops at primary categories that are subject
to direct sensory experience.
22/48
SLIDE 23 ESEVO Quality and Simplicity Frömel
A Concept is Part of an Implicit Theory
A category that is augmented by a set of beliefs about its relations to existing knowledge is called a concept.
◮ The set of beliefs relate a new concept to already existing
concepts and provide an implicit theory (i.e., a mental model) of the domain.
◮ The theory explains how the individual interconnects the
diverse concepts of the domain and understands their interrelationships.
◮ As a new domain is penetrated, new concepts are formed
and existing concepts are strengthened.
23/48
SLIDE 24 ESEVO Quality and Simplicity Frömel
Denotation versus Connotation of a Word
◮ Denoation: The association of a word with the essence of
the concept that is shared by the language community (denotational semantics).
◮ Connotation: The associations in addition to the essence
- f the concept that a word activates in an individual
speaker or some group of speakers – often referring to emotive or affective aspects of the concept. Example: Fireplace Denotation: a place where a fire can be maintained Connotation: cozy, romantic, personal experience
24/48
SLIDE 25 ESEVO Quality and Simplicity Frömel
Sameness of Concepts
Are the concepts formed by different people the same?
◮ The essence of a concept can be assumed to be the same
within a language community (denotation).
◮ Different persons will associate different shades to
meaning with a concept, depending on their individual concept base (the personal image) and the differing methods of concept acquisition (connotation).
◮ The vagueness of concept definition is of particular
concerns, when concepts are used at their boundaries.
25/48
SLIDE 26 ESEVO Quality and Simplicity Frömel
Basic-Level Concepts
◮ There is a natural level of categorization, neither too
specific nor too general, that is used in conversation and
- thinking. This categorization leads to basic-level concepts.
◮ Basic level concepts are usually represented in the
language by a single word. Example: furniture – chair – arm-chair
◮ Studies with children have shown that basic-level concepts
are acquired earlier than sub-concepts or encompassing concepts.
26/48
SLIDE 27 ESEVO Quality and Simplicity Frömel
Primary versus Secondary Concepts
In addition to basic-level concepts, we distinguish between:
◮ Primary Concepts: Those that are derived directly from
Examples: warm, loud, bright
◮ Secondary Concepts: Those which are abstracted from
Examples: cow, asset, wealth Basic-level concepts are not necessarily primary concepts.
27/48
SLIDE 28 ESEVO Quality and Simplicity Frömel
Abstraction Ladder: Bessy the Cow [Hayakawa, 1949]
- 8. Wealth
- 7. Asset
- 6. Farm asset
- 5. Livestock
- 4. Cow (basic level concept)
- 3. Bessy over time
- 2. Image of Bessy at an instant
- 1. Bessy at an instant (outside RT entity)
28/48
SLIDE 29 ESEVO Quality and Simplicity Frömel
Example: Celestial Mechanics
In Celestial Mechanics the world, with all its diversity, is represented as a single point of mass. The interactions of different planets (mass points) by the forces of gravity give rise to new complexities: the multi-body problem. At any given level of abstraction, the complexity grows until a proper conceptualization at a higher level of abstraction leads again to an abrupt introduction of simplicity.
29/48
SLIDE 30 ESEVO Quality and Simplicity Frömel
Understanding a New Concept
◮ Understanding a new concept is a matter of establishing
connections between the new concept and already familiar concepts.
◮ The isolation (abstraction) of characteristic features from a
set of existing concepts leads to the formation of a more abstract new secondary concept.
◮ A more abstract concept can only be understood bottom
up by generalizations from a set of a suitable collection of examples of already acquired concepts.
30/48
SLIDE 31 ESEVO Quality and Simplicity Frömel
Example: Concept of Counterfeit Money
A counterfeit money bill looks like an authentic money bill. In this situation examples and prototypes are of limited utility. In
- rder to understand the concept of counterfeit money, it is
necessary to understand and relate this new concept to:
◮ Concept of money, ◮ Concept of legal system, ◮ Concept of a national bank that is legalized to print money, ◮ Concept of cheating, ◮ ... 31/48
SLIDE 32 ESEVO Quality and Simplicity Frömel
Six Desired Properties of Scientific Concepts
◮ Utility: The new concept should serve a useful
well-defined purpose.
◮ Abstraction and Refinement: The new concept should
either be a basic-level concept or an abstraction or an refinement of a basic-level concept.
◮ Precision: The characteristic properties of the new
concept must be precisely defined.
◮ Identity: The new concept should have a distinct identity
and should be significantly different from other concepts in the domain.
◮ Stability: The new concept should be usable uniformly in
many different contexts without any qualification or modification.
◮ Analogy: Similarities with other concepts should be
pointed out to establish links to existing conceptual landscape.
32/48
SLIDE 33 ESEVO Quality and Simplicity Frömel
Ontology
”An ontology is a specification of the conceptualization of a domain of discourse.” [Gruber, 1993]
◮ A conceptualization offers an abstract (simplified) view of
a part of the world by providing definitions of the relevant shared concepts and establishing the relationships among them.
◮ An ontology outlines a domain specific vocabulary and
defines an implicit theory about a domain of discourse.
◮ An ontological commitment is an agreement to use the
shared vocabulary in a coherent and consistent manner.
33/48
SLIDE 34 ESEVO Quality and Simplicity Frömel
Compositionality of Concepts in Natural Languages
Is the meaning of a concept influenced by the context of its use? Compositionality: ”An expression makes a uniform semantic contribution to all the compound expressions in which it is embedded.” [Lahav, 1989] then it follows that compositionality of a language implies that ”every adjective has uniform applicability conditions across all compound expressions in which it is embedded” [Lahav, 1989]
34/48
SLIDE 35 ESEVO Quality and Simplicity Frömel
Examples for Non-Compositionality
◮ Red Apple: surface red, but not the internals. ◮ Red Melon: internal is red, but surface may be green. ◮ Red TV Set: frame (housing) red, no picture. ◮ Red Picture: picture red, not frame.
Compositionality simplifies composition, but restricts expressiveness.
35/48
SLIDE 36 ESEVO Quality and Simplicity Frömel
We Understand the World by Modeling
A physical system has a nearly infinite number of properties – every single transistor of a billion-transistor System-on-Chip (SoC) consists of a huge number of atoms that are placed in space have their own identity. We take the view that certain properties, such as determinism or complexity, can only be assigned to models of physical systems, but not to the physical systems themselves, no matter whether these physical systems are natural or man-made. We need to abstract, to build models that leave out the seemingly irrelevant detail of the micro-level, in order to be able to reason about properties of interest at the macro-level.
36/48
SLIDE 37 ESEVO Quality and Simplicity Frömel
Understanding a system
We want to understand a system for a purpose, i.e., to be able to
◮ use the system under normal circumstances, ◮ use the system under adverse circumstances (when
failures are occurring),
◮ maintain the system in case of failures within its
components,
◮ modify the system for a changed environment, ◮ build new applications.
The things we have to know about the system depend on the intended purpose.
37/48
SLIDE 38 ESEVO Quality and Simplicity Frömel
Heuristics for Understanding
◮
”In order to understand anything, you must not try to understand everything.” [Aristotele, 4th
◮ Don’t confuse the functioning of the parts with the
functioning of the system.
◮
”Complex systems will develop and evolve within an overall architecture much more rapidly if there are stable intermediate forms than if there are not.” [Simon, 1969].
◮ Users develop mental models of systems primarily on the
basis of the user-system interface.
◮
”If you can’t explain it in five minutes, either you don’t understand it or it doesn’t work” [David Jones].
38/48
SLIDE 39 ESEVO Quality and Simplicity Frömel
What Makes a Concept Difficult? [Feltovich et al., 2001]
◮ Concreteness vs. Abstractness ◮ Discreteness vs. Continuity ◮ Sequential vs. Simultaneity ◮ Mechanism vs. Organicism ◮ Separability vs. Interactiveness ◮ Universality vs. Conditionality ◮ Linearity vs. Non-Linearity 39/48
SLIDE 40 ESEVO Quality and Simplicity Frömel
One Definition of Complexity
”The complexity of the system is proportional to the number of elements it has, to the number of their interactions, and to the complexities of the elements and the complexities of their interactions.” [Gershenson, 2001].
40/48
SLIDE 41 ESEVO Quality and Simplicity Frömel
Cognitive Complexity
Cognitive complexity is concerned with the questions: How much mental effort is required in order to understand a given scenario for the given purpose by an identified user?
◮ The cognitive complexity of a scenario is proportional to
the time it takes for an average representative from the itended user group to understand the scenario.
◮ The time required for understanding will depend upon the
◮ Purpose of understanding ◮ Assumptions about the conceptual basis of the intended
user group
◮ Inherent characterisctics of the scenario ◮ Representation of the scenario
41/48
SLIDE 42 ESEVO Quality and Simplicity Frömel
Simplification Strategies
◮ Abstraction: The introduction of abstraction layers
whereby only the relevant properties of a lower layer are exposed to the upper layer – Structure and Behavior
◮ Partitioning: The partitioning of a system into nearly
autonomous subsystems (components). – Physical Structure
◮ Isolation: The logical and physical containment of
subsystems.
◮ Segmentation: The temporal decomposition of complex
behavior into smaller parts that can be processed sequentially (’step-by-step’) – determinism helps
42/48
SLIDE 43 ESEVO Quality and Simplicity Frömel
Expert Knowledge
◮ Experts have acquired higher-level concepts that assist in
structuring of present information.
◮ Immediate distinction between surface properties and
deep properties of a scenario.
◮ Automatic (unconscious) handling of irrelevant
information.
◮ It takes a long time to become an expert: 10
years [Simon, 1969]
43/48
SLIDE 44 ESEVO Quality and Simplicity Frömel
Credits References
Part III
End – Thank You!
44/48
SLIDE 45 ESEVO Quality and Simplicity Frömel
Credits References
Credits
◮ Images:
◮ https://www.flickr.com/photos/23397895@N08/ ◮ http:
//eyes.nasa.gov/images/eotss/bg-eotss.png
◮ http:
//sillylittlegiggles.com/stupid-questions/
45/48
SLIDE 46
References I
[Feltovich et al., 2001] Feltovich, P. J., Coulson, R. L., and Spiro, R. J. (2001). Learners’(mis) understanding of important and difficult concepts: A challenge to smart machines in education. In Smart machines in education, pages 349–375. MIT Press. [Gershenson, 2001] Gershenson, C. (2001). Complex philosophy. arXiv preprint nlin/0109001. [Gruber, 1993] Gruber, T. R. (1993). A translation approach to portable ontology specifications. Knowledge acquisition, 5(2):199–220. [Hayakawa, 1949] Hayakawa, S. (1949). Language in thought and action.
SLIDE 47
References II
[Lahav, 1989] Lahav, R. (1989). Against compositionality: the case of adjectives. Philosophical studies, 57(3):261–279. [Millett et al., 2007] Millett, L. I., Thomas, M., Jackson, D., et al. (2007). Software for Dependable Systems:: Sufficient Evidence? National Academies Press. [Pirsig, 1974] Pirsig, R. M. (1974). The art of motorcycle maintenance. William Morrow, New York. [Simon, 1969] Simon, H. A. (1969). The sciences of the artificial, volume 136. MIT press.
SLIDE 48
References III
[Zeki, 2002] Zeki, S. (2002). Neural concept formation & art dante, michelangelo, wagner something, and indeed the ultimate thing, must be left over for the mind to do. Journal of Consciousness Studies, 9(3):53–76.
