113 quiz 3

5.0(1)
studied byStudied by 35 people
learnLearn
examPractice Test
spaced repetitionSpaced Repetition
heart puzzleMatch
flashcardsFlashcards
Card Sorting

1/121

encourage image

There's no tags or description

Looks like no tags are added yet.

Study Analytics
Name
Mastery
Learn
Test
Matching
Spaced

No study sessions yet.

122 Terms

1
New cards

What is the difference between a horizontal and a vertical traceability link?

Horizontal — Within same phase

Vertical — Between phases

2
New cards

What is the difference between backward and forward traceability?

  • Backward traceability 

    • Why is this here?

    • Where does it come from?

  • Forward traceability

    • Where is this taken into account?

    • What are the implications of this?

3
New cards

What is the difference between a revision and a variant?

  • Revision = to correct, improve single-product versions

  • Variants = to adapt, restrict, extend multi-product version

4
New cards

What is the focus of requirements management?

Requirements changes

5
New cards

What is not a common source of requirements changes?

Traceability links

6
New cards

Backwards traceability provides the answer to…?

Why is this here?

7
New cards

A traceability link going from a use case to a source code file would be a __________ link

vertical, forward

8
New cards

The purpose of a _______ is to correct/improve a single-product version

revision

9
New cards

For which type of SW feature is it probably better to use AI?

One in which different content is recommended to different users

10
New cards

When should AI automate rather than augment?

When the task is boring and repetitive

11
New cards

For which type of SW feature is it probably better not to use AI?

One in which maintaining predictability is valuable

12
New cards

If Netflix’s recommendation feature suggests a movie you end up hating, this is a…

false positive

13
New cards

If Netflix’s recommendation feature includes more movies you don’t actually like, it’s probably optimizing for…

recall

14
New cards

Which of the following are standard features of requirements management tools?

  • database

  • traceability

  • change control and management

  • document analysis and generation

  • database

  • traceability

  • change control and management

  • document analysis and generation

15
New cards

True or false: Rational DOORS is the industry standard for RE tools and the one most widely used

false

16
New cards

What type of software would likely require you to follow a domain-specific standard for RE?

Self-driving car

17
New cards

The knowledge of RE standards is rather ______, and the known standards are _____ used

low, barely

18
New cards

Why do requirements change?

  • Problem keeps shifting

  • Stakeholders form a better understanding of system as time goes on

19
New cards

Problems with information management are…

consistency maintenance, change propagation, versioning

20
New cards

Definition of requirements/change mangement

process of anticipating, evaluating, agreeing on, and propagating changes in requirements doc

21
New cards

Source of requirement changes include

  • Requirement errors, conflicts, and inconsistencies

  • Evolving customer/end-user system knowledge

  • Technical, schedule, or cost overruns

  • Changing customer priorities

  • Environmental changes

  • Organizational changes

22
New cards

Features are ______ units

change

23
New cards

Feature changes yield…..

new system versions

24
New cards

Revisions aim to…

correct and improve single-product versions

25
New cards

Variants aim to…

adapt, restrict, and extend multi-product versions

26
New cards

A requirement is traceable if you can identify…

  • WHERE it comes from

  • WHAT/HOW it is used

  • WHY it’s there

27
New cards

Definition of traceability management

  • Identify, document, and retrieve rationale and dependencies/impact of requirements

  • Assess impact of proposed changes

  • Easily propagate changes to maintain consistency among req items + between requirement items and downstream software items

28
New cards

Traceability should be…

identified, recorded, and retrieved; bidirectional (forwards and backwards)

29
New cards

Forwards traceability describes…

source TO target

30
New cards

Backwards traceability describes…

target TO source

31
New cards

Horizontal traceability occurs within _______ phase

the same

32
New cards

Vertical traceability occurs between ______ phases

different

33
New cards

Backwards traceability answers the questions

  • Why is this here?

  • Where does it come from?

34
New cards

Forwards traceability answers the questions

  • Where is this taken into account?

  • What are the implications of this?

35
New cards

Some examples of TM techniques include…

Cross referencing, traceability matrices, feature diagrams (for variant link types), traceability databases

36
New cards

Process for change management is…

  • Define traceability policy (trade-offs, next steps)

  • Establish traceability links (which items to trace? which link types to use?)

  • Exploit traceability links (coverage analysis, support analysis, evolution support)

  • Maintain traceability links (update after changes)

37
New cards

Definition of requirements management is…

Process of managing requirements – focus on change to the requirements for a system

38
New cards

What are some principle concerns of requirements management?

managing changes to agreed requirements

managing relationships between requirements

managing dependencies between requirements doc and downstream documents

39
New cards

Requirements cannot be managed effectively without…

requirements traceability

40
New cards

If changes have ____ impact, you should rerun RE process

high

41
New cards

If changes have ____ impact, you should absorb changes (but be aware of risk!)

low

42
New cards

Scope creep happens when…

new functionality and significant changes that are presented after project requirements have been baselined

43
New cards

Best way to prevent scope creep is to….

just say no!!! (throwback to drug seminars)

44
New cards

Change management describes…

the procedures, processes, and standards used to manage changes to requirements 

45
New cards

In agile and iterative development processes, a requirements change request should (affect/never affect?) the current sprint iteration

never affect

46
New cards

In agile and iterative development processes, a requirements change request should get added to….

the product backlog

47
New cards

AI systems, as compared to non-AI systems, are ..(more/less?) data dependent

more

48
New cards

AI systems, as compared to non-AI systems, evolve… (more/less?) frequently

more

49
New cards

AI systems, as compared to non-AI systems, have …(more/less?) need for continuous monitoring

more

50
New cards

What are some questions to ask when determining whether or not to use AI?

  • Will the solution require AI?

  • Will the solution be meaningfully enhanced by AI?

  • Will the solution not benefit or be degraded from AI?

51
New cards

AI is better when…

  • Recommending different content to different users

  • Predicting future events

  • Personalization improves UX

  • UX requires NL interaction

  • Recognition of an entire class of entities

  • Detection of low-occurrence events that change over time

  • Agent or bot experience for a particular domain

  • UX does not rely on predictability

52
New cards

AI is NOT better when…

  • Maintaining predictability is valuable

  • Providing static or limited info

  • Risk of errors is too high

  • Complete transparency is needed

  • Optimizing for high speed and low cost

  • Tasks to automate are high-value to users

53
New cards

When thinking about AI, we should ask…

  • A: “Can we use AI to ____?”

  • B: “How might we solve____?

  • C: “Can AI solve this problem in a unique way?”

  • “How might we solve____?

  • “Can AI solve this problem in a unique way?”

54
New cards

Automation is successful when there is…

  • Increased efficiency

  • Improved human safety

  • Reduction of tedious tasks

  • Enabling new experiences that were not possible without automation

55
New cards

You should automate when…

  • People lack knowledge/ability to do the task

  • When task is boring, repetitive, awkward, or dangerous

56
New cards

True or false: When you automate, you don’t really need to add an option for human oversight (It’s helpful if it’s there, but not required)

False

57
New cards

Successful augmentation is measured by…

  • Increased user enjoyment of a task

  • Higher levels of user control over automation

  • Greater user responsibility and fulfillment

  • Increased ability for the user to scale their efforts

  • Increased creativity

58
New cards

You should augment tasks when…

  • People enjoy the task

  • Personal responsibility is required or important

  • There are high stakes

  • Specific preferences are hard to communicate

59
New cards

A reward function is…

a mathematical formula that the AI uses to determine right or wrong predictions

60
New cards

The reward function…

Determines action/behavior for which your system will try to optimize

61
New cards

Reward functions are… (major/minor?) drivers of the UX

Major

62
New cards

Precision represents..

the proportion of true positives correctly categorized out of all true and false positives

63
New cards

Higher precision means…

  • Higher confidence that model output is correct

  • Increased number of false negatives by excluding possibly relevant results

  • More accuracy, but less results

(e.g., Netflix only will recommend you 5 movies, but they’re all things you’re interested in seeing)

64
New cards

Recall represents…

the proportion of true positives correctly categorized out of all the true positives and false negatives

65
New cards

Higher recall means…

  • Higher confidence that all relevant results are included somewhere in output

  • Increased number of false positives by including possibly irrelevant results

  • Less accuracy, but more results

(e.g., Netflix will recommend you 100 movies, but you only want to see 5 of them)

66
New cards

Find good candidates for AI by…

  • Finding intersection of user needs and AI strengths

  • Assessing automation vs augmentation

  • Designing and evaluating the reward function

67
New cards

What can be supported by RE tools?

  • Elicitation – analyze textual artifacts

  • Documentation – generating and editing requirement work products

  • Modeling – primarily model editors

  • Management – store and retrieve, prioritize, trace, …

  • Validation – finding quality problems, simulators, model checkers, …

68
New cards

Support levels for RE include…

General purpose » Database level » language/method based

69
New cards

General purpose RE support includes things like..

word processors, spreadsheet tools, general-purpose graphic drawing tools, … (things that are used for other things but can also be used for RE)

70
New cards

Database-level RE support describes...

RE management tools for organizing, storing, retrieving, and tracing requirements

71
New cards

Language/method-based RE support describes things like...

state machine diagrams, UML + tools for supporting specific methods (validation with model-checking or checking doc for compliance)

(tools that support specific requirements language)

72
New cards

Features of requirement management tools include…

  • Database

  • Document analysis and generation

  • Change control and management

  • Traceability

73
New cards

The primary usage of AI in RE is…

processing natural language texts with machine learning

74
New cards

Places where AI’s NL processing can be used includes…

  • Finding and classifying (potential requirements, glossary candidates, smells)

  • Recommending (further stakeholders)

  • Analyze (automated impact analysis when requirements change)

  • Generating (requirements for a given problem, acceptance criteria)

  • Support (chatbots for autonomous interaction with a large number of stakeholders)

75
New cards

What are the three relevant dimensions of RE ethics?

  1. Ethics of profession (how to act ethically as a requirements engineer)

  2. Ethics of use (impact of requirements on user)

  3. Ethics of design (impact of requirements on system design)

76
New cards

What are some ways you could model structure and data?

E-R models, class models, data dictionaries

77
New cards

What are E-R models?

Models the relevant part of the domain using entity types, relationship types, and attributes

A straightforward mapping to relational database systems

78
New cards

What is a potential problem with E-R models?

They ignore functionality and behavior

79
New cards

What are class models?

  • Identifies entities in the domain that the system has to store and process

  • Map this info to classes, attributes, relationships, and operations

  • Represent requirements in a static structural model

80
New cards

What are data dictionaries?

Collection of detailed info about the data entities used in an application (composition, data types, allowed values, …)

81
New cards

Data dictionaries can be used to complement….

project glossaries

82
New cards

Static system models are..

info that a system needs to know and store persistently

83
New cards

Static domain models are…

(business) objects and relationships in a domain of interest

84
New cards

Activity models would be..

  • UML activity diagrams

  • Models process activities and control flow

  • Can model data flow

85
New cards

Data and information flows …

  • Models system functionality with data flow diagrams

  • Are rarely used today

86
New cards

The goal when modeling state and behavior is to..

describe dynamic system behavior

(how system reacts to external events, how independent system components coordinate work)

87
New cards

RE standards describe..

process/products + characteristics/attributes of well-formed requirements 

88
New cards

True or false: there is no requirements engineering body of knowledge document

True

89
New cards

The ISO/IEC/IEEE 2011 is a commonly used RE standards that is…

  • Heavyweight document-, process-centric standard

  • Does not work well for participatory and lightweight RE processes

90
New cards

Validation asks..?

Are we building the right product?

91
New cards

Verification asks…

Are we building the product right?

92
New cards

Analysis works with _____ requirements

raw (aka as-elicited from stakeholders)

93
New cards

Validation works with the (first/final?) draft of requirements doc

final

94
New cards

Validation complements the ________ phases

evaluation/analysis, negotiation

95
New cards

The goals of validation/QA during RE is…

  1. Detecting errors/flaws in requirements documents (no omissions, incorrect info, contradictions… etc) 

  1. Check adequacy of requirements and assumptions (Is requirements doc acceptable prior to further development?)

96
New cards

Tasks of validation/QA during RE include..

  1. Finding as many errors as possible

  1. Report defects, analyze causes, fix

97
New cards

Requirements review and inspections are carried out by….

a group of selected personnel

98
New cards

During requirements review, personnel should…

  • Read and analyze requirements

  • Look for problems

  • Meet and discuss problems

  • Agree on actions to address problems

99
New cards

True or false: Inspectors should be independent from original requirements authors

True

100
New cards

True or false: Inspectors don’t need to be representative of all stakeholders

False