24
Answers to Problems Abstract This section presents answers to the problems set at the end of each chapter (except for Chaps. 1, 15 and 16; no problems were set therein). Note that in most cases there is no single right answer; an answer is presented below with discussion pointing to other ways of doing things. No external references are made from this section; if a reference is needed it is made from the chapter in which the problem was set. Answers to the Problems of Chapter 2 The problem asks which, if any, of the given statements could be used as a claim in a Top Goal, i.e. as the basis of an argument. It also asked if any of the others can be reworded such that they can be claimed. 1. ‘‘Our Quality Management System is ISO 9001:2008 compliant’’ is a claim that could be used un-modified as a Top Goal. ISO 9001:2008 is an International Standard for Quality Management Systems; many companies make this claim, but I have yet to see one use GSN to support it. 2. ‘‘My Business Plan is complete and ready for review by the Board’’: At first sight, this looks to be two claims, but that is quite valid. There is nothing to stop your Top Goal claiming that both A and B are True. In this particular case, however, the Business Plan may well be ready for review by the Board, but it will not be complete until that review is finished, you have put in the changes for which they inevitably asked, and you have formally issued the document for use. ‘‘My Business Plan is ready for review by the Board’’ would be better. I expand on this claim in Chap. 5. 3. ‘‘This burial site is probably that of Rædwald, King of the East Angles’’: A valid claim, but it makes you look uncertain of your case. This statement may J. Spriggs, GSN—The Goal Structuring Notation, DOI: 10.1007/978-1-4471-2312-5, Ó Springer-Verlag London Limited 2012 169

Answers to Problems - Springer978-1-4471-2312-5/1.pdf · Answers to Problems Abstract This section presents answers to the problems set at the end of each chapter (except for Chaps

  • Upload
    doantu

  • View
    225

  • Download
    0

Embed Size (px)

Citation preview

Answers to Problems

Abstract This section presents answers to the problems set at the end of eachchapter (except for Chaps. 1, 15 and 16; no problems were set therein). Note that inmost cases there is no single right answer; an answer is presented below withdiscussion pointing to other ways of doing things. No external references are madefrom this section; if a reference is needed it is made from the chapter in which theproblem was set.

Answers to the Problems of Chapter 2

The problem asks which, if any, of the given statements could be used as a claim ina Top Goal, i.e. as the basis of an argument. It also asked if any of the others canbe reworded such that they can be claimed.

1. ‘‘Our Quality Management System is ISO 9001:2008 compliant’’ is a claimthat could be used un-modified as a Top Goal. ISO 9001:2008 is anInternational Standard for Quality Management Systems; many companiesmake this claim, but I have yet to see one use GSN to support it.

2. ‘‘My Business Plan is complete and ready for review by the Board’’: At firstsight, this looks to be two claims, but that is quite valid. There is nothing tostop your Top Goal claiming that both A and B are True. In this particularcase, however, the Business Plan may well be ready for review by the Board,but it will not be complete until that review is finished, you have put in thechanges for which they inevitably asked, and you have formally issued thedocument for use. ‘‘My Business Plan is ready for review by the Board’’would be better. I expand on this claim in Chap. 5.

3. ‘‘This burial site is probably that of Rædwald, King of the East Angles’’: Avalid claim, but it makes you look uncertain of your case. This statement may

J. Spriggs, GSN—The Goal Structuring Notation,DOI: 10.1007/978-1-4471-2312-5, � Springer-Verlag London Limited 2012

169

well be the final outcome of the debate; however, it would be better to startyour debate with a positive assertion, ‘‘This burial site is that of Rædwald,King of the East Angles’’.

4. ‘‘This painting should be attributed to Albrecht Dürer’’: again, a valid claim.You could be more assertive, ‘‘Albrecht Dürer painted this’’, but the form ofthe claim also depends on the purpose of the argument. If you are cataloguinga collection, and have an opinion about the Unknown Artist label given to oneof the pictures, use the ‘‘should be attributed to’’ form but, if you found thecanvas in your attic and you are trying to sell it, go for the assertive form.

5. ‘‘This equipment fulfils the essential requirements of the RTTE Directive’’ is avalid claim that would normally only be made in a situation where it is knownwhat the RTTE Directive is but, even so, it is better to expand the abbreviationon first use. Alternatively you could identify the directive by number, but it isstill not crystal clear (Chap. 3 contains a means of making it clearer). Alter-native expressions of the claim are, ‘‘This equipment fulfils the essentialrequirements of the Radio & Telecommunications Terminal EquipmentDirective’’ or ‘‘This equipment fulfils the essential requirements of EuropeanDirective 1999/5/EC, known as the RTTE Directive’’.

6. ‘‘Beryllia is a carcinogen’’: Yes, good one; succinct. This is another examplewhere you may need to add information to explain what a carcinogen is—orBeryllia for that matter. What can you assume about what your audienceknows? It is difficult for experts from different domains to have a meaningfulexchange of ideas because they tend to use different vocabulary, or use thesame word to mean different things. You need to be aware of the vocabularyof your intended audience. Are you presenting your argument to inorganicchemists, or to the general public? You will need to expand or contract yourdescriptions accordingly, but do not put extra information in the Goal; it goesin a new symbol, see Chap. 3.

7. ‘‘Hazard Identification and Risk Assessment’’: It is not unusual to find this onein a system safety argument, but it is not a claim. It is not valid as a Goal. Ifyou find a Goal like this, ask the author what they meant. It may be helpful tosuggest an appropriate re-wording; this can show that you have missed thepoint completely and prompt the author to strive for clarity. They may havemeant to claim that, ‘‘Hazard Identification and Risk Assessment was doneand documented’’, or ‘‘Hazard Identification and Risk Assessment showed thatthe risk is tolerable’’, or a host of other things.

8. ‘‘Assurance is provided that safety requirements raised on the software arevalid’’: Well, it is a claim, but a bit ‘‘round the houses’’. Presumably, in thiscase the author will provide an argument assuring us that the requirements arevalid. He, or she, will not be providing an argument showing that someoneelse has already provided assurance that the requirements are valid. We canthus delete the preamble, ‘‘Assurance is provided that’’. Also, note theambiguity due to the missing article, the claim is ‘‘requirements are valid’’, not

170 Answers to Problems

‘‘the requirements are valid’’. It could thus be interpreted to mean that onlysome of them are valid; surely, we want all of them to be valid. A better claimwould therefore be, ‘‘The safety requirements raised on the software are valid’’.This is still a bit strange, one would normally speak of software safetyrequirements; so the claim could be succinctly stated as, ‘‘The software safetyrequirements are valid’’. Half the number of words, but a lot clearer. Note that asimilar claim of software safety requirement validity is the subject of a problemset in Chap. 7, ‘‘The software safety requirements correctly state what is nec-essary and sufficient to achieve tolerable safety, in the system context’’.

9. ‘‘The GSN Symbol for a Goal is a rectangle’’: Yes indeed, but could you argueabout it? This is really a definition, rather than a claim. I consider what to dowith definitions in Chap. 3. Or, is it a fact, an axiom, or a self-evident truth? Ideal with those in Chap. 7.

10. ‘‘The colour of the sky’’ is not a claim. It could be made a claim by choosingan actual colour, ‘‘The sky is blue’’. Even now, this is not a good start for anargument; sky colour is not an invariant, it is different for night and day, forexample. It is different in different places; it depends on the weather, and soon. By the time you have finished preparing your argument the sky may haveturned grey. Think carefully: what is it that you want to argue, who is it thatyou want to persuade, and why? Once you have answers to these questionsyou will be well on the way to expressing a good Top Goal, ‘‘The daytime skyseen from the surface of Mars on a clear day is yellow–brown’’, for example.

Answers to the Problems of Chapter 3

1. If this were a real example, I would have changed the claim to, ‘‘PreventativeMaintenance Procedure PMP5 is fit for purpose’’ and would have providedmore detail as to what ‘‘keeping the power supply running’’ actually means. Butthat was not the question; using the information provided, Fig. A.1 is apotential solution. I could have split the left-hand Context into two; one tospecify what PMP5 is and the other to say where to find it. Of course, in someenvironments, ‘‘PMP5’’ may be sufficient reference, making it unnecessary tostate where it is documented. Note the Context specifying the factory powersupply; this is building on the definition of purpose, rather than the claim itself,but it is shown linked to the Goal, not the Context. Context does not havecontext of its own (unless it is an external document referenced in the Contextitself). It may have been clearer to combine this pair into one Context. In thissolution, I have assumed that the references specify versions, issue states, etc. Ifthey did not, I would add a Context to say, for example, ‘‘This argument is forPMP5 Issue 2 as applied to Power Supply build-state 7.2’’.

Answers to Problems 171

• Note that this last problem is not as contrived as it may seem. It is not unknownfor preventative maintenance procedures to be changed, or dropped altogether,by people who do not know what the purpose of the procedure was. I have alsoencountered the converse; I was told of a procedure that was being regularlycarried out to check a piece of equipment that was only there in case of anincident with some machinery that was no longer in service (in fact no longerthere).

2. This solution is very similar in concept to the previous one; in this case,allegedly, the Butler did It, so we need to specify both Butler and It, seeFig. A.2. Note that I have taken the fictional case reference and emphasized itto form a ‘‘tag’’. If you have a set of small arguments, such tagging makes iteasier to find the one you want.

PMP5 is

fit for purpose

PMP5 is a preventative maintenance procedure,

documented in <reference>

The purpose of PMP5 is to keep the factory power

supply running

The factory power supply is specified in <reference>

Fig. A.1 Potential solution to Problem 3.1

The Butler Did ItThe Butler in question is

that of Lord Symondsbury, Adams

The Butler is accused of stealing silver belonging to

Lord Symondsbury

Case Reference2011/537

Fig. A.2 Potential solution to Problem 3.2

172 Answers to Problems

Answers to the Problems of Chapter 4

1. The expected conclusion is that vinegar is an acid, as shown in Fig. A.3.

2. As well as a Context to present the fletchings definition given in the question, Ihave added one to explain tumbling, see Fig. A.4.

3. Note that, in Fig. A.5, I have included a defining Context for clarity.

Vinegar is an acid

Litmus paper goes

red in vinegar

Litmus paper goes

red in acid

Fig. A.3 Potential solutionto Problem 4.1

Fletched arrows fly without tumbling

Aerodynamic forces on the fletchings of an arrow in flight act to prevent it tumbling

An object is prone to tumbling when in

flight

Tumbling is rotation about any axis other than the

direction of travel

Fletchings are the fins attached near the back of

an archery arrow

Fig. A.4 Potential solution to Problem 4.2

Adelle is Estelle’s

granddaughter

Celine is Estelle’s daughter

Adelle is Celine’s daughter

A person’s granddaughter is the daughter of their child

Fig. A.5 Potential solutionto Problem 4.3

Answers to Problems 173

Answers to the Problems of Chapter 5

1. From the names, this appears to be a French family so, in Fig. A.6, I havephrased my claims in the French manner, rather than using apostrophes.

2. I revert back to using apostrophes in Fig. A.7, in line with the phrasing of theGoal re-used from Chap. 4. These two examples show that there is more thanone way of constructing an argument; we should strive to examine the alter-natives and pick the most compelling or the clearest for use.

Gabrielle is the great aunt of

Adelle

Gabrielle is the aunt of Celine

Celine is the mother of Adelle

A person’s great aunt is sister of their grandparent or the wife of

their grandparent’s brother, i.e. a parent’s aunt

Gabrielle is the wife of Frederic

Frederic is the brother of Estelle

Estelle is the mother of Celine

Fig. A.6 Potential solutionto Problem 5.1

Gabrielle is the great aunt of

Adelle

A person’s great aunt is sister

of their grandparent or the wife of their grandparent’s brother, i.e. a parent’s aunt

Gabrielle is Frederic’s

wife

Frederic is Estelle’s brother

Adelle is

Estelle’s granddaughter

Celine is Estelle’s daughter

Adelle is Celine’s daughter

A person’s granddaughter is the daughter of their child

Fig. A.7 Potential solutionto Problem 5.2

174 Answers to Problems

3. Context applies to the Goal to which it is attached and all of its Sub-Goals. Inthis example, it applies to all the Sub-Goals of a Goal. It is therefore reasonableto apply it to the Goal itself. As a general rule, apply Context as far down theGoal Structure as practicable; it is then clearer for the reader to understand yourmeaning. Invoking a standard, for example, at the top of the structure just addsconfusion if the subject of that standard does not appear until two levels downin the argument structure. Invoking the standard and introducing the subject atthe same level immediately sets the context and aids comprehension.

Answers to the Problems of Chapter 6

1. A Context is missing from the figure. It is needed to state that the three sub-divisions, as represented by the Strategies, cover all the requirements.

2. We can include Context in the Sub-Goal version to show that the three sub-divisions cover all the requirements, as shown in Fig. A.8. Alternatively, wecould add a fourth Sub-Goal to argue that there are only these three types ofrequirement.

Answers to the Problems of Chapter 7

1. Assumption A6.3a includes the preamble ‘‘It is assumed that’’; the symbol is tobe read as meaning that, so it is unnecessary to state it. It should have been just‘‘Vehicle traffic is in the junction’’. Examine that statement; whether it turns outto be true or false, it has no impact on the truth of the Goal. A6.3a is thus anunnecessary Assumption and should be removed. A6.3b contains an externalreference to supporting information; it should therefore be a Context. I already

All functional

requirements have been

successfully verified by

test or by demonstration

All requirements for

automatic functions have

been successfully

verified by test or by

demonstration

All requirements for

user-triggered functions

have been successfully

verified by test or by

demonstration

All requirements for

operator-initiated

functions have been

successfully verified by

test or by demonstration

Functional requirements

specify automatic, user-

triggered and operator

initiated functions only

Fig. A.8 Reducing span by using Sub-Goals to add levels

Answers to Problems 175

have a Context, so I have re-labelled it C6.3a and transformed A6.3b intoC6.3b. A6.3c appears reasonable, but as applied to the Goal, rather than theStrategy. I propose re-drawing the argument fragment as in Fig. A.9.

2. The basic answer is shown in Fig. A.10, in which I added a Context to saywhere the Goal statement came from. I should also have added Contexts to saywhat the system is and where the various requirements referred to are specified.For the next step, Fig. A.11, I add those and use the text of the notes (but I donot need to say ‘‘It is assumed that…’’ in the Assumption symbols).

C6.3bThe failure conditions are specified in the concept of

operations

G6.3On detection of any of

the specified failure conditions, all signal

lights are extinguished

C6.3a“All signal lights”

includes those for pedestrians

S6.3Argue over each

failure condition & any combination thereof

A6.3The system fulfils its

electromagnetic interference susceptibility requirements A

Fig. A.9 One Assumption remains

G1The software safety

requirements correctly state what is necessary and sufficient

to achieve tolerable safety, in the system context

J1NOTE 7

J

C1bNOTE 1

C1aSub-objective A of CAP 670 SW01,

“Requirements Validity”

C1cNOTE 8

A1aNOTE 2

A

A1bNOTE 3

A

A

A1cNOTE 4

A1eNOTE 6

A

A1dNOTE 5

A

Fig. A.10 Basic answer using the suggested shorthand

176 Answers to Problems

G1

The

sof

twar

e sa

fety

requ

irem

ents

cor

rect

ly s

tate

wha

t is

nece

ssar

y an

d su

ffici

ent

to a

chie

ve to

lera

ble

safe

ty, i

nth

e sy

stem

con

text

J1D

urin

g th

e so

ftwar

e de

velo

pmen

tpr

oces

s, fu

nctio

ns m

ay b

e in

trod

uced

whi

ch h

ave

repe

rcus

sion

s on

the

safe

ty o

f the

AT

S s

yste

m. T

hese

will

need

to b

e as

sess

ed a

nd if

nec

essa

ry,

new

or

chan

ged

Saf

ety

Req

uire

men

tsw

ill h

ave

to b

e ge

nera

ted.

J

C1e

The

se r

equi

rem

ents

will

incl

ude

requ

irem

ents

to c

ontr

ol h

azar

ds id

entif

ied

durin

g im

plem

enta

tion

C1f

The

set

of s

oftw

are

safe

ty r

equi

rem

ents

incl

udes

al l

softw

are

safe

ty r

equi

rem

ents

deriv

ed o

r ch

ange

d du

ring

the

requ

irem

ents

det

erm

inat

ion

and

desi

gnpr

oces

ses

A1a

The

sys

tem

-leve

l saf

ety

requ

irem

ents

are

deriv

ed fr

om a

haz

ard

and

risk

anal

ysis

of t

heA

TS

env

ironm

ent i

n w

hich

the

syst

em is

req

uire

dto

ope

rate

A

A1b

A n

eces

sary

and

suf

ficie

nt s

et o

f sys

tem

-leve

lsa

fety

req

uire

men

ts e

xist

, whi

ch d

escr

ibe

the

func

tiona

lity

and

perf

orm

ance

req

uire

d of

the

syst

em in

ord

er to

sup

port

a to

lera

bly

safe

AT

SA

C1a

G1

addr

esse

s S

ub-o

bjec

tive

A o

f CA

P67

0SW

01,

“Req

uire

men

ts V

alid

ity”

A1c

The

failu

re m

odes

whi

ch th

eso

ftwar

e m

ust d

etec

t and

miti

gate

in o

rder

tom

eet t

he s

yste

m s

afet

y re

quire

men

ts h

ave

been

iden

tifie

d e.

g. th

ose

failu

re m

odes

ass

ocia

ted

with

: oth

er s

yste

ms,

sys

tem

-sys

tem

inte

ract

ions

, equ

ipm

ents

, pre

-exi

stin

g so

ftwar

ean

d al

l use

r-sy

stem

inte

ract

ions

.A

A1d

The

failu

re m

odes

iden

tifie

d in

clud

ege

neric

failu

res

rele

vant

to th

e sa

fety

rela

ted

AT

Sap

plic

atio

n, e

.g. s

ecur

ityth

reat

s, lo

ss o

f com

mun

icat

ions

,and

loss

of p

ower

A

A1e

The

failu

re m

odes

iden

tifie

d (in

clud

ing

hum

an e

rror

s) a

re r

epre

sent

ativ

e of

the

oper

atio

nal e

nviro

nmen

t for

the

syst

em a

nd w

orkl

oad

on th

e sy

stem

oper

ator

sA

C1b

The

sys

tem

is to

sup

port

pro

visi

on o

f Air

Tra

ffic

Ser

vice

s (A

TS

); it

is d

escr

ibed

in <

refe

renc

e>

C1d

The

sof

twar

e sa

fety

req

uire

men

ts a

resp

ecifi

ed in

<re

fere

nce>

C1c

The

sys

tem

-leve

l saf

ety

requ

irem

ents

are

spe

cifie

din

<re

fere

nce>

Fig. A.11 Basic answer expanded using note text

Answers to Problems 177

3. The problem is that we have been too literal in our capture of the material andhave missed a significant point. This is a case in which a Customer (actually theCustomer’s Regulator) has told us what to argue; they want us to demonstratethat the claim is true for our software-based system. They stated assumptions,but these are not the conditions in which the claim is true, rather they are theconditions for it to be the right claim. We must not capture them asAssumptions; we must show in our argument (Fig. A.12) that they are valid.

It may have been better, rather than to have detail of a low-level processhighlighted right at the top of the argument, to have had a Justification in place ofGoal G1.3. That Justification would have the same wording as the Goal, but itwould also cross-refer to a lower-level Goal that provides the argument.

I have not captured all the Customer’s assumptions in the decomposition shownin Fig. A.12; to do that I will need to decompose Goal G1.1 another level, seeFig. A.13.

G1The software safety

requirements correctly state what is necessary and sufficient to achieve

tolerable safety, in the system context

G1.1The system safety

requirements correctly state what is necessary and

sufficient to achieve tolerable safety, in the system context

G1.2The software safety requirements

were derived from the system safety requirements, and that derivation was independently

checked

G1.3The Change Process includes a review of safety requirements, augmenting or modifying as

necessary to accommodate the effects of the change

C1aSub-objective A of

CAP670 SW01, “Requirements Validity”

C1bThe system is to

support provision of Air Traffic Services (ATS);

it is described in <reference>

C1cThe software safety

requirements are specified in

<reference>

C1.1The system-level safety

requirements are specified in <reference>

Fig. A.12 Revised top-level argument

178 Answers to Problems

Answers to the Problems of Chapter 8

1. It is conceivable that, if you were to adopt the S convention for numberingevidence, you could have two entities in your argument with the same number,the other one being a Strategy. This does not really matter to the reader, as it isclear from the geometry and the syntax which is meant; it can, however, causeconfusion in review or challenge. For example, someone may have written,‘‘What is the justification for S2.1.2?’’ Do you have to explain why you usedthat Strategy, or why that Evidence is pertinent?

2. We can use the same structure as we did for the Methuselah Report, as shown inFig. A.14. I would hope to see a bit more information in a real argument, seeProblem 3.

3. This part of the argument would be more (or possibly less) compelling if theactual result of the analysis were declared. If the report predicts a failure rate ofonce in a hundred thousand hours, I will be more confident that the claim is truethan if it had predicted exactly once in ten thousand hours. Of course, if the

G1.1The system safety

requirements correctly state what is necessary and sufficient to achieve

tolerable safety, in the system context

G1.1.1The system safety

requirements were derived from a hazard analysis and risk

assessment of the operating environment and the concept

of operations

G1.1.2The set of system safety

requirements is necessary and sufficient to maintain tolerably safe operations when in service

G1.1.4Human Error Effects Analysis has been carried out on the

system in its operating environment (including usability

and workload considerations)

C1.1b“Tolerable safety” is

defined in<reference>

C1.1aThe system-level safety

requirements are specified in <reference>

G1.1.3Failure Modes & Effects Analysis

has been carried out on the system in its operating environment

(including consideration of external and interface failures)

Fig. A.13 Decomposition of system requirements Goal

Answers to Problems 179

report were to predict once in a hundred million hours, I could be less confident,thinking instead that the people who prepared and approved the report may bekidding themselves!Another point of concern here is whether Fault Tree Analysis is valid in thissituation. The logical structure of such an analysis is universally applicable tocausal systems but, like in the Reliability Block Diagram example in the mainbody of the chapter, the failure rate calculations depend on assumptions thatmay not hold for this particular equipment. The use of the technique should bejustified, as should the competence of the analyst, the suitability of any toolsused to produce the results, and the provenance of the data.The argument segment, shown in Fig. A.15, also illustrates another problemwith evidence in some contexts: the potential for mismatch of units. When Imade the claim, I expressed the target in terms of hours but, when the analysisreport arrived, it gave the result in years. Although the widespread use ofstandard ‘‘SI Units’’ has reduced such problems, there are situations in whichother units, such as Knots, have been retained. I suggest either re-expressing thetarget in new units using a Context, as shown, or using a Justification to declarethat the target has been met, in this case because failures are predicted to occurat least eight times more infrequently than required.Also, is it sufficient to point your readers to a, potentially very large, report asevidence of a parameter attaining a target? It would have been better if I hadidentified the pertinent section of each report as Evidence.

G2.3.1The equipment will fail

no more often than once in ten thousand

operating hours

E2.3.1The Fault

Tree Analysis Report

J2.3.1This is argued in the

Fault Tree Analysis Report

J

Fig. A.14 Argument part redrawn using core GSN

180 Answers to Problems

E2.3.1.1The

Fault Tree Analysis Report

C2.3.1Continuous operation,

so ten thousand operating hours is just

over one year

G2.3.1.1Fault Tree Analysis

predicts an equipment failure rate of once in

ten years

G2.3.1.2The Fault Tree

Analysis tool used is fit for purpose

G2.3.1.5Fault Tree Analysis

was carried out on the current version of the

design

G2.3.1.3The use of Fault Tree

Analysis is valid in this case

E2.3.1.2Tool

Verification Report

E2.3.1.3Equipment Verification

Plan

E2.3.1.5Project Quality

Records

G2.3.1.4Fault Trees were

produced by competent staff

E2.3.1.4Staff

Competency Report

G2.3.1The equipment will fail

no more often than once in ten thousand

operating hours

S2.3.1Argument of a valid

analysis of the pertinent design by competent staff with

qualified tools

Fig. A.15 Argument augmented with new Sub-Goals

Answers to Problems 181

Answers to the Problems of Chapter 9

1. If you were to use the Goal to be Developed symbol to indicate that a Goal isdeveloped later on in the same document, it would cause confusion for thereviewer, who will not be expecting to follow that particular chain of thoughtfurther. It would cause even more confusion if your argument were to includegenuine undeveloped Goals. If you wish to include navigation information onthe diagram, use a Label; otherwise, put it in the text below the diagram. Use atabular or bulleted format if there are several such Sub-Goals.

2. I am sorry; I cannot give you an answer to this one. What you need to produceis a personal checklist that includes the detail you need. If you are naturallyvery methodical and/or succinct and accurate in what you write, you can prunethose areas of the checklist, but you may need to enhance other areas… Youmay also wish to add additional questions once you have read Chaps. 12 and 13on the problems of evidence collection.

Answers to the Problems of Chapter 10

1. Although now a deprecated symbol, this seemed an ideal opportunity to use aModel, see Fig. A.16. In a real argument, I would actually have included itlower down the Goal Structure, in the decomposition of Goal G3.3…

2. My proposed solution is set out in Fig. A.17.

G3Evidence of life is found in exoplanet

atmospheres

G3.1Atmospheric components

can be identified by spectrographic analysis as an exoplanet transits its star

G3.2Climate models can predict

the proportions of atmospheric gases both

with and without life present

G3.3Results from spectrographic analyses shows, to a high

level of confidence, the atmospheric gas proportions

that indicate life is present

M3Exoplanet

Climate Model <reference>

Fig. A.16 Potential solution to Problem 10.1

182 Answers to Problems

Answers to the Problems of Chapter 11

1. This is my suggested decomposition; I have split it into two diagrams(Figs. A.18 and A.19) for ease of fitting it on the page. If there are tasks toperform, they must be identified and specified so that training can be provided.That is not enough; we also need to run some evaluation trials so that Users canraise any concerns and we need to monitor performance when the system is inservice, and feedback any problems encountered to the System Authority.

When plants fix CO2 during photosynthesis, they take up 12C, 13C and 14C in the same proportions as in the

atmosphere

After a plant dies (or is eaten) the amount of 14C

declines by beta decay at a fixed exponential rate

A sample’s age can be derived by comparing the proportion of 14C in it with

that expected from the atmosphere

Calibration of the measurements to a known accuracy can be achieved by correlation with other

data sources

We can measure the age of an archaeological sample

of organic matter with a known accuracy

The half-life of 14C is 5730 years

Of the fifteen known isotopes of Carbon, only 12C, 13C and 14C

occur naturally

Fig. A.17 Potential solution to Problem 10.2

Answers to Problems 183

Answers to the Problems of Chapter 12

1. In the absence of any prior use data from other systems, I have to depend on theexperience gained with the component as it was deployed in the systemthroughout testing and evaluation. The usage environment should be verysimilar to that which will be experienced in operation, indeed I have had toargue that elsewhere in the assurance to justify the use of the test set-up. In a

C5.1Task specifications are in the system Master Record Index

S5.1Argue that all tasks have

been specified and validated, also that the users can perform them

G5.1.1

User {Task} has been identified and specified

G5.1.2The {Task} specification has been evaluated by representative users on

the prototype system

G5.1.3{User} has been

successfully trained to properly perform {Task}

>1 >1>1

Fig. A.19 Potential solution to Problem 11.1, View 2

G5

User tasks will be performed properly

C5Proper performance of tasks is defined in the

Human Factors Handbook

S5.1Argue that all tasks have

been specified and validated, also that the users can perform them

S5.2

Argue that performance levels will be maintained

or improved in service

G5.2.1There is a mechanism in

place to measure in-service user performance on {Task}

G5.2.2There is a mechanism in

place to capture in-service user feedback on {Task}

>1>1

Fig. A.18 Potential solution to Problem 11.1, view 1

184 Answers to Problems

real argument I would have bought out the number of hours of use thecomponent had been subject to. Here, in Fig. A.20, I have left the reader to getthat from reading the Evidence reports.

Answers to the Problems of Chapter 13

1. The key point is that, in the fish example, the counter-evidence challenged theTop Goal directly, whereas the test failure challenged a Sub-Goal way down theGoal Structure. In general, if a small part of your argument is refuted bycounter-evidence, it may only be that part that is wrong; it may still be possibleto argue for the Top Goal successfully. Look for a work-around, or (and this isthe better option) develop a different way of supporting your claims.

2. The example found in the fisherman’s catch was dead; therefore the Coelacanthis extinct, but the extinction event was a bit more recent than the previousestimate of sixty-five million years ago. OK, a silly example, but if you are areviewer of arguments, this is the sort of thing you should be looking out for.

E4.2.5.2.1System Failure

Logs

C4.2.5The component

deployed is N8112

E4.2.5.2.2System

Verification Summary

Report

G4.2.5.1The network component was developed by a well-

established network equipment manufacturer

G4.2.5The network

component is fit for purpose for the sales

office system

G4.2.5.2The component was in use

throughout system testing and evaluation without any network

failures attributed to it

J4.2.5.2The test environment is

representative of the operating environment,

see Goal 3.2.2 J

E4.2.5.2.3System

Evaluation Summary

Report

Fig. A.20 Argument based on the deployed component

Answers to Problems 185

Answers to the Problems of Chapter 14

1. I have spread my suggested solution across three figures; one for the Top Goaldecomposition (Fig. A.21) and one each for each of the Strategies (Figs. A.22 andA.23).

G1

Process 7 is fit for purpose

S1.2Argue over the

set of unwanted outcomes

G1.1

Process 7 achieves its stated objectives

G1.2Unwanted outcomes from Process 7 are

mitigated

C1.2“Mitigated” means avoided

or the associated risk is reduced

S1.1Argue over the set of objectives

C1bProcess 7 is documented

as Chapter 7 of the Decorating Manual

C1aProcess 7 provides a

method for properly applying wallpaper to interior walls C1c

Process 7 does not address quantity surveying,

preparation of surfaces, etc.

Fig. A.21 Potential solution to Problem 14.1, Top Goal

186 Answers to Problems

S1.1Argue over the set of objectives

G1.1.1Process 7 provides

methods for selecting and mixing paste

G1.1.2Process 7 provides

measurement, cutting and pasting methods

G1.1.4Each method has been tested and

deemed acceptable by a third party

G1.1.3Process 7 provides

methods for correctly hanging wallpaper

G1.1.3.1Process 7 addresses

starting from a vertical, and pattern alignment

G1.1.3.2Process 7 addresses

papering around doorways, windows, sockets, light switches and radiators

G1.1.3.3Process 7 addresses

papering around reveals, internal corners and

external corners

E1.1.3.3Sections 3.6 & 3.7

E1.1.1Section

2.1

E1.1.2Section

2.2

E1.1.4Specialist’s

Report

E1.1.3.1Section

3.1

E1.1.3.2.1Sections 3.2, 3.4 &

3.5

E1.1.3.2.2

Section 3.3

C1.1The objective of the process

is to give a high-quality finish by correctly pasting and hanging wallpaper J1.1

Process 7 Objectives are clearly stated in the

process document at Subsection 1.1

A1.1The precursor processes, e.g. surface preparation, are properly completed before this process is

started

J

A

Fig. A.22 Potential solution to Problem 14.1, S1.1

Answers to Problems 187

2. I asked you to modify the pattern in Fig. 14.1 to argue for the outputs of acomputer-based tool; see Fig. A.24 for my solution. In practice, such a toolwould be used to automate part of an existing process; if so, you should aug-ment the process argument with the tool argument, rather than replace it.

GXThe outputs obtained from {Tool} are correct

CXa{Tool} is documented

in...

SXArgue that the tool is fit for purpose, it has been used

properly and the outputs have been independently checked

GX.1

{Tool} is fit for purpose

GX.Y{Tool} was used by {Staff}, a competent

person

GX.Z{Output} of {Tool} was

checked independently

n > 0

CXbIndependent => an output is

checked by someone not involved in its production

n > 0

Fig. A.24 Potential solution to Problem 14.2

S.1.2Argue over the

set of unwanted outcomes

G1.2.1

Process 7 mitigates bubbles and creases

G1.2.2

Process 7 mitigates apparent seams

G1.2.3

Process 7 mitigates paste marks

E1.2.1Section

2.2

E1.2.3Section

4

E1.2.2Section

3.1

J1.2Unwanted outcomes were identified from Customer

Questionnaires and Quality Standards J

Fig. A.23 Potential solution to Problem 14.1, S1.2

188 Answers to Problems

Index

AApplications

Assurance Arguments, 1, 11, 13, 36, 41,53, 69, 79, 85, 89, 111, 115, 132,135, 147, 149, 167, 170, 184

Checklist Design, 35, 77Competent Person Arguments, 71, 131, 148Process Arguments, 71, 129, 143, 148, 166Tool Arguments, 68, 71, 140, 141, 166, 188

ArgumentAffirming the consequent (fallacy), 21Against a proposition, 9An invalid decomposition, 46Argument by contradiction, 56Challenge, 24Clarity, 25Clever, 93Commutativity of conjunction, 24Competence, See Applications, 139Contrarian Argument, 57Counter-evidence, 121Deductive, 26Definition, 2Fallacies, 21Generic, See Generic Argument, 103Great subtlety and cunning, 93Inductive, 26Missing evidence, 115Other Notations, 4Partition for Publication, 31, 53Partition with Away Goals, 144Partition with Justification, 53Partition with Strategy, 41Pattern, See Generic Argument, 103Premises, 20Process Argument, See Applications, 139Product Argument, 131

Proposition, 8, 19, 20, 155Safety Argument Re-use, 3Structuring an Argument, See Goal

Structures, 19Subverting a proposition, 9Tool Argument, See Applications, 139

ArrowsBroken, 107Cardinality, 106, 107, 124, 164MooN Relationship, 107, 165Open-headed (Contextual Reference), 12,

16, 17, 49, 55, 75, 159, 161OR Relationship, 106, 165Solid-headed (Thread of Argument), 19,

62, 156, 158Solid-headed, with bobble (Many

Relationship), 105, 124, 164Solid-headed, with O (Optional

Relationship), 106, 164Assumption

Connectivity, 55, 162GSN Symbol, 55, 161Labelling, 55Text Convention, 55, 161Validation, 55, 162

Assurance Argument, See Applications, 139Away

Context, 150Goal, 144Solution, 150

CChecklists

Additional review questions, 90, 101, 111,119, 127, 152

Argument ready for review, 77

J. Spriggs, GSN—The Goal Structuring Notation,DOI: 10.1007/978-1-4471-2312-5, � Springer-Verlag London Limited 2012

189

C (cont.)Authors’, 84Business Plan ready for review, 35Correctness of argument, 88Correctness of relationships, 87Correctness of symbols, 86Quality of presentation, 85Reviewers’, 88Task completion, 35Verification, 36

Claims Argument Evidence Trees, 4Coelacanth, 121, 127, 185Competent Person, See Applications, 139Confirmation Bias, 119Context

Away, See Modular GSN, 150Connectivity, 12, 16, 159Flow Down, 94Generic, See Generic Argument, 106GSN Symbol, 12, 13, 160Inheritance, 14, 160Labelling, 33Text Convention, 12, 14, 160Too much, 16, 60, 177

Counter-evidencePattern, 127Tests all passed, but one, 122Tests all passed, but six, 124Used to advantage, 125What is it?, 121

DDeprecated Symbols

Goal Developed Elsewhere, 66, 72Model, 16, 182

Disclaimer, 3Drake Equation

Formulation, 117History, 116Parameters, 117What is it?, 93, 117

EEvidence

Away Solution, See Modular GSN, 150GSN Symbol, 62, 158Labelling, 63Text Convention, 63, 158

ExamplesAll Men are Aristotle (fallacy), 21Applying Wallpaper, 136Argument by Contradiction, 56

Beryllia, 10, 15, 169Business Plan is ready, 30CAP670 SW01, 58, 176Car salesman, 70Carbon dating, 101, 182Cattery admission, 106Choreography, 139Competency, 131, 148, 180Contractual complication, 110, 152Counter-evidence used to advantage, 125Defendant is Guilty, 24Dürer, 10, 169Euclid’s Theorem, 56Fault Trees, 72, 180Fit for purpose, 134Fundamental Principle of GSN, 35Goal text, 8, 10, 44, 156, 169Human Factors, 111, 183Icosahedron, 53, 58ISO9001, 10, 169Justifying Evidence with a Justification, 67Justifying Evidence with explicit

argument, 68Justifying test failure, 124, 125Methuselah Report, 65Option C is the Best Candidate, 42, 50, 73Planning enquiry, 2, 9, 132Previous Use Argument, 69Process Argument patterns, 130, 132,

134, 138Protection system, 132, 152Radome repair, 135Rædwald, 10, 169RTTE, 10, 15, 169Site has Adequate Security, 25, 43, 51Socrates is a Cat (fallacy), 21Socrates is mortal, 20Specific Absorption Rate, 143, 146Staff Knowledge, 132Staff Skills, 133Tests all passed, 115, 122Tests all passed, but one, 122Tests all passed, but six, 124Tests not run, 115There is life on Exoplanets

See Exoplanet Example, 92Town needs a Bypass, 22Turing Centenary, 9Unit mismatch, 180University admission, 106, 107Wallpaper, 136Xanthippe, 64Yacht repair, 135

Exoplanet Example

190 Index

Climate model, 100, 182Doppler sensor, 95Drake Equation, See Drake Equation, 116Evidence of life is found in exoplanet

atmospheres, 98, 114Goldilocks Zone, 98Green Bank Equation, See Drake Equation,

116Introduction, 91Lack of evidence from SETI, 100, 116Some exoplanets are habitable, 97Spectrometer, 98Strategy for Top Goal, 92The first level of Goal decomposition, 94There are planets in other solar systems, 95Top Goal, 92Transit sensor, 96

Extra Symbols (not GSN)Navigation labels, 75Notes, 75

FFallacy

See Argument, 21

GGeneric Argument

Counter-evidence, 127Documentary Data, 110Generic Assumption, 165Generic Context, 106, 165Generic element symbols, 106, 165Generic elements to be developed, 166Generic Evidence, 163Generic Goal, 104, 163Generic Justification, 165Generic Strategy, 163Identifier, {X}, 104, 162Instantiation Table, 110Introduction, 103MooN Relationship, 166Multiple Relationship, 107, 165Optional Relationship, 106, 164OR Relationship, 106, 132, 165Strategy, 110

GoalAway, See Modular GSN, 144Developed Elsewhere, See Deprecated

Symbols, 66Generic, See Generic Argument, 106GSN Symbol, 7, 8, 156Labelling, 32

Text Convention, 8, 156To be Developed Symbol, 73, 105, 158True or False, 8, 156Vague, 16

Goal StructuresAd infinitum, 131Drafting strategy, 34Labelling, 32Layout and balance, 33Ordering of Sub-Goals, 23Reducing the span, 46Segmenting, 31Sequence of Argument, 19, 23Text Descriptions—are they needed?, 22Using ‘Yellow Stickies, 34

GSN OverviewArgument elements, 155Argument elements to be developed, 158Contextual elements, 159Generic elements, 104, 106, 162Modular elements, 144, 150

HHow to use this book, 2

I’Is Solved By’ relationship, 62

JJustification

Connectivity, 49, 161GSN Symbol, 49, 160Labelling, 50Text Convention, 50, 161

LLabelling Goal Structures, 32Layout and balance, 33Logic type, 26

MMnemonic Labelling, 33, 144, 147Model, See Deprecated Symbols, 16, 182Modular GSN

Away Context, 150Away Goal, 144Away Solution, 150Contract Module, 152Interface Problem, 151

Index 191

M (cont.)Mapping ‘away’ to ‘public’, 151, 152Modular Argument, 147Module, 147Public elements, 151

NNormalisation of Deviance, 125Numerical Labelling, 33

PPlanetarium, viiiProcess Argument

See Applications, 139Process to generate a Goal Structure, 45Product Argument, 131Public elements

See Modular GSM, 151

QQuotations

Boris Grushenko, 21Frank Drake, 117Juvenal, 131Monty Python, 2Richard Feynman, 69Spock, 95Woody Allen, 21

RReasoning, 26Relationships

See Arrows, 36Reviewing

An invalid decomposition, 46Authors’ checklist, 84–88Confirmation Bias, 119, 185Correct argument, 83Correct use of GSN Symbols, 81Correctness of GSN Symbols, 78

Quality of presentation, 77Reviewers’ checklist, 88Scope, 73

SSegmenting Goal Structures

for publication, 31SETI, 93, 116Solution

See Evidence, 62Strategy

GSN Symbol, 39, 157Labelling, 40Text Convention, 39, 157To be Developed Symbol, 74, 159Using for Emphasis, 42Using More Than One, 40

TTim Kelly, 44, 143To be developed

See Goal or Strategy,as appropriate, 73

Tool ArgumentSee Applications, 139

Toulmin Diagrams, 4

UUnified Modeling Language, 17, 147

VValidating Assumptions

See Assumption, 192Verifying checklists

See Checklists, 192

WWigmore Charts, 4

192 Index