niedziela, 19 lutego 2017

Discussing REQB mock exam questions



Hi All,

On 17.02.2017 07:40, […] wrote:

Hi Bogdan
I was wondering if you may help us because in one of the exams I found some different answers. You remember that I had the exam from Swedish board and also answers that I found on the Internet and in some cases the answers differ with yours, could you please double check these?

Thanks

_____________________________________________________________________
(13) Which of the following activities would primarily help to improve the quality of a Requirements Specification in an initial state?

a)  Make the specification formal as early as possible to avoid disturbances that can jeopardize the quality

b)  Executing System Test as early as possible to find discrepancies between stated requirement and actual outcome.
c)  Set clear statements of what a system should do rather than how it should do it

d)  Leave the development constraints to next phase, so they do not affect requirements quality

I have answer C here

"C" may have two meanings (and we do not know which one the question author had in mind):
  1. "what" = functional requirements, "how" = non-functional requirements;
  2. "what" = requirements, "how" = design, planned implementation details.
Whichever interpretation you choose, it cannot be "C": "1" is wrong (the syllabus rightly states that non-functional requirements are equally important as functional ones), and "2" is generally right, but not in any way specific for "an initial state" (regardless of whether "initial state" means early requirement's version, or early project stage).
Therefore, the right answer is "A", because "B" and "D" are impossible and wrong. However, some doubt may appear because in several places in the syllabus they say full formality is not always required nor preferable, but on the other hand, it states in many places that "early formal acceptance" is a good thing (please run "<Ctrl-F> formal" on the syllabus to check it).

_____________________________________________________________________

(15)   Which of the following will probably cause most problems when handling requirements?

a.    Prioritization of the requirements
b.    Too formal formulations
c.    Planned changes to requirements
d.    The fact that requirements are feasible

I have answer B here 

"B" is 100% wrong for two reasons: you seldom have problems with excessively formal description (the opposite is true - requirements specifications are commonly too vague), and even if they were too formal, the problem would be mild compared to prioritization problems. The syllabus says much on both subjects (check it, using " priori" and "> formal") but it does not say much on the relative levels of difficulty, sorry.

_____________________________________________________________________

(21)  Which one of the following is a main activity within the process Requirements Elicitation in Requirements Development

a)  Detailing known high-level requirements
b)  Defining and maintaining traceability of requirements
c)  Developing models of the business solution
d)  Elaborating business requirements into system/solution

I have answer A here (Syllabus mentions the answer A, the answer B belongs to Requirement Analysis) 

You and your answer is right! My mistake, SORRY! ("right" in the sense "according to the syllabus").


_____________________________________________________________________
 
(27) Which of the following statements about solution models is NOT correct?

a)  They serve as the basis for system design
b)  They are designed in parallel with the requirements model
c)  They describe various views of the system
d)  They serve as the basis for the estimate of work effort

I have answer D here.

Requirement models "serve as basis for the solution model" (syllabus, page 75), so they cannot be designed "in parallel" (i.e. simultaneously) - ergo, "B" is wrong (i.e. the right answer).

Both types of models [in reality, there is seldom any practical difference, but yes, you should forget about the reality as I told you] are basis for work effort estimation - first, requirements models, and solution models later on.

However - sorry for this! - in the next section on the same page the syllabus says that the "[business] solution model serves as a basis for the solution design". This is correct of course: solution model serves both as a basis for the solution design and for the work effort estimate.

So - yes - from these two sections the question author may have come to the (very wrong) conclusion that solution models are not used for work effort estimate. Hence their answer "D" here . Linguistics and psychology!


_____________________________________________________________________
 
(35)      Which of the following statements about requirements traceability is MOST correct [important]?

a)  It is extremely important to label requirements precisely in order to ensure good traceability

b)  Traceability provides a check on whether all requirements have been labelled precisely.

c)  The aim of traceability is to label interrelationships within a project precisely enough in order to comply with the requirements

d) Traceability is important to keep requirements stable and to ensure that they cannot evolve


I have the answer A here 

Concerning "A": the syllabus says "in order to ensure good traceability, it is important to label the requirements uniquely"; on one hand, the word "extremely" is excessive (both in reality and according to the syllabus") and the crucial word "unique" is omitted. So "A" is not fully correct nor the MOST important aspect of traceability.

The MOST important aspect (and, hopefully, the "MOST correct" answer according to the unusual figurative stylistics of this question) is its GOAL - to identify (and label) the interrelationships (between requirements and other artefacts) - therefore, the answer should be "C".

"B" and "D" are incorrect ("D" is simply WRONG, and "B" is very week).

_____________________________________________________________________
 
(36)      Which of the following statements about metrics is MOST correct [important]?

a.    Metrics must always be set in relation to reference data
b.    It is important to ensure that all metrics are selected

c.    Metrics in Requirements Engineering allow a qualified statement about the quality status of the system

d.    The lower the requirement rating, the greater the risk for the project

I have answer A here ( Syllabus also mentions here a similar answer) 

"A" is very true but very trivial, in the same sense as, for example, a sentence "Metrics is defined by the measurement scale it uses" is. True, but so what? "C" is the main GOAL of using metrics - to be able to "make quantifiable statement about project status and quality" (the preceding statement in the syllabus). And of course "quantifiable" metrics (the syllabus" makes the statement more "qualified" (the question). Quod erat demonstrandum.

_____________________________________________________________________
 
(40)  How exactly do process maturity models prescribe Requirements Engineering?

a)    CMMI and SPICE prescribe Requirements Engineering procedures
b)    CMMI and SPICE contain Requirements Engineering techniques

c)    CMMI and SPICE prescribe what Requirements Engineering must deliver, but not in detail how it is done

I have answer C here 

"B" is wrong (these maturity standards do not specify specific techniques), and "A" and "C" are both right: these standards tell us what [results] RE must deliver and how RE comes into development processes [RE procedures]. The choice is arbitrary. My choice of "A" rather than "C" is based on the assumption that "A" is more general and "C" is too detailed.

Summary: so what now? The problem is, you can almost never be 100% sure what the right answer is, because it depends on what a question author has in mind, and the questions are (like bad requirements) not sufficiently detailed to decide this! So guesswork and learning the way authors may think is the key to success here. Of course, you need to learn to guess what exam questions' authors think, and not Bogdan Bereza thinks . In this case (the SEQB questions) their authors must be Beata Karpinski and Ingvar Nordström. How probable is it that they are as well the authors of (some) real exam questions? I do not know.

Of course, there is as well some probability that certain answers in the "official" answer sheet are simply typing mistakes - this happens as well.


Please let me present you the uncensored version of the slide how these questions are created:
  • The question author takes a random piece / section of the syllabus and finds in it some complex / dubious / hard to remember statement.
  • The author rips this statement out of its context and changes a word or two in it.
  • This doctored statement becomes the "correct" answer, and three more statements are added to it (often sensible, but not from the syllabus)
  • To defeat such questions, the counter-measure it to read the syllabus slowly a few times (it is painful, but not deadly)
Cheers, Bogdan

Brak komentarzy:

Prześlij komentarz