Case study

Bigton City Car Parking

The large regional city of Bigton has a busy central business district (CBD). Over the last few years parking in the CBD has become a concern. One difficulty is that too many people overstay the time limits of the parking bays in the CBD's streets. Employing enough parking officers has also become a problem, so the Bigton City Council has decided to look for a technological solution to the problem.

City Engineer, James Jones, proposed that the city purchases and installs a commercially available solution that detects vehicles that have overstayed the time limit. However, after examining one such solution, the Council decided that it provided a lot of features the city would never use and it was too expensive. The Council asked James to investigate creating the city's own solution. Its only purpose would be to detect overstaying vehicles and notify parking officers where these vehicles are parked.

After some thought and investigation, James drew the plan shown below.

Parking Infringement Monitoring System (PIMS)

	A Car Detection Unit (CDU) in each parking bay detects the arrival (or departure) of a car. When an arrival (or departure) occurs the CDU sends a signal to a nearby Road-side Data Logger (RDL). The signal identifies the CDU and the presence (or absence) of a car.
	At set intervals, RDLs send the data they have collected to a dedicated computer at the Council's offices. This is done via the Internet, using the mobile phone network.
	The Council's dedicated computer determines which cars have stayed too long. It then sends the location of the cars and the overstay details to a parking officer's mobile phone.
	The parking officer locates each car and confirms that it is the overstaying car. If so, the officer gets the car's details, prints a ticket, and then sends this information to the Council's fines-payment information system.
	From time to time the City Engineer will access the dedicated computer to obtain statistics regarding the operation of the system.

During his investigation James found out that the Council gives all its parking officers a mobile phone for use on the job, which means that these do not have to be purchased for the system; they would just need some software written for them. He also found that all the other hardware needed for the system, including a portable Bluetooth printer, can be purchased 'off the shelf'.

The CDUs and RDLs operate automatically, but the dedicated computer would need software written for it. The computer can be connected to the Council's existing network, which would allow it to interact with the Council's fines-payment system.

To take his idea further, James hired programmer Kirsten Koder.

After interviewing James, Kirsten drew a use case diagram of the proposed Parking Infringement Monitoring System (PIMS). It is shown below.

The Car Detection Unit (CDU)

The type of CDU James has decided to use in the system is a small self-contained device about 5 cm long. It is designed to be placed just under the surface of a road. When the magnetic field around it is disturbed by a car, the CDU transmits a data signal that contains a 12-digit code as shown below.

Since 1990, each parking bay in Bigton has been allocated a bay number, as shown in a section of the city map below. Since the CDU ID number is programmed at the factory and cannot be changed, James has decided that the best way to provide its location is to match the CDU ID to the bay number.

Parking bays

Instructions for Section C

Answer all questions in the spaces provided. Remove the case study insert and read all the information provided before you answer these questions. Answers must apply to the case study.

Question 1

While looking at the plan drawn by James, the engineer, Kirsten noted that there would be four different data transfers. She also noted that one of the data transfers will not need TCP/IP.
Identify this data transfer and explain why TCP/IP is not needed.

Data transfer from CDU to RDL

Explanation

The transfer is not being routed through the internet like the other three, and to use the internet you must use TCP/IP for packet handling and routing. The transfer might use Bluetooth instead, since the CDU and RDL are said to be "nearby".

1+2 = 3 marks

State average was 52%

Most students were able to identify that the data transfer from CDU to RDL was the one that did not require TCP/IP. However, many students struggled to provide a technically accurate description as to why it was not required. TCP/IP is a protocol for Internet use and the use of the Internet is not required for this data transfer. Stating 'because it is not required' was not sufficient and simply restated the question.
The following is an example of a high-scoring response.

"Data transfer from CDU to RDL
Explanation: The data does not transfer using the Internet. Because TCP/IP is only for data transfer over the Internet the protocol is not required."

Question 2

To check that she had understood James' plan, Kirsten showed him the use case diagram. He expressed concern that the dedicated computer was not shown in the diagram.
State why Kirsten did not include the dedicated computer in her use case diagram.

The actors shown in the UCD are external entities that exploit functionality within the system.  Since the dedicated computer is the system it does not interact with itself from outside of itself!

1 mark

State average was 40%

Students had some difficulty with written responses regarding use case diagrams. As use case diagrams are new to the study, teachers are encouraged to provide more questions of this type throughout the year. Appropriate responses made comment about the dedicated computer not being included in the use case diagram because it was not being an actor (but part of the system).

Question 3

Kirsten will include the use case diagram in the Software Requirements Specifications (SRS) she is writing.
In addition, Kirsten will include a context diagram in her SRS.

State two functional requirements of the Parking Infringement Monitoring System (PIMS) that this context diagram would show. Give an example of each.

Functional requirement 1 - Calculate if cars have overstayed their parking time.

Example - ? For example, a car stays too long in a parking bay??

Functional requirement 2 - Transfer ticket details

Example - Send ticket data to the fines payment system. (??)

First problem: a context diagram will show no functional requirements. A context diagram doesn't show any processes: the system is a black box.  The CD will only show data flows between the system and its external entities. FRs might be suggested by the data flows, but the FRs themselves are not shown.

Second problem: how can you provide examples of specific functional requirements?
It's not as if FRs are a class, and there are instances of them in the UCD.
So if a FR is "print parking ticket", how do you provide an example of that?

Third problem: a big part of the system is calculating when cars have overstayed their time, but this FR is not shown in the UCD. The closest use case is "register car arrives and/or departures" which is different to calculating the time intervals between the events.

I'm confused.

4 marks

State average was 48%

A range of responses was accepted as long as the student could clearly articulate the functional requirements of the system. Functional requirements include:
• send data from roadside logger to central computer
• CDU detects the arrival/departure of a car
• transfer data from CDU to RDL
• send data to parking officers.
Functional requirements are statements about what the system is required to do; however, when responding to this question, many students confused functional and non-functional requirements. Reliability, robustness and maintainability were common answers, but these are all non-functional requirements.

Question 4

When writing the SRS, Kirsten identifies user-friendliness, response rates and maintainability as important non-functional requirements of the system.
Which of these three requirements is the most critical for the success of the system? Justify your choice.

Most critical non-functional requirement - Response rates.

Justification - since the allocation of parking tickets is strictly time-dependent, any lag in getting data on car arrival/departure times could lead to overstaying cars leaving a bay before the parking officer can get there to deliver the ticket.   A responsive system will react in a timely manner to critical events like these.

1 + 3 = 4 marks

State average was 46%

The most critical non-functional requirement in this situation is response rate. When justifying a choice, students who appropriately contrasted response rate with the other non-functional choices generally provided better responses.
It is important that students are conversant with the terminology used in the problem-solving methodology stated on pages 16–18 of the study design. It was apparent that many students did not recognise, or understand, the term 'functional requirement' – part of the analysis stage of the problem-solving methodology.

The following are examples of high-scoring responses.
Justification: The success of the system depends on whether it can inform parking officers of overstaying cars in time for them to give these cars a ticket. The system's response rates will therefore need to be quick enough to get this information to the officers in time for it to be used. User friendliness and maintainability do not necessarily effect whether the system fulfils its purpose, but in this case the response rates of the system are critical to its success,

Justification: Seeing as how the parking officer will need constant updates as the whereabouts of each overstaying vehicle, the response times between detecting an overdue car then the RDL sending this information through to the council's dedicated computer and then on to the parking inspector will need to be very short. How easy the program is to use is irrelevant because there is no handling by users, it is all automatic beside the information received by the officer, and the maintainability, whilst in its own right be an important part, is not considered as it is required but the devices are assumed to be very robust.

The analysis is now complete. Kirsten has submitted the SRS to James and he has told her to begin designing the software for the dedicated computer and the mobile phones.

Kirsten starts with the dedicated computer. It will deal with the use case 'register car arrivals and/or departures'. Kirsten has drawn the following data flow diagram (DFD) to help her with the design of the software.

1. I've always told my kids that each piece of data should have its own arrow. This DFD has multiple data items on the same arrow.

UPDATE - 28 June 2012 - after some research, I can't find an 'official' rule against multiple data items on a flow, but they are nevertheless VERY rare. In one nice slideshow on data flow diagrams I found an example of two data items on one arrow - but that was a higher level DFD, and the two data items were split up into two arrows in the lower level DFD that showed more detail.

2. Aren't double-headed arrows forbidden in DFDs?

UPDATE - 28 June 2012 - Yes. Data flows can only go in one direction!

3. Why are two arrows unlabelled?

Did someone change the DFD rules when I wasn't watching?

For the process 'Determine deadlineTime & cduBattery' she has developed the following algorithm. It updates deadlineTime and cduBattery in cduFile. The variable currentTime is set by the system and represents the current time of day in 24-hour time.

Question 5

The structure of cduFile is shown in the table below.

Using the information in the DFD, the algorithm and case study, complete the table to indicate the most appropriate data type for each field.

Select from: integer, floating point number, Boolean, array, string.

The deadlineTime data type is awkward. The most efficient type would have been 'time' but that was not allowed. Storing it like "1345" meaning 1:45pm is also clumsy and requires processor-intensive parsing. It would be far better to store it as a number of seconds since midnight, which comfortably fits in a short integer.

4 marks

State average was 74%

Only responses from the provided list were accepted. The VCE Software Development Study Design states the following data types: integer, floating point, Boolean, character and string. It is important that students undertaking IT Applications and Software Development are aware of the differences in each study's definition of data types.

Question 6

The updateCduFile algorithm uses a function called bayNumFind. It will search cduFile for the bay number that matches the given cduID.

a.    Kirsten plans to use a binary search in the function bayNumFind.
What does this tell us about how the data must be organised in cduFile?

cduID has to be sorted from smallest to largest.

2 marks

State average was 28%

Most students were able to identify that the file needed to be sorted in order to perform a binary search; however, many failed to identify the data item that should be sorted – cduID.

b.   Kirsten can choose to have cduFile either as a serial file or a random file. Which should she choose and why?

Choice - Random file.

Reason - Records in a random file can be retrieved far more quickly than those in a serial file because the starting byte of any record can be calculated with a simple calculation because every record has the same length. Serial files do not have to have their field lengths pre-defined, so their contents can be far more flexible than a random file, but one has to seek from the start of the file through each record until the required one is found.

2 marks

State average was 40%

Only a small number of students were able to adequately give a reason why a random file was better than a serial file in this instance. Random access files allow direct access to required records/data, so every record does not need to be read to find it as is the case with a serial file.

Question 7

For the process 'Determine overstay bays' Kirsten has developed the following algorithm. It goes through every record in cduFile and finds any that have overstayed. There is one record per parking bay.

Normally currentTime is set by the system, but for testing purposes Kirsten will set it to 1300 (1.00pm). To properly test the algorithm, Kirsten believes she will need to have four different values for deadlineTime. Complete the table below with samples that test the algorithm. State which line is being tested.

8 marks

State average was 41%

In this question students needed to identify key boundary conditions that would test the algorithm logic. Many students confused the variables deadlineTime with currentTime. An overstay message should occur if the deadline time was prior to 1300 (1.00 pm).

Students should be given the opportunity to demonstrate these skills in testing tables throughout the year. Creating their own test tables for assessment is not sufficient; sample algorithms with and without errors and practice questions from past papers should be completed where appropriate to ensure students are well prepared for these types of questions.

Kirsten now works on the mobile phone software. This software will be responsible for the three use cases associated with the parking officers.

Question 8

Messages containing overstay parking data are sent from the dedicated council computer to the mobile phone of the parking officer in the area. The phone software must process the messages in the order they are received.
Which data structure, stack or queue, should Kirsten use to store the messages? Justify your choice.

Choice - Queue

Justification - A queue stores incoming data and releases it in the order in which it arrived. A stack (usually) releases that last datum that arrived most recently, which is not the desired behaviour in this case.

4 marks

State average was 59%

The majority of students were able to identify queue as the correct data structure. It was clear that students had covered this material well as they were able to articulate the fundamental difference between queues and stacks, that being the order is which data is handled – first in first out (FIFO) for queues versus last in first out (LIFO) for stacks. In this case the tickets needed to be processed as they were received, so a queue was required.

The following is an example of a high-scoring response.
Choice: A queue structure should be chosen by Kristen
Justification: Kristen should choose a queue structure because it operates in a FIFO (first in first out) nature. Therefore the first message which will be sent to the officers mobile will be the first which is processed and dealt with which enable officers to deal with the infringements in the order in which they occur, as opposed to always having the first infringement processed last (like the operation of the stack) as offenders could never be punished as they are gone by the time the officer makes it on site.

Question 9

When the parking officers arrive at the parking bays they will enter the car's number plate into their mobile phones. The software must check that each number plate entered is in the correct format. Then each number plate is sent to the dedicated computer. Kirsten has proposed the following algorithm. The number plates can only contain numbers and letters and be no longer than 6 characters. The function length() is used to find the number of characters in a string.

I love the image of parking officers removing number plates from cars and entering them into their mobile phones.
Perhaps the examiners meant to say the officers entered the car's registration number?

*Thanks to Melanie for pointing out that because the test data is 6 characters long (not < 6), it fails the first IF test and drops down to the second numberPlateValid = False line without even checking for invalid characters. I'm not sure the examiners intended this. Were they, I wonder, if thinking that the actual outcomes were (1) valid, (2) invalid, (3) erroneously valid as I originally did?

4 marks

State average was only 18%  !!

When completing a testing table, students should ensure that the expected outcome and actual outcome are based directly on the algorithm in the case study. Many students wrote words such as 'true' or 'invalid', which showed understanding but did not reflect the messages displayed by the algorithm.

b.   Describe the error that this test data shows.

It automatically rejects number plates that are 6 characters long and does not even check them for valid characters.

1 mark

State average was 40%

The algorithm had one error based on the testing table: it does not check the length of the number plate correctly. A number of students responded with 'logical error'; however, the question required a description of the error shown by the test data.

c.    Write down the line of the algorithm that has caused this error.

If length(numberplate)>0 and length(numberplate)<6 then

1 mark

State average was 50%

The line of the algorithm that has caused the error was:

If length(numberPlate) >0 And length(numberPlate)<6 Then

This question asked students to write the line of the algorithm with the error. Students are encouraged to write the whole line of the algorithm, rather than parts of the line or creating line numbers for the algorithm.

d.    Write down the line as it should have been written.

If length(numberplate)>0 and length(numberplate)<=6 then

or

If length(numberplate)>0 and length(numberplate)<7 then

1 mark

State average was 45%

The majority of students found this question challenging. Algorithms and testing are both covered in two of the four Areas of Study and students need to develop their understanding of each. Students should practise these types of questions either in classroom activities using past examination papers or in assessment where appropriate.

Question 10

Kirsten is concerned about the storing of offenders' details if parking officers lost or even left their mobile phones on their desks at work.

State two features Kirsten should build into the mobile phone software to ensure that the data in the parking officers' mobile phones is secure.

For each feature, explain how it would help secure the data.

Feature 1 - password-protected login

Explanation - prevents unauthorised people using the software

Feature 2 - encryption of stored and transmitted data

Explanation - prevents unauthorised people 'sniffing' data travelling wirelessly. Also prevents people being able to read data stolen from the device's storage.

4 marks

State average was 69%

Students' familiarity with mobile phones was in some cases a limitation to them providing accurate responses to this question. A number of students simply described a list of mobile phone safeguards – for example, a timed screen lock– rather than what should be built into the software to ensure the security of the data.

The following is an example of a high-scoring response.
Feature 1: Password authentication
Explanation: This would only enable the authorised people to access the data who knew the password. Therefore, the data could not be easily stolen.
Feature 2: Encryption
Explanation: Even if the phone was compromised the data would be unreadable to everyone except the person or people it is intended for eg by using public/private key authentication. This means that the data even if it is removed from the phones can't be accesses by unauthorised people. Therefore, only the council and parking inspectors need to be able to decrypt this data.

Question 11

A friend, Jose, argues with Kirsten that if she uses single letters for variable and procedure names, her program code would be 20% smaller and hence she would be producing more efficient software.

The names would be meaningless and would require constantly referring to a data dictionary to find out what the variables and procedures contained or did. It could also easily lead to errors by using the wrong variables.

2 marks

State average was 50%

Students needed to demonstrate an understanding of efficiency in their explanation. According to the study design glossary 'measures of an efficient solution include the speed of processing, the functionality of the solution, the ease of use of the solution …' Full marks were awarded to responses that focused on processing and ease of use.

The following is an example of a high-scoring response.
It is the processing of lines rather than reading variable and procedure names that take up processing time, so any efficiency gains would be too small to be measurable. Furthermore, single letter variables and procedure names make code hard to understand making development and maintenance more difficult.

b.   State one other action, besides using meaningful names, Kirsten could perform to make her program code more programmer friendly.

Lots of useful, relevant internal documentation.

The following is an example of a high-scoring response.
Kristen should include meaningful internal documentation.

1 mark

State average was 75%

Most students were able to show that a naming convention and internal documentation were essential elements for the maintainability of the code by the programmer or a future programmer, and that variable length had little, if any, impact on the efficiency of the software.

Question 12

During the development of the software, Kirsten finds a module developed by the Fast Fines company. If she could use this module she would not have to develop the module herself.

Explain Kirsten's legal obligations if she wishes to use this module.

The Copyright Act (1968) says she would need to obtain permission from the owner of the intellectual property before using the code.

2 marks

State average was 65%

Some students failed to provide a detailed response to this question. Students should ensure that they read questions carefully and provide sufficient depth and detail in their responses to ensure the question is answered as asked. To gain full marks students needed to make reference to the legal obligation, namely intellectual property through copyright, and an action that could be taken in order to act legally such as seeking permission or paying a fee for copyright permission.

The following is an example of a high-scoring response.
As the module was created by the Fast Fines Company, it is legally protected by the Copyright Act. If she wishes to use the code she would have to contact the company and ask their permission, as well as most likely pay a cost to be able to use it.

Question 13

Kirsten is sure that all of the errors are now out of the software and that it is ready to be used by the Bigton City Council. The final testing process is to complete the User Acceptance Testing (UAT), but James is not entirely sure what this is and how he will be involved.

a.    Explain the process of UAT and how James and the other staff will be involved.

UAT is testing the opinions of people who will actually use the software once it is released. This is important because these people will have to use it, and can offer valuable feedback from a experienced user's point of view. It is not done by developers, who are too close to the system and may not see its faults, or managers who would be unlikely to know what is important from a regular user's perspective.

3 marks

State average was 47%

Most students struggled to provide a detailed response, limiting their responses to statements about 'checking the system does what is intended'. In addition, many students confused user acceptance testing (UAT) with evaluation.
User acceptance testing occurs in the development stage during testing activities, generally as the last testing step before implementation and user sign-off that the programming solution meets the requirements in the software Requirements Specifications (SRS). The UAT would then be evaluated for its effectiveness as part of an evaluation strategy.

For full marks students needed to address what is involved in user acceptance testing, James' role in this and how other staff would be involved.

The following are examples of high-scoring responses.
The user acceptance testing is done at the end of development and requires the Bigton City Council staff such as James to test the solution to make sure it operates in the intended manner. They will use numerous tests such as accuracy tests, response rate tests to make sure that it operates correctly before signing off on the solution.

A UAT is the process of finding out if the software meets the requirements and the expectations of the end user. The parking staff will be involved in testing the mobile software and each of the other staff members will test if their section functions correctly. Once testing is completed the staff will report back to Kirsten any problems.

b.   Provide an example of a criterion that the new PIMS must meet to pass the UAT process.

It must be easy for the parking officers to use.
It must be reliable and responsive.
Its output (onscreen and printed) must be clearly readable.

1 mark

State average was 45%

A variety of responses, including their form, were appropriate to this question. For example, some responses stated a criterion in terms of a question; others identified if the criterion was a measure of efficiency or effectiveness and then provided an example.
The following are examples of appropriate responses.
The PIMS must be able to send overstay data to the mobile phones of parking inspectors.
Effectiveness – more overdue cars are being found and more fines are being issued.
Does the system send the location of overstayed cars to a parking officer's mobile phone?

Question 14

To evaluate the success of the PIMS, Kirsten needs to collect a range of data.
Suggest a strategy that she might use to acquire data relating to the accuracy of the parking fines that are issued.

• She could check council logs to count how many complaints there have been by motorists about incorrect fines being issued.
• She could also investigate court records to find how many times motorists successfully challenged incorrect fines.
• She could advertise for people to give feedback on the accuracy of parking tickets they have received.
• She could interview parking officers about how often the system has indicated a fine should be given when the officer knew a fine was not actually deserved.

She would not get a car to enter/leave a bay and leave it there for various times, and observe whether tickets were issued by the system - this would be testing, not evaluation.

3 marks

State average was 33%

This question required students to suggest a strategy to evaluate the accuracy of the parking fine issued. A number of students missed the key word 'accuracy' and provided a general answer. As in previous years many students found it difficult to develop a strategy. A strategy is a method by which an activity is carried out; this would include a set of steps or procedures, i.e. What will be done? When does it happen? Who does it? etc. Evaluation questions can be found on previous exams and these questions should be used as appropriate in classwork and in assessments.

For full marks responses needed to consider what data needed to be collected, who was collecting it and from where, and what type of information was being generated to measure the accuracy of the solution.

The following are examples of high-scoring responses.
After approximately 6 months, Kirsten should attempt to collect data as to the accuracy of parking fines produced. To do this Kirsten should consult documents, or logs, letters, emails, telephone correspondence (possibly notes taken by the council when on the phone to clients). All of this information will show how many complaints there have been from customers over the last 6 months to determine how well Kirsten's solution is working. She could also survey customers to see if they notice improvements or agree with the accuracy of the fines. This could be done through the post or email or online survey.

Kirsten may wish to count the number of complaints from motorists in relation to being overcharged in their parking fines. If this number is lower than the previous system then it is an indicator of success although it should be noted that more parking fines are likely to be issued with the new system increasing error amounts.
Kirsten should review any documents relating to the public's complaints with regards to the fines that were issued. She could survey affected people as to whether the parking fine was accurate when they came to pay their fine, bearing in mind that some people will lie about whether they deserve parking fines. She should also interview some parking officers as the whether they have noticed inaccuracies in the new system

Question 15

The key users of the new PIMS developed by Kirsten are the parking officers. The Council wants Kirsten to develop a training program to teach them how to use the new software.

The case study does not say how many officers need training. If there are not too many (e.g. a dozen or fewer), they could be trained in a single group with practical hands-on experience.
If there are many of them, a group lecture could introduce the system to them all, and selected officers could be trained as trainers to teach the rest.
If they were all trained at the same time, it would need to happen outside of normal work hours so people don't get away with illegal parking.

2 marks

State average was 65%

Most students were able to provide a reasonable response to this question. Student generally responded with, 'train the trainer' or 'onsite group training'; however, a range of responses were appropriate as long as they were applicable to the
organisation.

b.   Suggest how the Council could measure the suitability of this program.

They could test trained officers in the use of the system and see how many mistakes the officers made. If trained offers used the system accurately and quickly, the training would be considered a success.

1 mark

State average was 40%

Suitable responses included a method of data collection (for example, a survey, interviewing, a written test or observing) and what was to be collected (for example, questions such as was the training sufficient or who was the data to be collected from – the parking inspector) or when this would occur (for example, at the conclusion of the training program). Many students misread the question and assumed the word 'program' was the software implemented, not the training
program discussed in 15a.

Question 16

The introduction of the new PIMS at the Bigton City Council will affect a number of stakeholders. Outline the conflict that may occur between the Council and each of the following stakeholders.

Parking officers

If fewer parking officers are needed with the new system, some of them might lose their jobs. This would cause a considerable amount of stress and hardship.

Some officers might resent having to learn a new system - especially some older experienced people who have used the old system for years.

Motorists using parking bays

Motorists might believe they are being watched by a 'Big Brother' and resent the intrusion.
They might not believe or trust the evidence of the the sensors and the ticketing system, and might want to challenge its accuracy.

They might resent having their rates used to pay for a system that will fine them more efficiently!

They might be annoyed that the new system is fining them for offences they never got caught for in the past (e.g. overstaying for a short period of time).

2 marks

State average was 30%

Most students found this question challenging, with many unable to clearly describe a conflict. An answer such as 'they may feel annoyed' was not sufficient. The following is an example of high-scoring response.

Parking officers: The parking officers may resist the council's changes and be angry about the new system since they weren't warned or told why. This could result in for example, strike action or an increase in sick days. Motorists will complain that the council is deliberately attempting to raise revenue from PIMS and that the software is unfair/unjust.

END OF QUESTION AND ANSWER BOOK