Sandbox‎ > ‎Archive‎ > ‎IPT 2008-09‎ > ‎Blair's page‎ > ‎

Blair's updates!

Below you will find the contents of my classwork for Information Processes and Technology, sorted by chronology descending. For an index of this work, please refer to Schoolwork.

2009-12-01 students.jamesruse ?

posted Nov 30, 2009, 12:20 PM by Unknown user   [ updated Dec 14, 2009, 4:25 AM by Eddie Woo ]

Data inputs and outputs

student.jamesruse.nsw.edu.au could be used for student collaboration and discussion. Here is a context diagram:
 
Possible uses for this domain
  1. Permission notes. As of December 2009, permission notes are still provided to students on paper exclusively; digital copies are rarely available, with the exception of some co-curricular groups such as ISCF and JRAHS Cadet Unit. As permission notes on paper have a tendency of being misplaced, it is highly effective to have digital copies available, saving the school and its staff the need to print extra copies. It is known that some students scan permission notes and then distribute them by email anyway.
  2. Addressing the students. It is sometimes necessary to make announcements to the student body, as the daily Day Sheet demonstrates with its often-repeated “students are to consult the main noticeboards for a list of unexplained absences, whole-day and partial”. Regarding this particular example, it is proposed that having both the announcement and the list of unexplained would save time during school and decrease the manual labour required (posting up paper documents on noticeboards).
  3. Inter-student communications. The intranet could provide an official forum for students to discuss school-related topics. Such forums have been trialled, such as the We Can Change The World forum, and have been demonstrated to be feasible. 
Challenges faced & possible solutions

Ensuring usage
  • If the intranet is to be relied on for the purposes listed, then all students must be using it. 
  • This challenge was faced by the officers of the JRAHS Cadet Unit, who had difficulty implementing their section of the intranet as cadets did not habitually check it. The solution used by the officers was to effectively make it the responsibility of the cadets to check the intranet, reminding the constantly until they understood a need to constantly check it. This same solution could be adopted for students.jamesruse.nsw.edu.au.
  • Another possible method of increasing usage is to use RSS feeds and email subscriptions, delivering content immediately to inboxes of RSS readers and email accounts automatically. This means that they do not have to check “yet another website”, but simply have to check something that they already check regularly (their inbox).

Moderation

  • As the discussion forum will be a system officially approved by the school, it will need to be moderated to keep content politically correct and appropriate.
  • Moderation can be automated, such as through the filtration of unwanted words, e.g. profanity. This is easy to implement and does not need human presence; however, it is often ineffective alone. Furthermore, lists are often too restrictive, adding words that may be politically-correct in the correct context.
  • Moderation can take place in the form of designated administrators, such as responsible students (like SRC/Prefects). However, such a system is prone to bias and human corruption.

2009-10-20 Questions and Answers

posted Oct 19, 2009, 11:35 PM by Unknown user   [ updated Oct 20, 2009, 9:21 PM by Eddie Woo ]

I asked Mr. Woo some questions about the course today (in the morning class and then in period 1). Here are the answers (as I remember them), because:
  • I find that I absorb new stuff easier if I have to describe them;
  • I need some practise with HSC keywords;
  • Some of you other 10IPT'ers might find this useful, and we want to come first, right? :D
If you find anything wrong with what I have typed here, please let me know. We're all in this together.

1. Identify the purpose of packet switching, and explain how packet switching works. (Communications)

Packet switching exists to allow a message to be transmitted in a more efficient way in a network where there are multiple possible routes from the sender to the recipient. It works by splitting a data transmission into groups of bytes called packets. The sizes of these packets are determined by protocol (e.g. 4 bytes per packet). These packets are sent simultaneously, each on a separate route to the recipient. They will arrive in an arbitrary (random) order, but can be utilised by the recipient in the correct order by identification features in the packets.

2. Clarify the possible meanings of the terms "relationship", "entity", and "attribute". Propose the most suitable term to refer to things they represent (in a HSC exam), and justify your response. (Databases)

Much like how the term friendship refers to the way in which two friends are connected, a relationship describes the way in which one table in a relational database is linked to another one. They are the links between tables, and can be one-to-many, many-to-one, one-to-one, or many-to-many, as described by the Board of Studies IPT syllabus. While the definition given by the Jacaranda1 textbook ("a relationship is a table or file created by joining together data from different databases") can be considered valid in an abstract way (in that a relationship is required for two tables to share data), it appears to be conventional to refer to relationships as the links between tables, and not the tables2 involved.

An entity is a specific tangible person, place, or thing. Databases store data about entities, such as students in classrooms, classrooms in schools, cities in states, states in countries, products sold by a department store, etc. The Heinemann3 textbook implies that an entity could also be the name given to a table that contains data about a specific type of entity (e.g. a table about students could be termed the "Students entity"), a notion that is supported by the syllabus' intepretation of tables as "implementations of entities". It appears that this convention is occasionally followed by the Board of Studies4.

An attribute describes a known feature of entities about which a relational database stores data. This notion is supported by both the Heinemann3 textbook ("attributes are the same as fields in a flat file database") and Wikipedia2; however, the Jacaranda1 textbook proposes that "an attribute is a field copied from a parent database", implying that a field in a relational database is only an attribute if it contains data from another table. Nevertheless, the Jacaranda textbook adds that "attributes are the columns in a relationship table shown in list view", which is implicit of support to the definition of an attribute as being any field in a table in a relational database.

In a HSC exam, it is most suitable to refer to tables as tables, links as links (or relationships if the questions is about one-to-one, one-to-many, many-to-one, or many-to-many relationships), and fields as fields, due to the ambiguity from different textbooks. Referring to fields with data from other tables as attributes, however, is ideal, as no definition of attributes presented by any of the textbooks would contradict this.

3. Propose which network topologies can utilise collision detection, and justify your response.

All major network topologies (star, bus, ring, point-to-point, hierarchical, and mesh) can use collision detection, because collisions will still occur regardless of efforts to minimise their abundancy (e.g. protocols such as token rings in ring networks). However, as collisions are significantly more common on bus networks, it is bus networks that use them the most.

4. Describe intelligent agents.

Intelligent agents are decision support systems that make decisions and implement the best decision in each case autonomously, rather than proposing an array of possible decisions to the end user. An example of an intelligent agent is a search engine that displays what it believes to be the most relevant result. Intelligent agents are still a form of decision support because humans ultimately still have the power to prevent the intelligent agent from implementing decisions.

Some miscellany:
  • When questions in Decision Support Systems part of Section III talk about "neural networks", e.g. "describe how a neural network works", they are always talking about artificial neural networks.
  • If there are a bunch of computers pointing to and from a circle at the center of a network, do not assume that the circle represents the central server of a star network - it may be representing the ring of a ring network or, indeed, the network itself
  • A primary key is a key field; a foreign key is a primary key when it's an attribute in another table
  • Normalise databases by drawing a schema
  • "Mobile technology" refers to mobile phones
  • Certainty factor (CF) is only for probability
  • If a question asks you to draw relationships between items, don't talk about items in isolation
  • A byte can store:
    • One ASCII character
    • Any number from 0 to 255
  • Try to answer S&E issues questions from differing perspectives in terms of people and groups of people
  • Date is stored as the "date" data type (thank you Jason and Komal)
  • Decision tree going down a page is fine

1 Information Processes and Technology: HSC Course (Ware, Cheleski & Chivers, 2001), page 48
2 Wikipedia proposes that these tables should be called relations, not relationships
3 Information Processes and Technology: HSC Course (G. K. Powers, 2000), page 44
4 A database schema is said to consist of "relationships, entities and attributes" according to the answers to question 4 of the 2001 Board of Studies paper; this same idea is presented by the syllabus

2009-09-15 QBE, Google Maps crowdsourcing

posted Sep 18, 2009, 5:16 PM by Unknown user   [ updated Sep 18, 2009, 6:17 PM by Eddie Woo ]


Query By Example (QBE)

A query is an operation to match criteria. QBE is a competitor/complement to SQL from the 1970s - while QBE and SQL compete as methods of query, QBE relies on SQL to operate.

Field Example
Due date

Price

Importer

Manufacturer
Spain
Ingredient
Olives

This QBE form is sent to a QBE Parsing Enegine (parsing means to interpret, and an engine is something that converts input to output) for analysing and ultimately the construction of an SQL query. In this way, QBE is effectively a shell for SQL.

Advantages of QBE include:
  • Less time needed to make a query
  • Less technical training required for use

Google Maps crowdsourcing

For further details, see IPT: Live Traffic Data

Crowdsourcing is the practice of relying on the end users of an information system to provide content – usually each user provides a small portion. It is a form of collaboration, and is generally beneficial to all who are involved. The compiled content is usually provided back to the contributors. A familiar mainstream example of Crowdsourcing is the system of collaborative editing used on the website Wikipedia. It is a play on the word outsourcing.

Hardware used in this system include:
  • Google web servers
  • GPS devices
  • Visual display units
  • Satellites
  • Transponders
  • Cabling
  • Base towers for mobile phone cell networks
  • Wireless network interface cards
People in the environment include vehicle drivers who send data about the traffic but do not use the content provided by the system. Participants include system developers, maintainers, engineers, and employees of Google corporation. End users include everyone who uses the system, using computers and using GPS devices.

Data flow diagram:


The resulting data could be rendered inaccurate under these circumstances:
  • A car is intentionally moving slowly on a road, and it is the only car with the Crowdsourcing to Google Maps capacity on that road
  • Google intentionally corrupts the data (e.g. under pressure from governments)
  • Technical problems / faulty technology  – miscalculation, incorrect use of units, incorrectly calibrated clocks (and hence incorrect timestamps for calculation)
  • Miscommunication – traffic is colour-coded, but what criteria is used to colour-code is questionable, as it may cause problems if colour-coding is relative, and two very different roads are being compared.
  • Confusion with satellite imagery (colour-coding)
Social issues and ethical issues raised by this system include:
  • Privacy. Although Google insists that data is largely stripped of any personally identifiable data, whether or not they fulfil this promise is unknown to end users and people in the environment. It is an ethical issue if Google stores personally identifiable data without consent.
  • Accuracy. As aforementioned, there are many ways in which the data can become biased or inaccurate. If the data is indeed inaccurate/biased, it is unreliable; however, end users may be unaware and continue to trust the system. This may raise social problems – for example, an ambulance may rely on the system to find the quickest route to an emergency, but then discover that the system was faulty, and ultimately fail to arrive to the emergency on time.
  • Cost. This system requires a lot equipment and software, and relies heavily on infrastructure. Google insists that it will provide the service for free, but whether or not they will be able to continue to do so under a changing economy is a social issue.
  • Public safety. Terrorists could easily use this system to target roads as part of a strategy to cause mass destruction. For example, they could target congested roads to maximise the amount of casualties, or use these roads to limit the speed of which emergency vehicles can access an accident site. This is an ethical issue.
  • Road safety. This system could easily encourage people to look at their GPS screen while driving, and hence cause accidents as a result of vehicle drivers failing to look at the road.
  • Changing nature of work. Many radio stations currently provide a similar service in the form of “traffic updates”. These people may find their jobs endangered.

2009-09-10 League tables questions

posted Sep 9, 2009, 9:30 PM by Unknown user   [ updated Sep 14, 2009, 3:49 PM by Eddie Woo ]

A league table is a table that ranks organisations or groups in order of their achievements.

Reasons for:

  • Changing nature of work - league tables provide feedback for schools, allowing a school to check if it is meeting standards and whether or not the school needs to work harder
  • Decision support - league tables provide feedback to the government as to which schools need more funding/support (perhaps down to individual faculties e.g. English, Maths)
  • Reward - league tables provide well-deserved recognition of outstanding schools
  • Decision support - league tables can help students decide which school they should attend
  • Transparency of data - league tables may defend against corruption by exposing schools which are not meeting standards

Reasons against:

  • Data reliability - league tables may not be an accurate reflection of a school's quality, as the values (weightings) of criteria are subjective
  • Data reliability - if the tests used to obtain the league table data are poorly-written or do not test for the right course outcomes, the final product would also be poor
  • Data completeness - a league table is based only on basic skills such as numeracy, literacy; it would be unable to take into account other factors
  • Privacy - the privacy of individuals may be at risk depending on how the data for league tables is collected (particularly if the names of individual students are recorded)
  • Public safety - as it is apparent that there is a great deal of strong public opinion against league tables, the publication of league tables could potentially results in a violent lashback from the affected communities
  • Cost - it may be extremely costly to produce the league tables, and it can be argued that money spent on league tables could be better spent elsewhere
  • Diversity of education - one of the strongest arguments against league tables is that they encourage teachers to only teach students how to do well in basic skills tests, hence limiting their ability in other fields and producing a generation with a very narrow set of skills
  • Unnecessary - it is maintained that "reliable school performance information is already available"
Emerging trends in information technology:
  • Ease of collection - the data is very easily collected using modern computer technology, the use of which is an emerging trend; this is in contrast with traditional methods such as paper record-keeping
  • Ease of analysis - data collected with computer technology can easily be manipulated and analysed; rank information can be generated in mere seconds
  • Collection of personally-identifiable information - there is an increasing trend for organisations, companies, and groups to gather personally-identifiable information; hence, people affected by the publication of league tables are justifiably paranoid that personally-identifiable data will be collected
Propose an argument for/against:
 

2009-09-03 Trial exam review

posted Sep 2, 2009, 4:52 PM by Unknown user   [ updated Sep 14, 2009, 3:30 PM by Eddie Woo ]

Question 6

The best option is query by example because it saves time by removing the need to type the SQL query with full, valid syntax.

(Exam technique: N/A, theory was not taught)

Question 21

The interactions part of (b) was too general, as it was simply a generic interaction of how the components of the system interact in any given system. It should have been related specifically to the system in question, e.g. "the purpose is achieved by means of the participants collecting statistics and photographs about the runners, using the timing devices and camera, and then uploading them to the website"; examples are good, even if the HSC keyword is "describe".

Subquestion (c) was asking for judging the success of the system as a replacement of the old system, as it had labelled it as "the new information system".  Criteria should therefore have reflected the advantages of a computer-based system; e.g. wider audience, saved paper, increased convenience.

(Exam technique: theory should be tailored to the scenario in question, new systems should be judged as replacements, not products)

Question 22

The calculation in (a)(i) was perfect, except I had foolishly forgotten to convert to kilobytes (despite naming the calculation "size in kilobytes of the bitmap"). To convert to kilobytes, divide by 1024.

(Exam technique: avoid silly mistakes)

Question 23

The data flow diagram in (b)(i) could be improved by having two separate entities instead of "Mobile phone or computer" as a single "entity". Also, despite asking for a data flow diagram of "when a user sends a single tweet", the question had mentioned followers; mentioning them would have earned the extra marks.

(b)(ii) had called for a "critical analysis"; to critically analyse something, one must discuss the relationships between items. Therefore, more marks could have been gained by discussing how issues relate to each other, e.g. virtual communities act as catalysts for globalisation, emerging trends may counter the insincerity of messages (e.g. using videos), censorship may destroy or counter the formation of virtual communities.

Part (c) was done as if the system was the laptops and the network combined; however, the question had actually subtly excluded the laptops. More care should have been taken during the reading of the question. Also, points should not have been phrased as rhetorical questions.

(Exam technique: all the given information should be used if possible, critical analysis requires discussion of relationships between items, more attention to what the question is asking for)

Question 24

In (a)(i), the layout of the decision tree was more like that of a flowchart; a decision tree has questions on the side, and options (not necessary boolean Y/N options) on the tree itself. Also, assumptions were made as to what the system would check for; this was counter-productive.

In (a)(iii), it would have been much more rewarding to mention more types of analysis. Only calculating was mentioned. Other types include sorting, searching, labelling, graphically representing, identifying patterns and trends. Mr Woo's specimen answers also suggest comparison as a type of analysis.

Part (b) included enough issues, but fundamentally did not answer the question properly in terms of HSC key words - a discussion needs to include both points of view, and should not be an argument for a certain point of view.

(Exam technique: revise diagrams properly, use the arsenal of lists logically instead of instinctively, and remember that a discussion must be two-sided)

Question 26

In part (c), I used the phrase "it seemed to make sense to". This is too informal; a better phrase would be "it was an apparent, sensible and logical decision to", or an equivalent. Part (e) could have been improved by listing similarities as well as differences, because the question had clearly stated compare and contrast.

(Exam technique: compare and contrast is two-sided)

General exam technique advice

  • Spelling - incorrectly spelt words look very bad
  • More detail is often required - e.g. different perspectives, additional points
  • DNAQ stands for Did Not Answer Question; always answer the question given
  • Information systems that replace manual systems must be better if to be compared using criteria
  • The HSC Keyword Analyse requires that relationships between individual items are established
  • Social and ethical issues are two-sided (positive and negative effects)
  • Do not phrase issues as questions - in the worst-case scenario, you can say "it is questionable"
  • Revise decision tree layout
  • Don't extrapolate/assume - e.g. the Q24 decision tree (where many people made up conditions)
  • Don't be colloquial - it is a formal assessment
  • Extra social & ethical issues
    • Cost
    • Reliability (CB vs NCB)

2009-08-18 Past paper - 2008

posted Aug 21, 2009, 5:34 PM by Unknown user   [ updated Aug 21, 2009, 5:35 PM by Eddie Woo ]

See the file attached.

2009-08-11 Pre-trials throughts

posted Aug 11, 2009, 2:54 AM by Unknown user   [ updated Aug 11, 2009, 4:20 AM by Eddie Woo ]

  • If someone makes a bad decision on the advice of the decision support system, it is the person's fault?
    • What is the main point of a decision support system? I would say it is to support decisions, not make them.
      • Jacaranda textbook, page 185 - "Note that decision support systems assist in decision making, but they do not make the final decision, this is up to the person using the system"
      • Jacaranda textbook, page 186 - "[Decision support systems] support the decision makers rather than try to replace them. The person using the system makes the final decision. The nature of many problems means that some human judgement is still required in their solution."
      • This appears to be contradicted by page 201 - "Can DSSs be designed to take the extra step and make judgements for us? They can if they are using artificial intelligence."
    • However, Jacaranda textbook, page 186 lists the spelling/grammar checks in word processors as a drill-down analysis because it "shows working" (as in mathematics)
      • If a student handed in a written work that was corrected automatically using this sort of DSS, and it had mistakes like, "eye used my pea sea", the student would be blamed despite the fact that the student was incorrectly advised by the DSS.
      • Stock market example - the stockbroker wouldn't be able to blame his/her DSS for bad stock decisions as the stock market is complex to the level of artificial neural networks
    • Question is answered on page 224 - "The end user is ultimately responsible for the decisions that he or she makes with or without the help of a DSS"
  • Jacaranda textbook, page 185 - "A common abbreviation for decision support system is DSS." - therefore, "DSS" should be an acceptable abbreviation in exams?
  • Drill-down analysis is important because it allows the end user to made an evaluation of the advice provided by a DSS based on how the advice was constructed.
  • If we get a question that requires us to draw a user interface or expert system shell, do we need to use a ruler?
  • How are GDSSs decision support systems? The examples on page 223 aren't GDSSs, they are collaboration techniques.

2009-04-08 Knowledge Engineering

posted Aug 3, 2009, 3:39 PM by Unknown user   [ updated Aug 6, 2009, 12:26 AM by Eddie Woo ]

Engineers are...

  • Civic engineers
  • Chemistry engineers
  • Mechanical engineers
  • Molecular engineers
  • Aerospace engineers
An engineer is a person who understands the principles of a certain field and builds things in accordance with those principles.

Knowledge engineering...

  1. Acquisition, i.e. collecting - many kinds (procedural/instructive, declarative/facts); many forms (text, numbers, audio, images, video); many sources (experts and experience/mistakes)
  2. Organising, e.g. organising by pixels would be problematic in a system that tries to identify art; some sort of analysis must be done (e.g. finding similar matches for colours, basic shapes, proportions, etc)
  3. Validation - to verify and validate knowledge, making sure that it is correct and up-to-date (similar to the information process of processing)
  4. Design - to create rules of inference, understanding cause & effect

Answers to questions about knowledge engineering

For further details, see IPT: Knowledge Engineering
  1. Knowledge acquisition is complicated because of the diversity of knowledge. There are many kinds of knowledge (e.g. procedural or declarative), many data types (text, numbers, audio, iamge, video), and many sources (which may contradict each other). Furthermore, this knowledge is often very abstract, and difficult to express in a way that machines will be able to interpret and make decisions based upon.
  2. Knowledge validation is important for the obvious reason of making sure that the knowledge is correct; a decision support system that using incorrect or outdated information will produce incorrect or outdated results.

2009-07-29 DSS Spreadsheet

posted Jul 29, 2009, 6:05 AM by Unknown user   [ updated Jul 29, 2009, 6:06 AM by Eddie Woo ]

Prototype attached.

2009-06-16 Expert Systems and Neural Networks

posted Jun 15, 2009, 3:30 PM by Unknown user   [ updated Jul 17, 2009, 8:27 AM by Eddie Woo ]

How Expert Systems Work

Expert systems have a knowledge base of rules (domain-specific heuristics) that govern situations and hence inform decisions.

  • Domain-specific: about one specific area
  • Heuristics: stategies that always seem to work

Expert systems are designed to imitate experts. Why would you mimic human experts with a computer system when there are real humans? Well...

Advantages:

  • Consistency can be an advantage
    • Repetitive situations can be responded to with consistency, regularity, predictability
    • Repetitive situations do not bore them or waste their presence
  • Decisions are clear as they come from pre-programmed rules and logic instead of intuition
  • Greater processing speed with simple data
  • Handle more data to superhuman levels of memory
  • Data can be stored indefinitely and can be accessed instantly
  • Humans can die, and humans are considered less disposable than machinery
  • Humans may be incapable in emotional stress
  • Greater availability in time and space + multiple users
  • Can replace original human expert(s) after their death(s)
  • Diligence (expert systems do not forget factors)
  • Persistence (machines do not sleep, they sleep mode!)

Disadvantages:

  • 'Dumb' as in expert systems lack common sense and intuition (Google Maps)
  • Boring-ness (cannot make creative responses)
  • Humans can think faster in complex, abstract situations
  • Humans learn
  • Machinery requires an electrical supply
  • Interpretation/inaccuracy (expert systems are limited by the situation data input - ambiguous)
  • Stubborn (does not automatically adjust or adapt to changes)

How Neural Networks Work

To be continued...

1-10 of 49