INFO20003: Database Systems

10% MST, 3x10% assignments, 60% exam

I really liked Renata as a lecturer. She gives us easy assignments (Based on the bell curves she releases), but otherwise I like the way she runs the course and content and of course teaches.

Doing well at the assignments will lead you much more comfortably into the exam, as there is a lot of content. Overall I'd say this subject is wide but shallow, so if you were taking more memory subjects, you would kill this subject as long as you had some programming aptitude and basic maths, it's highly memory based for a large percentage of the marks.

Only difficulty I had with the subject was that Conceptual modelling needed more examples from tutors/lecturers of the thought process of tackling those problems so everyone comes up with slightly different answers, I never had the knack for getting the same answer they did (or a better one). It was a subjective topic because interpretation and design depends on the person, but some designs are better than others. Otherwise the other 2 assignments were easier to full mark as they had right or wrong answers.

This subject could be dry to a lot of students, but I think Renata has made the subject pretty good, apparently she came in 2016 Sem 2. It really makes you appreciate how the fuck anything on the internet even works without massive amounts of lag. It covers interesting topics and has a nice balance of theoretical algorithms parts and practical parts.

Also recommend getting David Eccles as a tutor as he writes the exam, might be scary but you'll thank me later.

Database Modelling
SQL
Query Processing and Optimisation
Concurrency, Transactions, Distributed Databases, Data warehousing
Database Administration
NoSQL introduction
Also would like to point out, doing both this subject and SWEN20003 in the same semester is definitely manageable.

Yes

Renata Borovica-Gajic (primary) , David Eccles (minor)

Yes, 1 sample exam but a decent amount of practice questions in general

4 Out of 5

None

2x1 hour lectures, 2 hour tute and lab

2018 Semester 2

H1

The projects were a little different than described in the 2015 handbook, apparently we were unable to do team projects due to some technical difficulties. We had this instead:

10% Mid-Semester Test (week 7)*
2% Assessed Workshop (creating a simple database-driven website with PHP, week 7)
14% Assignment 1 (creating a large-scale data model, due week 10)
14% Assignment 2 (building a bunch of SQL queries of varying complexity, due week 12)
60% Exam in exam period

From the looks of the 2016 handbook entry, there are still no team projects, but the assessment structure has been altered somewhat.

* The in-lecture MST was aborted because Mitchell didn't bring enough test papers for everyone. It was replaced with a take-home test due about one week later.

Database Systems is on the road for the Computing and Software Systems major, as it is a prerequisite for the capstone subject COMP30022 IT Project. It might also be considered by anyone interested in information systems. However, based on my experience in 2015, I would recommend avoiding this subject for a few years if possible, until the coordinators figure out how to run a subject coherently.

For the most part, there was less than zero structure in this subject's delivery. Expectations were poorly communicated, and constantly changing. Lectures and workshops were a waste of time. This meant that it was almost impossible to study independently, because nobody knew what we were meant to be studying. Projects were deliberately ambiguous (to 'simulate the real world', despite how in the real world you get to ask for clarification from clients; they don't say 'I can't tell you how to do the project' and force you to make assumptions about what they want), and the MST (the one that was cancelled) and exam were incredibly unbalanced (requiring highly specific knowledge and also containing large modelling tasks better assessed through a longer-term assignment).

Above all, INFO20003 was utterly intellectually dissatisfying; a real shame considering the potential of databases as a genuinely interesting topic of study. The theory of relational databases is inherently mathematically precise and completely logical. The systems that implement that theory are filled with clever algorithms and data structures worth studying. Data modelling in itself may not be a black-and-white problem-solving discipline, but there are logical approaches and sound principles that can be used to construct and evaluate quality data models. Instead, INFO20003 abandoned this rich content in favour of a shallow 'learn by example' approach that skipped almost all of the interesting theory underying the topic of database systems.

I'm not sure if it would have been possible, but I wish I had looked into taking INFO90002 Database Systems & Information Modelling in its place. The coordinator for this graduate-level version is Greg Wadley, who also gave 4 of the INFO20003 lectures. His lectures stood out as a shining example of how good this subject could have been.

If you're stuck taking this subject, and it hasn't had its deep fundamental issues addressed by the time you take it, my advice is to attempt to take your learning into your own hands as much as possible.
- For data modelling, focus on nailing the underlying principles of the relational model. These principles were never mentioned (whether the lecturer was aware they existed or not I don't know) but they were the basis for everything that we learned and did, and once you understand them, data modelling is only as hard as learning the notation and applying your common sense to interpret case studies.
- For learning SQL, there's no shortage of online tutorials, from W3Schools to Codecademy's short course and I'm sure there are plenty of MOOCs and other resources I never tried.
- For the rest of the subject (the conceptual stuff), use your usual strategies (making notes, flashcards, or practice problems, whatever) and try to dig a little deeper than what's mentioned in lectures because the exams sure tested beyond the level discussed.In short, find a good learning resource (see textbook section), because if you're not relying on lectures, it should be entirely possible to survive this subject's endless frustrations.

Yep, with screen capture. Slides were also provided on LMS. Some whiteboard work was carried out during lectures and Mitchell swore you were missing out by not being there in person, but you really weren't.

The majority of lectures were delivered by the course coordinator Mitchell Harrop (the first 7 weeks, and also week 12 for revision). Mitchell's topics included introductory concepts, 3 weeks on data modelling, and also SQL, data model quality, and using databases with websites. Mitchell is a relatively new lecturer and his inexperience really showed.

Then there were 4 lectures on miscellaneous topics by Greg Wadley, including normalisation, transactions, and select database administration topics. These lectures were put together really nicely.

There was a guest lecture on Data Warehousing (a.k.a. big databases) from Sean Maynard, the previous lecturer for this subject. I was glad he was not still the lecturer for the subject.

And finally, 3 lectures on relational algebra and query processing by Linda Stern, however this was Linda's last semester lecturing before retirement and I believe these topics are being removed from the course for future semesters. Linda's lectures were organised nicely but her explanations sometimes missed points and weren't always so satisfying.

No past exams or practice exams were available.

We were only given a small list of incomplete exam style questions, aimed to give us a vague idea of what kind of questions to expect.

There was also a large set of SQL query practice questions with sample solutions (kind of like a problem booklet in a maths subject, with about 45 questions total, which was okay since SQL was only a small part of the course).

We were also referred to 'everything in the lecture slides and workshops'. However, the questions on the exam mostly demanded more than what had been presented in lectures or required in workshops, and rewarded independent study.

1/5

We didn't really stick to one textbook but there were a few official recommended texts: Modern Database Management (Hoffer) and Database Systems Concepts (Silberschatz). Mitchell told us that one of the changes planned for future semesters was to standardise more to the Hoffer textbook, so in future years that one might be a bigger deal.

Another notable textbook was Data Modelling Essentials (Simsion), authored by an ex-UoM Database Systems lecturer who quit his job as a high-profile academic to study screenwriting at RMIT and is the author of The Rosie Project and its sequel The Rosie Effect; books which are being made into movies soon! Anyway this textbook was relevant to a small section of the course on evaluating the quality of data models.


Unfortunately there seems to be a trend of verbosity in Database Systems textbooks, which can contain a lot of text (700-1000+ pages) but not say very much at all. I personally found there was no need to waste so much time to communicate such an inherently logical topic. So I've listed some alternative recommendations here that might end up being preferable!


First up, I found a much more concise resource in Database Fundamentals, a free book from IBM. It talks about a different Database Management System than the one studied (DB2 instead of MySQL) but the first few sections cover the essentials of relational databases really nicely, and without wasting your time with pages of useless fluff. Worth reading carefully!

I found the Wikipedia articles on database systems concepts comprehensive and insightful, probably because it's such a well-established field. Generally, though, searching the web proved a good way to answer any questions that came up during study.

I'll also take this opportunity to plug my own set of notes, available on StudentVIP for $19 (I get $14.25 ). I put these together in the lead up to the exam; they're are the result of my efforts to piece together the subject into a coherent whole. I'm also open to responding to PMs if anyone is really struggling.

Two 1h lectures
One 2h workshop

2015 Semester 2

94