The due date is 20/11/2024, Wednesday Week 11 Trimester 1.
So there’s a handful of things to do for my literature review before the due date. And oh boy, is it due soon. It feels far away but it’s actually due in two weeks and I have two weekends to do it.
The good news, however, is that it doesn’t look that bad.
A pretty annoying thing is my literature review grade is reliant on avoiding internet source. Since this is quite a new topic, a lot of the useful information is on the internet. Haha, it even says “not reviewing any recent work (within, for example, the last 5 years”. It clarifies it depends on the field. It is funny that the highest grade criteria “anticipated to be seen in a journal review paper”. Ah well.
Here are the requirements:
- Literature review (12-15 pages)
- Prerequisite information
- Existing literature and what research is missing
- Problem statement informed by gaps, and aims and outcomes? weird
- Project plan (5-7 pages)
- Break the project into components and define the function of each component. Justify the time allocations
- Timeline - goals of Thesis A and Thesis B
- Required training and upskilling
- Project dependent preparations (1-2 pages)
- Evidence of some upskilling (language, software, tools, specific knowledge required)
- Preliminary results
Then some grades for formatting, English, figures and citations. Note that part of the project plan requirements is “the student has acquired all the skills”. So I’m going to need a really tiny demo using MLIR probably… Maybe I can make an incredibly lightweight version of Lingo DB that just lets you import data and select?
Okay, so, cool. Basically I can take all those page counts and increase them by 67%. A good page count to aim for is probably about 35-38.
So, topics I want to go through in the literature review
- Literature review
- Compiler knowledge: What is JIT, what alternatives are there, what JIT compilers are there
- How do popular on-disk databases prepare queries
- How do most in-memory databases prepare queries
- What are the bleeding edge ones like?
- Project Plan
- This really depends on feedback from my seminar, which I haven’t done yet…
- Project dependent preparations
- Benchmarking - how impactful is JIT here anyways?
- Very small demo of MLIR
My timeline for doing this in a reasonable time… So if this thing is 8000 or so words, I think writing about 4000 for the literature review will go quite fast. That’s maybe ten hours. Project plan is really only 1-2 hours, and project dependent preparations is immensely long. The literature review and project plans are worth the most, but I can’t concretely start the project plan until I know the feedback of my seminar. Which is fine, I should be able to get that scheduled this week. Project-dependent preparations is probably 20 hours, and weirdly doesn’t depend on the seminar too much.
The best bang for buck is doing the literature review since it’s quite brainless to work on. Read. Write what it says. Fantastic. Project plan is very low effort, but needs clarification so I’m a bit blocked. Project dependent preparations will be a massive time sink for how much actually shows out of it.
Now let’s take all of these numbers and add fat, which should give me a good value for how much effort this is going to be. 10+2+20=32, double it to 64 hours.
Great. Let’s dig into it.