"You should start writing a blog!" They say.

I don't think this blog is going to last very long once things get hasty during the year. Alas, it is something to began with.

This research project diary is inspired by a student's blog I found on the web sometime after I was in uni. Interesting stuff. It is then I realised that I did not find any blogs on people doing their final year project and I thought to myself:

"Wow, isn't this a great opportunity to start writing a blog! Maybe it'll help people who is looking into what the part 4 project is about!"

I don't remember myself being so altruistic.

I doubt this blog is going to be as interesting or as technical as the students whom I had read. So for now I shall designate this place as a meditative space to help me cope with the LITERATURE REVIEW that's going to haymaker me into the south wall. 

Perhaps that's it, keeping a blog and talking about what I have read will help me digest the papers I am going to read. Or, it is more likely that I am going to be soo caught up in reading them that I won't have time to write this blog. Yeah, gonna look into this colossal task in more detail once uni finally began.

Part 4 projects are a whole year's worth of investment, so it is incredibly intimidating. When we read through the projects on offer from the faculty, we'd have to be very sure that the project is feasible for us. While I am sure the supervisors have every consideration in the project being doable by the students, the projects are still are vastly different in scope.

• First, there are the industry-sponsored ones. Those are usually the supervisor's own efforts in connecting industry interest to our academic endeavours. This kind of project is often very practical in professional conduct and the topic is often very "onto" the current scene. However, this comes with great responsibility.
• We avoided any collaboration or continuation of previous years' work. Collaboration has similar responsibilities as the above and extending someone else's work feels underwhelming for our final year project. Think about it: It is a project left entirely to the devices of two students + their superior. If anything, we'd like it to be a trial of character — something accomplished by us, authentically. It would seem that to me and my partner, we'd like to keep things to our own (like the loner COMPSYS students we are).
• We also avoided "pure-research" projects. These projects often have few deliverable or very non-consequential ones. The primary activity is just to research a topic critically and present our findings. Very boring. Even for a research project.
• And then there are just projects that try to accomplish way too many objectives. First and foremost, our final year project is a research project. While it is natural to have some deliverable at the end of it, there are a certain few that bites off more than one can chew. That is speaking for me though.
• Also, we dodged every machine learning related project. We joked that our machine learning lecturer in our early years gave us a run for it!

In the end, while confining to our interest in low-level digital system development and embedded systems, we managed to pick out the following for our top 5 allocations. Did I mention that we had to contest with other students?

1. #42 Programmable Co-Processor for Security-Enhanced Software Implementation on Embedded Systems
2. #77 Dynamic Cryptosystem: the Case of Robust Security on Reconfigurable Platforms
3. #41 Application-Specific Processor for Embedded AI Applications
4. #32 AR/VR Perceptual Training (collaboration)
5. #83 Real-time defect detection in manufactory lines

Now the top two is the ones that we truly wished to do. The supervisors are really desirable (nice and most importantly: they're good at teaching!). #42 and #77 and both very in-depth investigations into developing a system to deal with security — a very catchy topic. #42 is the independent design of a co-processor, sweet. And #77 is an investigation into cryptography down in the hardware level, intriguing. 

I even got into a panic mindset trying to outsmart the supervisors to get allocated our top two allocations. Scouting out other students, there were a few who liked to do #42 and #77 as well. When I imagined this battle in my head, I had thought about whether it is okay to not include other popular projects? What if people thought we were okay with doing our latter three picks, so they'd allocate us that? Should we have made our selection more acute so that it seems like we desperately set our to lock us into the top two picks? For example, if our latter three picks are also the popular ones with many people placing it as their top picks, would they be more inclined to give us our top pick just to ensure that we can at the very least be allocated with a project on our wish-list? So that we don't fall in the trap of: as long as everyone is OK, it'd be fine?

Well, that's just a paranoid rant. 

Now we did mention that we avoided every machine learning project, and #41 is related to extending an instruction set architecture to better accommodate DNN algorithms. It is alright, the topic has many industry appeals and is a very important endeavour in trying to optimize performance — it is what we do. However, on the contrary, it is also very boring compared to #42 and #77. We had set out to try and avoid some really research-heavy projects and the sound of simply dealing with looking into DNN is giving me a fat head...

We got allocated project #41.

翠花！上酸菜！这汤太TM苦了！