The Autopsy Report

Mein Gott... it's... dead...

^ 'Voluntary' ambient music for your reading experience.^[1]

Hi guys!

Just to start off, this isn't your usual autopsy report like "oh no, this guy died because someone shot him (?)", instead, this is inspired by the Republican party's report after the 2012 election. In other words, I'm going to talk about what went right, what went wrong, and what I learnt. :)

Note that half of this was written during a 3 hour ride through the 'scenic' Malaysian landscape, and the other half was written while I was sick. So take what I say with a pinch of salt.

The good stuff

I think first and foremost, the goals and intentions of the Cooking project were able to address a very specific issue for people in my position, i.e. students. Paper isn't cheap, and so if you want to print a bunch of exams, you should print them in booklet form. This saves you money, and saves paper for someone else to use. While there are many, many, many options that offer this, they are often polluted by advertisements or paywalls! This leads one to start to worry about how much they can trust these websites. For most people, this can be very off-putting, and thus they end up concluding that the whole procedure simply isn't worth their time.

In addition to this, those who use Latex may be familiar with the many tools that accompany it. One that I found very useful was pdfbook2, which is a simple Python program that converts pdfs to booklets with added but not essential specifications. Perfect, right...? Well, I would say it has a few barriers of entry.^[2]

For example,

• It is CLI (command line interface) only. The terminal is a rather unfamiliar concept to most people, making it difficult to use without prior experience.
• You have to type in file names, individually. This can be very cumbersome if you have a lot of files you want to convert!
• The program itself is packages with a lot of unnecessary stuff! It takes a lot of digging to just get what you need.

Now this is not to say the program is bad. In fact, it is quite efficient and works perfectly for what it is designed for: as a component of bigger and more complex programs.

This is why I made Cooking, along with a little inspiration from my favourite Bento :)

In the end, Cooking was quite a success. You simply needed to gather your files, oppen the website, drop them in/submit them and voila! Booklets, with no intrusive ads, and the code being open sourced so you could look under the hood. I'd say this project was a strong improvement on its predecessors.

The bad stuff

Unfortunately, there are definitely some flaws with Cooking. Most glaringly of all, it's a Flask app. That means the website is connected to an external computer, in this case one of Google's, which it uses to process the pdfs into booklets. Good, right? Not really. It's not like turning a pdf into a booklet is a monumental, Herculean task. It can definitely be done on your computer. And thus, this link to Google is a critical redundancy, for a variety of reasons.

• You need to be connceted to the internet to use Cooking.^[3]
• In the event Google goes down (unlikely, but let me cook), Cooking goes down too.
• As the website owner, I actually have to pay for people using Google's resources. :(
• It's definitely less environmentally friendly to wake up Google's computers as opposed to utilising the user's computer.

Now, there are some ways to rectify the whole Google issue. As Cooking is a Flask app, I can use Docker to create a portable image that can be run by anyone! This essentially packages everything needed to run Cooking on a computer into a reasonable (≈500MB) sized file that can be shared, i.e. through Docker Hub.

However, this introduces a new problem. What even is Docker? How do I use it? For the average person using Cooking, the whole Docker setup is a major barrier of accessibility. This cuts away at one of the key philosophies behind this project, that is;

"You shouldn't have to use Arch, btw, to use Cooking."

This is what makes Cooking a bad project. It is dependant on third parties. When we consider how others use our code, its crucial to consider accessibility and agency.

These things should not be mutually exclusive! If you want to be free of third parties like Google, you should not have to learn how to use Docker. Likewise, if you just want an easy to use tool, you should not have to trust that third parties are managing your data with care and transparency.

This is the critical flaw at the heart of the Cooking project, and one that I should have considered more thoroughly. In the process of learning all these new things, i.e. making a Flask app, using Google's Cloud Run, using DOcker, I had forgotten to consider accessibility. The whole project started because it wasn't reasonable to have to install Python, TeX-Live, etc, for booklets. Instead, I've ended up having to use yet another third party software (Docker) for local usage.

What I learnt

Clearly, the whole Cooking project has been a mix of good and bad. So, what did I get out of it?

Well, a few surface-level observations can be made,

• I learnt how to make a Flask app.
• I learnt how to use Docker.
• I learnt how to use Google's artifact registry.
• I learnt how to use Google's Cloud Run.
• I learnt what Cygwin is (yay).

However, if I think a little harder, I can say that a lot more was learnt as well.

• I learnt the importance of planning. As soon as I had thought of the idea for Cooking, I knew I wanted to use Python for the file processing. I mistakenly believed this to be easier. This decision is the root of all my future problems. Had I spent the time to think how I would approach this project, I would have considered using Javascript, as Choyu pointed out. As one would expect, Javascript is perfect for this situation. Instead of needing Google or Docker, all you need is a web browser. Easy!
• I learnt the necessity of reassessing goals and purposes of a project. As I worked on Cooking, I realised how easy it was to get carried away on trivialities, and to go on unproductive side quests. Keeping the main goals of the project, i.e. making life easier for students, allowed me to refocus and continue work. It also allowed me to overcome the impending sense of technical debt, as I was able to move Cooking in a direction that ultimately brought benefits to people, despite its many flaws.

In conclusion

That certainly was a lot of talk. I hope you were able to sit through all that. I find it useful after working on an idea like Cooking to take the time to reflect on what actually happened, and what I gained from it. This is that reflection. This also marks the official end to the Cooking project. I might make small updates just to see if I can actually make Cooking a 100% offline thing, but it's unlikely. I'd say its current state is more than enough.

I hope that you learnt something (?) from this rather long-winded rant. I also hope you enjoyed Sorgemarsj :)

Footnotes

• ^[1] Sorgemarsj over Rikard Nordraak (Funeral March in Memory of Rikard Nordraak), EG 107 (version for wind ensemble)

• ^[2] Not to say it isn't great for what it's designed for. I doubt it was intended as a standalone product! Instead, it was probably made so that other programs bundled with TeX-Live can use it.

• ^[3] This is an inherent flaw in most website, except for ones where all the code is stored in the HTML file. Then, all you need to do is download it once and then you can use it forever! This is what the main page uses as of now, thanks to Choyu's code! (Yes, that means you can download the html file and use it without internet.)(On second thoughts... no. Unfortunately you still need internet for this page to work. At least its just downloading some js libraries!)