Learning Minecraft, surprisingly.

My work has scheduled me without asking my availability to teach a pair of students the their next level of Python classes in a newer format that uses Minecraft Education Edition. I'm not really sure if I really like it or not. There are parts of it that have extremely rough edges, some of which I think are because of the company I'm working at, and others because of Minecraft Education Edition seems to be a not-so-great way to teach. I'm worried about having to wrestle the kids away from the game that I am literally putting them right in front of. There are a lot of really mixed incentives with the program, I think. I've played around with the lessons that Microsoft has made, and they are ... not rudimentary, exactly, but they show off the constraints. Every task they have you do is in a world and it seems like it's good because it makes logical sense, but then again you are programming a little robot to do things that if you weren't constrained would be so much easier for you to go out and do it yourself.

But the curriculum I'm teaching is the more advanced stuff, for kids who have already learned these concepts that Microsoft's lessons have been teaching. And so what exactly is the point?

I say this, but really, I think this might be helpful to these particular students because the thing that I find most of my students have a hard time with is being flexable in their thinking. Even if all they did were drawing with blocks, it's still going to need them to learn to think a little differently.

The thing that's really bugging me, though, is that this uses the relatively new Python Notebook editor. The Notebook editor is just stupid.

For one, the editor is just a web browser integrated into minecraft. It's not well integrated, either. You need to click on it before it will register any inputs, so if you want to, for instance, click on a piece of your code. you will have to open the editor with the "c" key, and then click somewhere in the editor before it will actually allow you to click on your code. To run any given piece of code, you will need to open the editor, click on it, then click a second time to actually run it. If you want to see the code in action, you can theoretically press escape to close the editor, but in reality your input is being captured by the browser so you will have to click on the "X" close window button - twice. So basically, everything feels like it's barely working because it doesn't actually do what you want to do the first time.

Beyond that, there is almost zero advantage to using notebooks over a basic editor. For one thing, you cannot save your own projects. Officially, at least; the program tells you when you start a new project it will not save, but it's persisted every time in the world I've created. But even then, the actual usefulness is gimped. For one, you cannot write plain text anywhere. The only option you have is to write code. It seems like they are using the format to integrate lessons - that's how they do it in their example teaching worlds - but for the life of me I cannot find a way to write my own and include it in a world. It seems like you are supposed to be able to write rich text and insert gifs and images and even have multiple pages, but none of that functionality is exposed to the user. The only things I have found - by accident - were that you can press "a" when not in an editor and a new editor will appear somewhere, press "x" and some editor will be deleted, and sometimes pressing "z" will bring back an editor that was deleted. So basically, my only real option to teach my students is to be sitting next to them and walking through everything bit by bit.

I have been learning how to read code in larger projects. Since my programming experience has only been in school, I have not had to take on a larger codebase that I need to modify and debug. The project I am working on is a fork of someone's personal project, so code comments are minimal. So I have been slowly piecing together all the project libraries it is using and for which. I have a rough understanding of the overarching concepts, so now I am creating a flow chart of the different app pages and trying to figure out which methods are called and when.

Didn't realize that my previous comment was at the near end of the previous thread, and since I wasted far too time during work it takes more effort than you'd think to change lyrics: enjoy. It's based on '2 days into college'. Gotta vent frustrations having to deal with a frontend you barely control.

2 days into backend

I'm two days into backend, and I'm three stories behind.
There's this client, let's name them Client.
They say it works just fine.
But it doesn't really sit with me quite right.
'Cause they don't see what things are really like.
And I keep accidentally triggering a new set of functions in the middle of the night.

I write the code tired, and I write the code with bloat
And I didn't write the code that fired, but I'll just drink coffee cold.
See I don't like hacking things but must make this function as I've been told.
So I just write around and keep my head down
And hope my career unfolds.

And my seniors keep telling me that I'm coding so well
I try to trust them honestly, I find it hard to tell
If I need Copilot or I need Git
To try a lot or just a bit
To tell myself, I code the line in
"It's fine, I'll just hack it all in"

I tend to forget, my position here is quite young
In a way this career of mine has only just begun
And I've got time
I've got time

I'm two days into backend with a busy, busy mind
That feature we call Frontend is an enemy of mine.
My colleagues say I'm crazy say I took it way to far.
'Cause the frontend passed it over as it doesn't take the JSON as part.

I'm only two days into backend and my desk is now a mess.
There's just so much that I want to build that I have not build yet.
There's just so much I want to see and far too little developed.
On my mind, it runs so further away, it's easy to forget.

That to all end users it looks like I'm doing so well.
I try to see it honestly, I hack it all to well.
If I've code fully or there's something I omit.
I need to sleep after each commit.
They said "Push", I said "alright".
I think I won't, I maybe might.
I probably should just Fetch it slow.
I'll name the branch well, but coffee that I know.

The one thing that's important above everything else.
Is to learn not to push before the code is tested well.
I try believe it when I say
If it's meant to work it'll work anyway
I'll be true
I've got time.
I've got time.

That's where I'm at, to be honest.
Just two days
Two days into backend.

I built a NAS

Background

I picked up an HP Elitedesk 800 G4 Mini to use as a home server almost a year ago, and set up Home Assistant OS through Proxmox (RIP tteck). I had wanted a NAS for a long time, and bought that model because you can install 4 drives into it (2 m.2, 1 SATA 2.5" SSD, 1 m.2 drive in the network card slot with an adapter). But the entire project went on hold for a couple reasons:

• The OEM caddy for the 2.5" drive physically blocks one of the m.2 slots. Every post that I've seen where someone tries to use both involves 3D printing a new caddy, which I don't have access to.
• I didn't have any large spare SSDs available to put in the server, and SSD prices seem to have climbed back up since I last bought.
• All the adapters I found to fit an SSD into the network card slot wifi card slot required a chassis modification I didn't want to do.

Fast forward to this past January, and I stumble on a post of someone stuffing 4 SSDs into the same computer I have while managing to find an adapter that fits an SSD into the wifi card slot without chassis modification. This was a kick of motivation that made me realize the very obvious fact I don't need a NAS for bulk storage and OS drives for bulk storage, so I can move my largest capacity SSDs into the NAS and replace my OS drives with much smaller capacities. I still did need to buy one SSD, but that was better than three. And while I didn't want to mangle the chassis of the computer, I was fine with mangling the SSD caddy to make everything fit.

The Build

This should've been simple. Just:

1. Move the data off my 2TB drives
2. Cut out the bottom from the SSD caddy to make space
3. Install the SSDs
4. Set up the NAS software
5. Profit

It wasn't that simple.

Moving my data off my existing 2TB drives was a journey in realizing my only spare storage capacity was scattered across multiple smaller SSDs, trying to delete as much data I didn't need, attempting to compress 100s GB of data, noticing after 2 hours that the compressed archive was bigger than the data I was trying to compress, realising that compressing compressed data (i.e. .mp4, .mp3, .jpeg, etc...) yields little to no space savings, and finally settling for spreading out those files anywhere they would fit.

Modifying the caddy was initially pretty straightforward, and I was pretty quickly able to get everything installed onto the motherboard so I could boot the device. And... the SSD in the network card slot wasn't detected by the server. I figured this might be a case where plugging in one SSD disables PCIE lanes elsewhere on the motherboard. So I emptied out all the drives I just installed, and repeated the cycle of installing one drive and booting to check that it showed up. All the drives were functional. Assuming that maybe I had a faulty connection at the start, I reinstalled all the components into the chassis, closed it, and booted it up again. And the network card SSD is gone again! It turned out that installing the 2.5" SSD caddy seemed to cause the network card SSD to malfunction. I'm pretty sure the caddy was shorting out the SSD, because it eventually died. I ended up cutting the caddy in half to guarantee that no part of it could get even get close to my storage devices, which seemed to solve the issue.

The rest was pretty smooth sailing. OpenMediaVault was very easy to set up on Proxmox, and seems to work fine after using PCIE passthrough to give it access to the SSD controllers.

What did I learn

• I still don't have the mastery over home server administration that I would like, but getting stuff running is a step in the right direction.
• If you aren't using bluetooth or wifi, you may be able to repurpose your network card slot as extra storage.
• It's probably worth having a couple spare terabytes of storage available in case I need to move stuff around.
• Bridging contacts on your SSD is bad and might kill it :D

I built a standing desk I bought a surge protector

I built a standing desk to help me be less sedentary when I work from home. Thinking about cable management threw me down a rabbit hole about surge protectors, and learned some basic facts that I probably should've known years ago:

• Surge protectors are differently rated (in Joules [J]) based on what you're trying to protect. My understanding is that you want 3000J+ for PCs.
• You should think of your surge protectors as a consumable that needs replacing every couple of years.
• It might be worth going for metal (instead of plastic) to reduce the chance of the surge protector catching fire (no clue how necessary this is, I might've been reading documentation from a manufacturer trying to upsell me).
• If you're doing cable management under a desk, you might want to look into rack mount surge protectors since the plugs come out the sides instead of straight down.

I'm working on an infinite canvas [drawing / note taking] app that I call Mitten. I'm trying to design an elegant algorithm to support a truly boundless canvas (infinite zoom and pan). I thought I could solve this quickly but it's proving quite hard. There are so many pitfalls (floating point issues, how to uniquely identify a point within the world, data structure, camera system, etc).

About a week ago came across Dick, an esoteric language, and got sniped into making an interpreter for it in (zsh) shell script.
Well, at that point it was just rubber-duck with ChatGPT to think about how to do that.

Only today did I build out the interpreter script and a few sample programs.
https://github.com/master5o1/dick-sh