Containerizing some thoughts.

Previously - 2025 Q1 - Spoon Shortage

It's fair to say for a lot of the last few months, we've felt stalled - we've been ill, work has been changing at a rapid pace, and the weight of the world is getting us down a lot more than it used to. A worrying trap of browsing the same few websites again during non-working hours, starting projects and not finishing them.

And the worrying thing is, when we've reread our post for Q1, a lot of those same sentiments have been expressed.

But at the same time, progress is progress, no matter how slow.

Around 2016 or 2017, we had thoughts about what it would take to move from first line support to a DevOps engineer. The certifications, the career moves that might be needed - before the pandemic pushed us towards prioritizing stability while we can, as the best way to help others.

The other day, while we were making a pull request to work's Terraform repo, we were hit with the realization that we had arrived and not even noticed it, largely because it happened so gradually. Keeping a curious mind and just learning about things when the opportunity came up.

q1 goals

In Q1, our big focus was Furality Somna. We ended up customizing 9 different avatars. And then also learning the basics of video editing, because doing something like this needs showing off.

Check it out y'all: furality_somna_fits.mp4

[image or embed]

— Dex 🔞 (@dex.nullpat.ch) Jun 6, 2025 at 13:10

But all of these... took a while to coalesce. Furality started on June 5th, and even our main avatar didn't have basically any work done to it until May 27th.

The output of git log, showing multiple commits relating to Furality Somna between May 27th and June 10th
Understanding when things happened is a great reason to put stuff in version control, among the other reasons

We bought some makeup and some plus size skirts... but we haven't had opportunity to wear them. Again, still worried about the situation at home, but the prep work is there. [1]

We've taken time out to visit Scotland and the Peak District. We've only seen small parts of both, but they're both beautiful places that we'd like to see more often.

That cert work wants us to get? We've got a training course scheduled in a couple of weeks.

And we've kept up with the idea of trying new things when it comes to restaurants and the like; to order things we don't remember getting at them before. [2]

looking forward

We've made our big expensive tech purchase for this year. We're planning to get another article out about that one, because hey, gotta justify it somehow.

Ordinarily after Furality, we have some sort of Tetris effect induced dream that provides some idea for our yearly theme for next year, and we slowly start transitioning that. That didn't happen this time round - perhaps the con being dream themed broke our streak, perhaps our sleep being interrupted caused some problems. But Furality did still provide some inspiration.

Let's revisit our recent article about our 2D workflow for customizing VRChat avatars for a second.

But we're not good enough at Blender to do more than the basics, and we count UcuPaint as more than the basics. (It also feels like it's something that makes more sense with a graphics tablet).
...
We still consider ourselves scared of Blender, in a way. We don't touch 90+% of the things it can do for fear of things breaking; we've had avatars break all animations with simple tweaks because the T-pose was somehow different between what was in Unity and what was in the provided Blender file.

Blender is a big complicated piece of software, with more functionality hidden behind obscure keyboard shortcuts than we will ever know.

But it also isn't brain surgery. If something goes wrong, there's backups, there's version control, there's reverting back to provided base files if we need to. During the process of getting avatars ready for Somna, we became a lot more comfortable with the thought of kitbashing things from different avatars. And one of the outfits for one of the avatars edited is really annoying to edit without some way of working in 3D.

A UV map showing a '150' decal spreading over both the top and bottom of the UV map - because it's one continuous unit rather than separated into UV islands
This is what happens if you don't mark UV seams - so the pride badges are on the front of the jacket, and the 150 decal is on the shoulder. I think for this one, we had to rely on Poiyomi's decal placing tool to get them lined up properly, and then copy these things that should be vectors back to the original texture as overlaid raster images.

And while browsing Furality's dealer's den, we saw a lot of vendors offering renders of your avatar. And with every one of those we saw, we thought "why can't we do that?"

We're definitely not doing a big complicated animation any time soon. Our struggles with anatomy make modelling from scratch probably not on the cards for the moment.

But taking something someone else has rigged, posing it, and rendering a basic scene? That feels like something we should be able to do.

So, next avatar we customize, I want to at least try customizing the texture in UcuPaint, and making a render to introduce them. [3]

And that brings us on to the broader point.

spilling the t

For a while, we've thought about ourselves in the sense of the "T shaped professional" - approximate knowledge of many things (the horizontal), deep knowledge of some speciality areas (the vertical). A lot of people are the same way (after all, it is the perfect shape).

For the next 18 months, rather than making the horizontal wider or the vertical deeper, we want to make the vertical wider - to turn more approximate knowledge into concrete skills.

We know of Kubernetes, and we know that work is going to increasingly move towards it as an application platform. It makes sense to get a cheap ex-business PC or two where we can upgrade the RAM and see how it does - and in turn, perhaps that'll be the thing to get us to try a lot of cool self hosted apps that we've heard about and then never set up, or to fix the way this site and ❤️'s site are deployed.

We take a lot of photos, but they're primarily out of the camera, default settings - whether that's our phone or the VRChat camera. We tend to do this because photos are a big part of how we remember things - and as such, the plain shot with everyone there clearly is of more use to future us than the artistic shot with manual focus - still, it would be nice to be able to do more things when the opportunity presents itself. To actually learn some editing skills.

I don't know what the catchy title for next year will be yet (compared to the Year of New Adventures, which came very naturally). Possibly "Year of Focus", possibly "Year of Refinement", absolutely not "Year of T" in case it gives the wrong idea of what hormones we want running through our body. But we've got a while to decide.

q3 goals

Some goals for this quarter. We might not hit all of them, but it would be good to try.

  • Buy one new avatar base, edit using UcuPaint and do a render.
  • Set up one node of Kubernetes "cluster" for homelab - potentially including staging version of this site + ❤️'s site, a better git solution than just relying on a fileshare, and maybe something for observability because that's going to be an increasing part of the role going forwards.
  • Get the certification work wants us to get.
  • Spend less time on social media, including non-educational YouTube use.
  • Take another cheap coach trip somewhere interesting, and take some nice photos while we're there - possibly even buying Halide to see how it impacts.

footnotes

  1. (If anyone would like to forcefem us the next time we're at a furry con, please do). ↩︎

  2. Helped by a very generous friend that has encouraged us to indulge more when it comes to days out, rather than getting whatever's cheap at Greggs/Poundbakery. ↩︎

  3. Foxtober is coming. Which also makes for a pretty convenient quarterly deadline for these things too. ↩︎