3D modeling just doesn’t suit me, so I’ve been experimenting with ways around it 🫡
Maybe because I’m used to coding, I prefer building things with text rather than operating a GUI.
Having source code that compiles into the thing you wanted feels like the principle of leverage — getting maximum effect from minimal force.
And the biggest perk is that with code you can do bulk replacements easily, and you also get diff management through git, which is really nice.

The above is a prototype-level domain-specific language tentatively called Hitoshina - Model, where I tried creating two cubes as a test.
It defines cubes b1 and b2, with code that places b2 on the right side of b1.
I’m still in trial-and-error and nothing is finalized, but I’m thinking about whether I can do modeling with syntax along these lines.
Mechanically, when you compile Hitoshina - Model code, it generates Universal Scene Description code as an intermediate language, which is then imported into Blender to ultimately produce a dae file.
Like the relationship between Markdown and HTML, my idea is that you’d code on one screen while seeing the modeling result rendered in real time on the other, building things that way.

In addition to the modeling experiments above, I’m also reimplementing the SL API in Rust. I can now log in with 2FA, upload a local script, and check the compilation result locally.
I’ve also gotten it to apply local scripts to scripts inside rezzed objects, so I’m scheming to make item production much more efficient 😃

On a different note, I’ve completely stopped taking photos lately.
Creating an avatar I’m satisfied with and taking photos in pretty places — that dress-up-doll style of enjoying Second Life — I did that for about three years, and I’m pretty full on it now 😌

Even so, Second Life has a wide embrace and lets you do all kinds of things, so it looks like I’ll be able to enjoy trial-and-error for quite a while yet ☀️