UBI was supposed to be a mascot. Why was it an egg?

The aperture needed a resident

I wanted UBI to have a personality. A face, initially. Something that belonged to UBIKITOS rather than a generic chat bubble borrowed from every other website.

The aperture logo was the starting point. That mattered: the mascot should feel connected to the existing identity. But turning a geometric mark into a friendly character proved to be a rather effective way of discovering how many things can go wrong with two eyes and a mouth.

The eyebrows looked wrong. Some expressions read as sad or menacing. Bigger eyes created an entirely different problem. I wanted happy, excited and challenged; the results occasionally seemed to be auditioning for a less welcoming role.

By the time the direction changed, the better question was whether UBI needed a face at all. What if the character was a lava lamp inside the aperture? A moving material with a recognisable temperament, rather than a set of facial features trying very hard to look friendly.

The brief became wonderfully specific

The material needed to be shiny black vinyl. Glossy, three-dimensional and non-metallic. The shapes should deform, merge, separate and shed tiny droplets. The aperture would contain the character.

This is easy to describe when the desired image is already clear in your head. It is harder to obtain from a succession of approximations that each get one part right and another part wrong.

There was a visible circle inside the aperture. Then there were blobs that looked too much like eggs. The shape needed sharper reflections. The movement needed more deformation. Making the object darker did not make it more convincing; making it shinier did not automatically give it a fluid silhouette.

“Less eggs” became a remarkably efficient piece of art direction.

It sounds imprecise, but it identified a real visual problem. A pair of smooth ovals moving around a small space does not necessarily look like liquid. The shape itself has to change in a way that suggests material gathering, stretching and redistributing.

A polished surface still needs a shape

One of the useful distinctions in the conversation was between the material and the motion. They were separate problems, even when a single image made both look wrong.

A glossy material depends on reflections that give the surface readable volume. If those reflections are soft or muddy, the object can feel blurred. If the material is interpreted as metal, it can lose the black-vinyl quality I was asking for.

The silhouette has another job. It gives the object weight and character before the viewer studies the highlights. A beautifully rendered egg is still an egg. Improving the reflection does not solve a shape that never does anything interesting.

That is why the brief kept returning to deformation. I wanted the material to feel alive without needing a face. Its personality would have to emerge from the way it moved, gathered and relaxed. The aperture could supply the stable identity around it; the liquid inside would provide the changing presence.

Blender joined the conversation

I said Blender was available. Eventually it was open, and the task shifted from describing a look to getting the rendering script to run in the application.

The dated screenshots make this part of the story very concrete. One error concerned the active object in Blender's context. Another came from adding a modifier when the required active-object context was missing. A later error involved a render-engine name that the installed version did not recognise.

Then there was a missing world setting, followed by a render that could not be saved to the relative output location because that location resolved onto a read-only filesystem.

It was a small parade of compatibility and setup problems, each arriving after the previous one had been addressed. The material was the creative goal; getting the application to interpret the script correctly was the immediate work.

There is no need to romanticise that sequence. It was not a dramatic breakthrough. It was a series of very specific obstacles between an idea and an image file.

The black vinyl finally appeared

The conversation includes a rendered image that looked gray and matte, followed by a much darker glossy result with clear highlights. I had also clarified that the material should not be metallic.

That distinction mattered because “shiny” describes an appearance without specifying the kind of surface. A black plastic or vinyl object and a metal object can both reflect light, but they do not communicate the same material character.

The glossy black image was closer to the requested finish. It was still an oval form, so solving the material did not finish the mascot. There was also the question of its background. A square visible around the asset breaks the illusion that the character belongs inside the aperture.

Transparency therefore became part of the practical result. The blob needed to sit within the site's existing design without carrying its own little rectangular stage around with it. This is the sort of unglamorous asset detail that can decide whether an otherwise good render works in an interface.

“Where?” is part of a user experience

The later messages from the evening are short and familiar: “where?”, “how?”, and “I click and drag and nothing happens”.

Those are not failures of ambition. They are signs that the instructions did not yet connect with the application state in front of me.

The guidance had moved into Blender shape keys and proportional editing. The interface distinctions mattered: selecting the right key, entering the right editing mode, and changing the mesh rather than merely selecting an object. A technically correct instruction can still be ineffective if it assumes the user is looking at a different panel or working in a different mode.

That was an especially relevant lesson while making a mascot for a product meant to help people. The person receiving guidance should not need to reverse-engineer the guide's assumptions. “Click the button” only becomes useful when both sides mean the same button in the same place.

The screenshots helped because they replaced a guessed interface state with something visible.

The unfinished part was the interesting part

By the end of this dated sequence, the glossy material and transparent still image were further along. The fluid animation remained the next task. The later thirty-second animation work belongs to a later entry; it would be misleading to move that success back into this evening.

That leaves a properly untidy founder story. The visual direction improved through objections that ranged from specific technical problems to “still blurry.” The script encountered several application errors. The final material came closer to the brief before the movement did.

I still think the decision to remove the face was the interesting turn. A mascot does not have to imitate a person to feel like a character. Motion, timing, material and a stable silhouette can do much of that work.

For UBI, the aperture offered a home and the liquid offered personality. We had made progress on the surface. The next job was to persuade the egg to stop being an egg.