A portrait can be a good fit for the visual language of a site and still be poorly served by the way the site displays it.
That became clear while reviewing the Archive. The images were being fitted into a consistent frame. The layout looked orderly, but parts of the work were disappearing outside it.
The effect was doing its job. The container was making the wrong decision about the image.
This is the tension I keep returning to with SYSTEM BOOT. The screen has a strong visual character: phosphor green, scanlines, dark surfaces and the language of older computing interfaces. Those choices give the site a recognizable atmosphere. They also need to make room for the things people came to see and read.
A consistent frame can hide a different image
A landscape screenshot and a portrait do not ask the same thing of a layout.
Put both inside a fixed-height frame and one may look perfectly natural while the other loses its composition. A face sits too close to an edge. A detail disappears. The visitor sees an editorial decision that the original image never made.
There are places where a crop is useful. A directory thumbnail can give a quick indication of what a record contains. But the main image in an artwork's detail view needs more room to explain itself.
For the Archive, the useful change was straightforward: let the image fill the available width and let its height follow its proportions.
That means the page has to accept different image heights. I think that is a reasonable trade. The directory provides consistency through its structure and labels. The work inside a record can keep its own shape.
A centered crop removes part of the portrait.
Width fits the column. Height follows the image.
The treatment should have a way back
The Archive's image treatment brings photographs and artwork into the same environment as the interface. At rest, the images are subdued and largely monochrome.
That creates continuity, but color can be part of the work. It affects mood, contrast and the relationship between details. A treatment that changes those relationships should leave the original easy to inspect.
On a desktop, hovering over an image reveals its color. Keyboard focus does the same. Clicking or tapping opens an enlarged view in original color.
Those actions serve different situations. Hover offers a quick comparison. The enlarged view gives someone time and space to look. On a phone, tapping provides a route that does not depend on hover existing.
I want the visual treatment to introduce the image, then give the viewer access to the source.
Motion competes for attention
A moving detail is difficult to ignore, which is exactly why it deserves restraint.
A brief signal at the moment an interface changes can help explain what happened. Continuous movement beside a paragraph asks for attention throughout the reading experience, even when it has nothing new to communicate.
The question I find useful is specific: what should the person notice at this moment?
During an entry sequence, motion can establish the environment. While someone reads an article or examines a portrait, the content needs that attention. The same effect can be appropriate in one place and intrusive in another.
Timing, repetition and placement all matter. So does responding to a reduced-motion preference. Turning down movement should leave the route through the site understandable and complete.
Test the work underneath
It is easy to review an effect in the context where it looks best. A luminous mark on a large dark screen makes a convincing specimen.
The more useful checks are less flattering. Read a long paragraph. Inspect a pale image. Open the same page at phone width. Switch to the light theme. Move through the controls with a keyboard.
Those checks reveal where the treatment starts making decisions on behalf of the content. A glow may soften a small label. A dark overlay may flatten an image. A frame may look balanced while obscuring the part someone wants to inspect.
I do not take that as a reason to abandon the visual language. It gives me a specific adjustment to make: reduce an effect, change its timing, let an image resize, or provide a clearer route to the original.
The test is whether the work remains legible and available inside the environment I have built around it.
Let the image set a limit
The recent Archive changes did not require a new identity. They required a better relationship between the identity and the material.
The interface still has its phosphor atmosphere. The images retain their proportions. Their color is available. The enlarged view gives them more space.
That is the balance I want: enough character to make the place recognizable, and enough restraint to let each piece of work be seen.
An effect earns its place when it helps someone notice, understand or use what is there. When it starts obscuring that work, it needs to change.
