OpenAtlas

I slept over it and tried today things out. If nobody has a better idea, I will write the initial manifest manually to get things done. But the more complex the manifests get (e.g. more pages, annotations, metadata), the more error-prone it will get.

• We are not limited to Debian packages if they are "just" JavaScript libraries (or similar) and can be installed via our calmjs/npm workflow
• Nevertheless, underlying libraries may be required. With the update to Debian bookworm around new year much more current (and possible more) libraries should be available

• We can just try some things out, you don't have to find the perfect solution on the first try
• Don't worry too much about how to manage the manifests. In a first step these could be uploaded as additional files via the user interface. We deal with how to manipulate them when we figured out the rest.

I also added our frontend team as watchers. Maybe they can help with this.

Libraries
I know, that we can use JavaScript libraries. The most promising is manifesto, but since the documentation/tutorial site is down, I don't know.

Manifest
The manipulation part is not of my concerns right now. But in order to view any image in an IIIF viewer, there has to be a manifest that points to that image.

Solution
The first step is, that OpenAtlas will provide an on-the-fly manifest of that image. This, I will create by "hand" (so putting together the JSON file). Then the image can be viewed in a viewer with the appropriate licence tag, reference and description.
The second step (there will be a new issue about that) is, that the user can create with some (external) tools a manifest and can upload it to an entity (e.g. place, NOT file). In this manifest, the user can add a sequence, canvas and so on which is linked to that entity.

I think we most likely will talk about this in person.