Private Static Sites¶
Not every useful website needs to be public.
Some of the most valuable sites are private working environments: committee handbooks, operational notes, internal references, procedural documentation, research collections, or personal archives.
For these purposes, static publishing remains remarkably effective.
Working assumption
A private site should optimise for clarity and usefulness rather than visibility.
Why static¶
A static site ultimately consists of files.
That simplicity matters operationally.
Static systems generally involve:
- fewer moving parts
- smaller attack surface
- easier backups
- easier migration
- predictable behaviour
- simpler hosting
For many forms of documentation and reference material, this remains entirely sufficient.
Why private¶
The modern web strongly encourages publication, visibility, engagement, and analytics. Much useful information does not benefit from any of those things.
Private sites allow a calmer model.
They can contain:
- draft procedures
- internal guidance
- committee documents
- reference material
- working notes
- archival collections
without needing to become public-facing products.
Privacy
Private does not necessarily mean secret.
Often it simply means “written for the people who actually need it”.
Current infrastructure¶
The current hosting model is intentionally restrained.
Components¶
| Layer | Technology |
|---|---|
| Source | Markdown |
| Build system | MkDocs Material |
| Hosting | Linux |
| Transport | Tailscale |
| Reverse proxy | Caddy |
| Deployment | Git hooks |
The sites themselves remain fully static.
Tailnet-only publishing¶
Many sites are available only inside a private Tailscale network rather than on the public internet.
This creates several advantages:
- simplified exposure model
- reduced attack surface
- no public indexing
- low operational overhead
- ordinary browser access
The result behaves more like a private reference library than a conventional web application.
Why MkDocs¶
MkDocs Material occupies an interesting middle ground between documentation tooling and lightweight publishing platform.
It provides:
- strong navigation
- fast static output
- search
- good typography
- structured organisation
- low operational complexity
Most importantly, it treats content as files rather than database entries.
Design philosophy¶
The visual layer draws heavily from older technical interfaces.
This is less about nostalgia than about operational clarity.
Older systems often assumed:
- sustained attention
- keyboard use
- long reading sessions
- dense information
- visible structure
Modern interfaces frequently optimise for atmosphere instead.
Interface preference
Interfaces should support concentration rather than compete for it.
Folder structure¶
The overall system is organised around plain directories and visible source material.
Typical structure:
The intention is that the archive remains understandable independently of any particular application.
Deployment model¶
Deployment is intentionally simple.
Workflow¶
Edit locally
↓
Commit changes
↓
Push to bare Git repository
↓
Post-receive hook
↓
Static rebuild
↓
Updated site
No database migration is required. No application state needs to be preserved.
The site itself is ultimately reproducible from source files.
Operational preference¶
A recurring theme throughout the system is legibility.
A good infrastructure system should allow its own structure to be inspected directly:
- folders should look like folders
- source should remain visible
- build steps should be understandable
- hosting paths should be explicit
- archives should remain readable
This is as much an operational preference as a technical one.
Current direction¶
Current work includes:
- refinement of private publishing workflows
- long-term archival organisation
- DOS/Borland-inspired interface systems
- improved mobile administration workflows
- integration between notes, documents, and websites
- stronger typography consistency across PDF and web output
The emphasis remains on systems that are quiet, maintainable, and durable.
“The best technology is often the technology that quietly disappears.”