Protection, honestly

A module that starts with what it cannot do

The publication has a content-protection module. Before describing what it does, it is worth being straightforward about its ceiling, because most writing on this subject is not.

Text on a web page cannot be protected. It has to be in the HTML for search engines to read it, which means it is available to anyone who views the source, fetches the page with a command-line tool, or opens developer tools. That is not an oversight in this implementation. It is the deliberate consequence of wanting to be findable, and no amount of scripting changes it.

Anything claiming otherwise is selling something, or has not thought it through.

What it actually does

Within that ceiling, the module raises the cost of casual copying and makes captures attributable.

Blocked: selection, copy and cut, right-click, dragging images out, and printing — including print-to-PDF, which is the one people reach for first.

Deterred: saving images, and casual screenshotting. A faint tiled watermark sits across the content, so a screenshot is visibly from here. It is not obtrusive when reading and it is unmistakable in a capture.

Impossible: preventing operating-system screen capture, or hiding the text from view-source. Stated plainly in the module’s own documentation, because a security feature that overstates itself is worse than one that does not exist — it produces false confidence.

Production only

The protection is active on the live site and inert during local development.

That is not an oversight either. If it were always on, the preview would be unselectable and unworkable, and I would spend my time fighting my own tooling. The consequence is that its behaviour cannot be verified in a local preview, so it is verified against the live site instead — the only place the claim is actually being made.

What it is for

Not to stop determined copying, which it cannot. It is for the ordinary case: someone who would casually select-all, copy, and paste an article somewhere else without attribution. Friction plus a visible watermark changes that from thoughtless to deliberate, and most of it was only ever thoughtless.

Meanwhile the copy notice — the line that accompanies a copy attempt — points at the original. Someone who wants to quote a passage is given a pointer to where it came from rather than a silent failure.

Configurable, and off by default in spirit

Everything here is settings in the site configuration: whether it is on, the watermark text, its opacity, its size and spacing, and the notices. It can be turned off entirely without touching code.

I have kept it on, with the watermark subtle enough that most readers will never consciously notice it. The moment it becomes something a reader has to work around, the trade has gone wrong — the goal was never to make reading harder.