PDFs look like simple documents. They aren't. Under the hood, a typical PDF is a structured container holding pages, fonts, images, metadata, and sometimes more interesting things. Here's the gentle tour.
Pages and content streams
Each PDF page is a separate object holding content streams — instructions like 'draw this text here, in this font, this size, this colour'. Pages don't flow into each other; each is self-contained. That's why reordering pages is easy and editing text mid-paragraph is hard. The page is more like a printed sheet than a Word document.
Fonts, images, and resources
Pages reference shared resources: fonts, images, colour profiles. Two pages using the same logo don't carry two copies — they share one reference. Two pages using the same font don't double the file size. This is part of why PDFs stay reasonable even with lots of pages. Embedded fonts ride along so the file looks identical anywhere; Flint preserves them on save.
Metadata, attachments, and hidden bits
Most PDFs carry metadata: author name, creation date, original software. Some carry attachments — other files tucked inside the PDF like email attachments. Some carry annotations as a separate layer above the page content. Older PDFs sometimes have scripts (JavaScript), now usually blocked by viewers. Before sharing sensitive files, strip metadata and attachments — a final save in Flint usually does this.
FAQ
Can I see the structure of a PDF?
Some advanced viewers expose the object tree. For everyday use, you don't need to. Knowing the layers exist is enough — pages, resources, metadata, annotations.
Why are PDFs so editable in some tools and not others?
Editing requires understanding the content streams and being able to re-emit them safely. Some tools just render; some genuinely edit. Flint does the latter, which is why it can change text in place.
What's the maximum PDF size?
The PDF spec allows huge files (gigabytes). Practical limits come from software, not the format. Most PDFs you'll meet are well under 100 MB; compress PDF handles outliers.
Pages, resources, metadata. Once you see the layers, PDFs make sense. Open one in Flint and poke around.