What is a PDF, really?
Everyone uses the format. Almost nobody is told what it actually stores — which is why so much PDF software behaves in ways that seem arbitrary.
A PDF describes a page, not a document
PDF stands for Portable Document Format, published by Adobe in 1993 and an open ISO standard since 2008. The name undersells what makes it different. A Word file stores a document: paragraphs, headings, a structure that software can reflow. A PDF stores a set of instructions for painting a page — put this glyph at this coordinate, draw this line from here to there, place this image in this rectangle.
That is the whole design goal. A PDF looks identical everywhere because it leaves nothing to interpretation: the positions are absolute, the fonts are embedded, the page size is fixed. Print it in Tokyo on a machine that has never seen the font and you get the page the author saw.
Everything people find frustrating about PDFs follows from that same decision. The format is precise about appearance and almost silent about meaning.
Why there are no paragraphs inside a PDF
This is the fact that explains the most confusing behaviour. A PDF has no concept of a paragraph, a heading, a column or a table. It records characters at coordinates. When you copy text out of a PDF and it arrives scrambled — a two-column page interleaved line by line, a table collapsed into a run of numbers — nothing is broken. The software is reconstructing structure that was never stored, by looking at where things sit on the page and guessing.
Good extraction guesses well: it groups characters into lines by their vertical position, splits columns by finding the gutter, rejoins words broken across a line break. Bad extraction concatenates characters in whatever order the file happens to list them, which is why the same PDF gives clean text in one tool and gibberish in another.
It is also why editing a PDF is genuinely harder than editing a Word file, rather than just less well supported. There is no sentence to edit. There are glyphs at positions, and changing one means deciding what should happen to every glyph after it.
Text PDFs and scanned PDFs are barely the same thing
Two files can both be PDFs and have almost nothing in common inside.
- A text PDF — exported from Word, a browser, or a design tool — contains real character data with embedded fonts. You can select the text, search it, and extract it. These files are usually small, because text is cheap to store.
- A scanned PDF is a photograph of a page wrapped in a PDF container. There is no text in it at all, only pixels. Selecting text does nothing, search finds nothing, and the file is large because it is really an image. Adding a text layer to one of these requires optical character recognition.
What else the file is carrying
Beyond the visible page, a PDF holds things people rarely think about. There is an Info dictionary with a title, author, and the software that produced the file, plus creation and modification timestamps. There is often an XMP metadata block saying similar things in more detail. There may be bookmarks, form fields, annotations, attached files, and a JavaScript action or two.
Most of that travels with the document when you send it, and none of it appears on the page. The author field in particular is worth checking before a file leaves your organisation — it is frequently the name of whoever originally created the template, not whoever sent it.
The versions with letters after them
You will occasionally meet a PDF with a suffix. These are subsets of the format, each one banning whatever gets in the way of a particular job. PDF/A forbids anything that stops the file rendering identically in fifty years — no external font references, no encryption, no video. PDF/X constrains colour and fonts for commercial printing. PDF/UA requires the structural tagging that a screen reader needs.
A file can be a valid PDF and fail all three. That is normal, and only matters when someone specifically asks for one of them.
Frequently asked questions
What does PDF stand for?
Portable Document Format. Adobe created it in 1993 and released it as an open ISO standard (ISO 32000) in 2008, which is why any software can read and write PDFs without licensing anything.
Why can't I edit a PDF like a Word document?
Because a PDF stores characters at fixed coordinates rather than paragraphs. There is no text flow to insert into — changing a line means deciding where every following glyph should move, which the file gives no guidance on. Editors work around this by patching a line in place or by rebuilding the page.
Why is the text in my PDF not selectable?
It is almost certainly a scan — an image of a page inside a PDF container, with no character data at all. Running it through optical character recognition adds a text layer that makes it selectable and searchable.
Why did copying text out of a PDF scramble it?
The file records characters at coordinates with no notion of columns or reading order, so the software has to infer the order from the geometry. Multi-column layouts and tables are where that inference most often fails.
Are PDFs safe to open?
Generally yes, but the format supports embedded JavaScript and attached files, which have been used as attack vectors. Modern readers sandbox or disable those by default. Treat an unexpected PDF from an unknown sender with the same caution as any other attachment.
Does a PDF store who made it?
Usually. The Info dictionary records the author, the producing software, and creation and modification dates, none of which appear on the page. It is worth checking before sending a document outside your organisation.