What Is PDF? Structure, History and Conversion Tips

When you email a contract, you want it to open on the other side with the same layout, the same fonts and the same page breaks. PDF (Portable Document Format) was built for exactly that. Today most documents that are considered "final", from invoices and court rulings to research papers and boarding passes, travel as PDFs. This article explains where PDF came from, what is inside the file, which versions exist, and what is kept or lost when you convert a PDF to another format.

What is PDF?

PDF is a file format that describes every page of a document with a fixed appearance. In editing formats such as Word, text "flows": narrow the window or change the font and the lines rewrap. In a PDF, the position of every character, line and image on the page is decided in advance. The file carries the text, the fonts (usually embedded), the images and the vector graphics in a single package, so the document renders the same way on a computer that does not have the original application, on a phone or on a printer.

That design has a cost: PDF is an output format. It does not always contain semantic information such as paragraphs, headings or tables. Often it only contains instructions like "draw these glyphs at this coordinate". Keeping that distinction in mind explains why editing a PDF, or turning it back into a Word document, is hard.

A short history

PDF goes back to a 1991 proposal by Adobe co-founder John Warnock called "The Camelot Project". The goal was to capture documents from any application and view or print them on any machine. According to Adobe, the effort had become PDF by 1992, and PDF and the Acrobat software that reads it were released on June 15, 1993.

For its first fifteen years PDF was an Adobe specification, and it gained features release by release: interactive forms (AcroForms) in 1.2, transparency and XMP metadata in 1.4, and layers, object streams and cross-reference streams in 1.5, published in August 2003. The real turning point came in 2008, when PDF 1.7 became the international standard ISO 32000-1:2008 on July 1, 2008 and control of the format moved to an ISO committee. PDF 2.0 (ISO 32000-2) followed in 2017, with a corrected second edition in December 2020. Since April 2023 the PDF Association has offered ISO 32000-2 as a free download.

For a sense of scale: in a June 2023 blog post Adobe said its customers had opened more than 400 billion PDFs in Acrobat products over the previous year. That figure only covers Adobe products; built-in browser viewers and other software are not included.

What is inside a PDF file?

Open a PDF in a text editor and the first line shows a version marker such as %PDF-1.7. The rest of the file has four main parts:

1. Header: the first line, declaring the format and version.
2. Body: the "indirect objects" that make up the document. Each object has an object number and a generation number. Pages, fonts, images and content streams are all objects.
3. Cross-reference table (xref): lists the byte offset of every object, which lets a reader jump straight to any page without scanning the whole file. Since PDF 1.5 this can also be stored as a compressed "xref stream".
4. Trailer and end of file: a dictionary pointing to the document catalog (/Root), followed by startxref (where the xref begins) and the %%EOF marker.

A simplified example looks like this:

%PDF-1.7
1 0 obj << /Type /Catalog /Pages 2 0 R >> endobj
2 0 obj << /Type /Pages /Kids [3 0 R] /Count 1 >> endobj
3 0 obj << /Type /Page /Parent 2 0 R /Contents 4 0 R >> endobj
4 0 obj << /Length 44 >> stream
BT /F1 12 Tf 72 720 Td (Hello) Tj ET
endstream endobj
xref ...
trailer << /Root 1 0 R /Size 5 >>
startxref ...
%%EOF

The actual page content lives in the content stream. In the example, the BT ... ET block is a text object, Tf selects the font and size, Td sets the position and Tj paints the text. Notice there is no notion of "paragraph" or "heading" here, only positioned glyphs. Even a single visual line is often split into several fragments.

Another PDF feature is the incremental update. When you add a comment or a signature, software can append new objects and a new xref section to the end of the file instead of rewriting it. Digital signatures rely on this mechanism to keep the signed content intact.

PDF profiles and sub-standards

Alongside ISO 32000 there are restricted PDF profiles for specific purposes. The best known is PDF/A, designed for long-term archiving. Its first part, PDF/A-1, was published as ISO 19005-1 in October 2005. A PDF/A file must embed all fonts, define colors in a device-independent way and include XMP metadata, while encryption, JavaScript, audio, video and external dependencies are prohibited. That is why archives and public bodies often ask for PDF/A.

A more practical distinction is between text-based PDFs and scanned PDFs. A file produced with "Save as PDF" from Word or a browser contains real text you can select, search and copy. A sheet of paper run through a scanner is just an image inside a PDF. Getting text out of it requires OCR (optical character recognition).

Where is PDF used?

• Legal and official correspondence: contracts, petitions and rulings, helped by digital signature support.
• Accounting and e-commerce: invoices, receipts and bank statements.
• Publishing and print: print-ready files, e-books and catalogs.
• Academia: papers, theses and lecture notes.
• Forms: fillable applications and declarations.
• Archiving: records that must be kept for decades, using PDF/A.

Strengths and weaknesses

Which programs open PDF?

You rarely need to install anything. Chrome, Edge, Firefox and Safari all include a PDF viewer. Preview opens PDFs on macOS and Edge does so on Windows. For advanced editing, form design and signing, Adobe Acrobat is the common choice. Microsoft Word can also open a PDF and try to turn it into an editable document, with results that depend on how complex the file is. On Linux, Evince, Okular and the Poppler tools are widely used.

Common problems and how to fix them

Text copied from a PDF comes out scrambled

A PDF may store text in drawing order rather than reading order. In a two-column paper the lines can interleave, and some fonts are embedded without Unicode mappings, so accented characters may look broken. Converting the PDF to TXT or Word usually gives a cleaner result than copying from the viewer.

The file is encrypted and will not open or convert

A PDF can carry two kinds of password: an open password and a permissions password (for editing or printing). No converter can read a file with an open password without knowing it. If you have the password, open the file, remove the protection, save it again and then convert it.

I cannot search a scanned PDF

The file contains images, not text, so search and copy do not work. You need OCR software to add a text layer.

Fonts look different from the original

If the creating application did not embed the fonts, the viewer substitutes similar ones. Enabling "embed fonts" when exporting, or saving as PDF/A, prevents this.

Tips for converting PDF files

Converting a PDF means rebuilding a fixed page layout as a flowing document, so even the best tool has to make educated guesses. Here is how our converter handles it: when you convert PDF to Word, the file is uploaded to our server, the open-source Poppler tools read each text fragment together with its position, and the lines are merged back into paragraphs. Character formatting such as bold and italic, and the reading flow, are largely preserved, but multi-column pages, complex tables and positioned boxes may not match the original exactly. Uploaded files are deleted from the server within 30 minutes.

• If you only need the text, converting PDF to TXT gives the cleanest result: formatting is dropped and readable plain text remains.
• If you want to keep the page look on the web, PDF to HTML places the text at its original positions on the page.
• Scanned PDFs contain no text, so our converter cannot extract editable text from them. Run OCR first.
• Encrypted PDFs cannot be converted; you will see a message asking you to remove the password.

Going the other way is easier. With Word to PDF, DOCX, DOC and RTF files go straight to PDF through LibreOffice with no intermediate step, so page layout, headers and footers, tables and images are kept. If you used a font that is not installed on the server, it may be replaced with a similar one. Sources such as HTML, Markdown, TXT and UDF are laid out on A4 pages and printed to PDF by a browser engine; UDF to PDF, for example, is a practical way to share UYAP documents.

Frequently Asked Questions

When was PDF created?

The idea goes back to John Warnock's 1991 Camelot proposal. PDF and Acrobat were released on June 15, 1993, and the format became the international standard ISO 32000-1 in 2008.

What is the difference between PDF 1.7 and PDF 2.0?

PDF 1.7 is the last Adobe version, standardized as ISO 32000-1 in 2008. PDF 2.0 (ISO 32000-2) was developed entirely within ISO, published in 2017 and corrected in a 2020 edition. In everyday use both open in the same programs.

What is PDF/A and when should I use it?

PDF/A is a restricted PDF profile for long-term archiving. It requires embedded fonts and forbids features such as encryption and JavaScript. Choose it when a document must open the same way years from now, or when an institution requires it.

Why does the layout break when I convert a PDF to Word?

Because a PDF usually has no paragraph or table information, only positioned glyphs. The converter has to rebuild lines and paragraphs from those fragments. Simple single-column documents convert very well; multi-column or table-heavy pages may need manual cleanup.

Can I turn a scanned PDF into editable text?

That requires OCR. Our converter does not perform OCR; it only transfers text that actually exists in the PDF. Run the scanned document through OCR software first, then convert it.

Is it safe to convert my PDFs?

Document conversion runs on our server. The file is used only for the conversion and is deleted within 30 minutes. For highly confidential material you may still prefer not to upload it to any online service.

Sources

• Adobe: What is a PDF? (The Camelot Project and the birth of PDF)
• Adobe Blog: Celebrating 30 years of digital transformation with Adobe Acrobat (June 15, 2023)
• Library of Congress: PDF 1.5 format description
• Library of Congress: PDF/A-1, ISO 19005-1:2005
• Wikipedia: PDF (history and file structure, secondary source)
• Poppler PDF rendering library