Converting Word to PDF: fonts, options, accessibility

Every spring, a fiduciary office in Zug sends out dozens of cover letters with the annual accounts. They used to go out as Word files. Then a client called: on his computer the letter was two pages long, the logo had shifted and the table had broken apart. Since then, everything goes out as PDF. Only nobody in the office knows exactly which of the options in the save dialog actually do something.

Creating a PDF takes seconds. Whether the result looks the same everywhere, can be read by a screen reader and contains no traces of revisions is decided in those seconds.

Why PDF and not Word

Microsoft gives three reasons for PDF: the file looks the same on almost all computers, it is smaller, and it conforms to an industry format. As examples, Microsoft lists CVs, legal documents, newsletters and files meant for reading and printing.

Microsoft is also clear about what a PDF does not do: many programs, including Word, can open and edit PDF files. Sharing as PDF alone does not guarantee that the recipient cannot change the file. A PDF is a format for consistent appearance, not write protection.

The options when saving

In Word, the route is File › Save a Copy or Save As, the PDF file type and, for the settings, More options › Options. Microsoft describes these possibilities:

Option	What it does
Page range	Output only some pages as PDF
Publish what: Document or Document showing markup	Include tracked changes in the PDF or not
Create bookmarks using	Turn headings or Word bookmarks into PDF bookmarks
Document properties	Carry author, title and other properties into the PDF
Document structure tags for accessibility	Make the document easier to read with screen readers
ISO 19005-1 compliant (PDF/A)	An archiving format so the document looks the same later
Bitmap text when fonts may not be embedded	Output text as images so it looks the same
Encrypt the document with a password	Only people with the password can open the PDF

For a cover letter, three of these rows are decisive: the markup, the document properties and the structure tags.

Fonts: why a PDF can look different

A PDF only looks the same everywhere if the fonts used travel with it. Microsoft describes the option “Bitmap text when fonts may not be embedded”: if a font cannot be embedded, the PDF uses images of the text so it looks like the original. If the option is not selected and the font cannot be embedded, the PDF reader may substitute another font.

That explains some of the surprises: a house font that may not be embedded appears at the recipient’s end in a substitute font, with different line breaks. The bitmap variant keeps the appearance, but at a price: text as an image cannot be searched, copied or read by a screen reader.

Which fonts a PDF actually contains is shown in PDF programs’ document properties; Adobe lists font information there as a category of its own.

Tracked changes and hidden details

If Track Changes is on in Word, the question when saving is whether they should go into the PDF. Microsoft describes choosing “Document showing markup” under Publish what to include them, or “Document” to leave them out.

Even if markup does not end up visible in the PDF, a step before is worthwhile: Word’s Document Inspector. According to Microsoft, it finds comments, revisions, versions, document properties, hidden text and more, and removes them on request. Microsoft recommends running it on a copy.

Accessibility: structure tags

A PDF that looks good is not automatically readable by everyone. People who use a screen reader need structure: what is a heading, what is a list, in which order is it read, what does an image show?

Microsoft describes the option “Document structure tags for accessibility” for this and recommends running Word’s Accessibility Checker before creating the PDF. The tags can only carry over what the Word document contains: headings as styles rather than bold text, alternative text for images, real tables rather than tabs. A cleanly built document yields an accessible PDF with almost no extra effort.

PDF/A for documents that stay

The option “ISO 19005-1 compliant (PDF/A)” creates, according to Microsoft, a document in an archiving standard so that it looks the same when opened later on another computer. For a cover letter this is rarely needed. For documents kept for years it may be worth considering — PDF/A is a topic of its own, with its own rules.

Quality or file size

In several Office programs the save dialog offers two levels. Microsoft describes them like this: if high-quality printing is needed, choose “Standard (publishing online and printing)”; if file size matters more than print quality, “Minimum size (publishing online)”. The difference is felt mainly where documents are printed and contain photos or fine graphics.

For a letter with a logo, the smaller level is usually enough. For a document with plans, photos or fine graphics that will be printed, the larger one is worth it.

Common surprises and their cause

What you see	Common cause
Different line breaks, different font	A font was not embedded and was substituted when read
Text cannot be selected or searched	Text was output as images (bitmap text) or the document was a scan
Comments or coloured changes in the PDF	“Document showing markup” was selected
No outline in the sidebar	“Create bookmarks” was not selected
An odd title in the window	Document properties were carried over with an old title
The file is unexpectedly large	High-resolution images and the “Standard” level

Each of these causes can be fixed in the save dialog or in the Word document itself — usually faster than repairing the PDF afterwards.

A short routine

For the fiduciary office in Zug, a five-step routine is enough:

1. Build the Word document with styles and run the Accessibility Checker.
2. Run the Document Inspector on a copy if the document had comments or revisions.
3. When saving as PDF, choose “Document” rather than “Document showing markup”, switch on structure tags, and include document properties only if they are right.
4. Open the PDF: does it look as expected? Can text be selected?
5. Glance at the document properties.

The PDF tools we are building will convert Word documents to PDF, processed in Switzerland, in memory and without any file being kept. The quality of the result still depends on the source document: a well-built Word document makes a good PDF, whichever way it is converted.