Preface
I'm not going to sugar coat it. I don't like PDF as a format. While omnipresent on the web, the format is — justifiably — criticized for its poor usability and accessibility. It had its time in the spotlight but we should be moving. Its shortcomings are numerous and, unfortunately, invisible to most users.
About the PDF format
Portable Document Format (PDF) files are self-contained and pack whatever assets they need to display as designed. That is so say they do not require fonts, images or other external files to distribute. It all gets bundled in one neat package that's cross-platform and cross-device compatible, making it an attractive solution when the goal is to display a document exactly as it was originally designed.
The format has been around since the early 90s. In 2007, Adobe placed it in the hands of the International Organization for Standardization (ISO).
Modern PDF documents can now be produced in a variety of standards: PDF/UA for accessibility, PDF/A for archival, and a handful of others for design and print. The standard PDF format is, by a wide margin, the most popular and the default profile used by most text editors.
Accessibility concerns
They're really difficult to remediate
Remediating an inaccessible PDF is often a complex (and mind-numbingly tedious) process that requires a solid grasp on its semantics and the proper hierarchy of tags. While Acrobat does provide a handful of tools like the auto-tagger and a validator, fixing the issues the tools uncover is going to require some expertise.
Even with the knowhow and the skillset, the hierarchy of basic elements like lists and links are not intuitive and differ from other similar markup languages like HTML.
Limited customization
In general terms, you can't really open a PDF and quickly modify it to suit your needs or preferences, and this is by design. The format offers consistency and conformity, regardless of the platform or device its viewed on.
A practice I've seen from users with dyslexia is to replace or alter font stacks to increase legibility. This is easily done in most browsers and while some PDF readers do allow that level of customization, the options are generally limited and can cause reflow and layout issues.
Poor mobile experience
PDF documents are typically designed on the desktop, for the desktop. This results in the user zooming in and out, and panning left to right to read, which not only gets irritating quickly but is a major accessibility hurdle.
Terrible fallbacks for screen readers
Unless meticulously tagged, screen readers can really struggle to interpret PDF documents. An untagged PDF is difficult to use and can cause screen readers to interpret the document as "blank." As a workaround, users can perform an OCR, but the end result is rarely worth its salt.
This is less of an issue with the advent of artificial intelligence, but its still a hurdle screen reader users often have to overcome.
Forms are puzzling
Even when properly designed, PDF forms aren't great for accessibility. The format offers little support and a limited toolkit to work with. At best, an input can be assigned a short, plain-text tooltip, which leaves formatting information and other instructions off the table. There is no way to group a series of fields or divide a form in section other than with headings, which isn't ideal.
While some readers offer basic error validation, they generally do not offer anything beyond notifying the user of empty required fields, or offer regex-based validation. This limits the feedback that can be provided to a boolean state; either it's okay, or it's not, but the error message generally can't be customized.
Usability and other concerns
Beyond accessibility concerns, the PDF format has a handful of usability and other concerns that need to be addressed.
Indexing and tracking
Because PDF readers are external modules, it's nearly impossible to capture detailed analytics on the use of PDF documents other than the number of downloads or hits.
Additionally, search engines prefer HTML to PDF documents as HTML provides more contextual information. Landmarks elements such as <main> and <nav> provide an outline of an HTML document and its features, whereas PDF documents can only divide content into articles and generic sections.
Difficult and expensive to maintain
Once a PDF document has been published, maintenance can be difficult and requires specialized, licensed, and expensive software. Even with the software, it's often impossible to modify a document's structure without an extensive amount of work as the layout tends to shift around to accommodate the changes.
It's considerably easier to generate a new PDF from the source material than it is to maintain an existing documents. However, the originals are often difficult to track within an organization or are, in my experience, often discarded once the document is published.
Lack of documentation
When it comes to the accessibility remediation PDF documents, there's surprisingly very little useful information available other than the deeply technical ISO specifications and syntax guides. You'll find a sea of companies willing to fix your documents for you, and plenty of general how-tos, but detailed information on the tag tree and other techniques is very sparse.
Other concerns
Other less impactful considerations are the significantly larger file size of PDF documents compared to HTML, and the security concern of attaching documents to emails — particularly forms — that could contain unencrypted personal information.
Why are PDF documents still popular?
Primarily because they're easy to generate. The option to convert, export, or print to PDF is baked into most office applications and allow the user to quickly and painlessly generate pristine-looking documents. It's significantly easier to generate a PDF for the web than it is to code the document in HTML and CSS.
They also provide a sense of security. Given how difficult it can be to modify a PDF and that doing so requires specialized software, there's a general sense of security that comes with the format. Many seem to be believe that once a PDF is exported, it cannot be modified and is therefore safer than other formats.
When to use a PDF
In short, you should not be using a PDF to disseminate information. Consider producing your documents in HTML, which is far more user friendly, lightweight, and is fully compatible with assistive tools. If offline portability is a factor, consider an epub — essentially an HTML package — as an alternative to PDF.
The PDF format excels at one thing: retaining a document's original design, regardless of the platform it's being viewed on. This is particularly useful for archival, visual or graphic design, for blueprints and plans, and for some legal documents. Otherwise, stick to HTML and maximize the accessibility and the usability of your documents, but please don't use a PDF.