Filling a PDF form on a phone is still awful. Here is what we did about it in Kernstra

by•

Try filling a W-9 or a patient intake form on your phone inside a web app. The page zooms in when you tap a field and never zooms back. There is no Tab key, so you pinch around looking for the next box. The date field wants a format it does not tell you. The total does not add up because the form's calculation never ran.

PDF forms were designed for Acrobat on a desktop, and many web viewers show the fields without the behavior behind them. We spent a lot of time on this in Kernstra, our PDF SDK for web apps, so I wanted to share what we ended up with and hear how others handle it.

What the viewer does with a form:

  • It fills the fields the PDF already defines, so a filled form is still a real form in Acrobat or any other reader. Text, check boxes, radio groups, dropdowns, list boxes and signature fields.

  • It understands what kind of field it is. A field the form defines as a date gets a date picker. Currency and percentages show formatted ($1,234.50). Zip codes, phone numbers and masked fields only take what fits. A value outside the field's range is refused with a message.

  • Calculated fields work. Sums, products, averages and "Qty * Price" style formulas are calculated when the form opens and again whenever a field changes.

  • It does all of that without running the document's JavaScript. The standard Acrobat form functions are recognized and handled by the SDK itself, so a form's scripts are never executed in your app.

On a phone or tablet:

  • Tapping a field does not make iOS zoom the page.

  • A bar with Previous, Next, "Field 3 of 11" and Done replaces the Tab key. It moves up with the keyboard and scrolls the focused field clear of it.

  • The layout is measured from the element you give the viewer, not the browser window, so the same configuration works in a full page, a split pane or a mobile web view.

For developers: you can prefill a form from your own data with one call (), listen for every change, hold back the toolbar's Download until required fields are filled, and save the result flattened or still editable. Everything runs in the browser and the document is never uploaded.

What it does not do: it does not build forms or detect fields on a flat scan. The fields have to exist in the PDF. A field driven by custom JavaScript behaves as a plain field. Signing a signature field is an electronic signature, not a certificate-based digital one.

You can try it with your own form, no account needed: (there is a phone and tablet preview in the top bar).

Question for the builders here: if your product has users filling PDFs, do you keep them in the PDF or rebuild the form in HTML and generate the PDF afterwards? I am curious which one people regret less.

26 views

Add a comment

Replies

Be the first to reply

Have a question or a thought to share? Add a comment above to start the conversation.