Thoughts on precision and layout in XML-based PDF publishing

This started while I was writing academic papers. I could usually get acceptable PDFs, but having precise control exactly where it mattered most was always frustrating.

Most tools can produce PDFs, but once you compare layout precision, typography, usability, and long-term stability, the differences show quickly. Many workflows don’t provide satisfying results when you start measuring details closely.

I went through most of the usual XML-to-PDF stacks: DocBook with XSL-FO and HTML routes, DITA and DITA-OT pipelines, Apache FOP, Prince XML, Paged.js, Typefi-style systems, and a range of enterprise tools. I also spent time with TeX-based workflows. They are incredibly powerful and impressive, but in practice the complexity and friction often push people to admire them more than actually use them day to day.

Where things tend to break down is layout fidelity. Font rendering depends on ambiguous units or web-engine limits, styling has a steep learning curve, and many systems feel disconnected from traditional DTP practices.

After running into the same issues for years, I ended up building my own solution focused on predictable PDFs, stable pagination, and consistent typography. People coming from both the XML world and traditional tools like InDesign seem to feel comfortable with it, which was exactly the gap I was trying to explore.