- PDF is a successful technology for preserving visual fidelity in documents but was not designed for machine reading.
- The format's origins are in PostScript, which focused on printing with geometric precision rather than semantic structure.
- PDF gained ubiquity due to its ability to make documents look identical across different systems.
- Attempts to extract structured data (like text, tables, or reading order) from PDFs require complex reconstruction and inference.
- There are many types of PDFs (born-digital, scanned, tagged) that behave differently, complicating parsing.
- Table extraction from PDFs is particularly challenging due to the lack of inherent table structure.
- Tagged PDF standards exist to add accessibility but are inconsistently implemented.
- The ecosystem around PDFs includes numerous libraries and tools to address its limitations.
- PDF's history shows how a good solution for one problem becomes problematic when repurposed for other uses.