Initially PD4ML was seen not as a HTML capture, but PDF authoring tool, which uses a subset of HTML/CSS to define document layout. As our current HTML renderer has a number of limitations, we do not expect it will convert an arbitrary web page perfectly – it expects the source HTML is designed or refined bearing in mind PD4ML.

The situation is going to change with forthcoming PD4ML v4 release. Unfortunately a .NET port from Java will take some time and be published later.