Tagged PDF support creates a lot of small simple dictionary objects. The management and serialization of these is adding a lot of overhead. This, for instance is the main reason why Pod::To::PDF::API6 is twice as slow as Pod::To::PDF.
I've begun an experiment to add specialized support for StructElem objects and their serialization to ObjStm (object streams), with the idea of integrating them into PDF::Tags for writing purposes only. Work has just started on branch struct_elem.
work dropped on struct_elem branch in favor of pdf-dom #2. I've decided to implement a general purpose native DOM that will target serialization to object-streams for use by PDF::Tags and others.
Tagged PDF support creates a lot of small simple dictionary objects. The management and serialization of these is adding a lot of overhead. This, for instance is the main reason why
Pod::To::PDF::API6
is twice as slow asPod::To::PDF
.I've begun an experiment to add specialized support for StructElem objects and their serialization to ObjStm (object streams), with the idea of integrating them into PDF::Tags for writing purposes only. Work has just started on branch
struct_elem
.