I like the idea of the generated PDF being a self-contained thing. From it, we would be able to get the source .typ file used to generate it (along with maybe the typst version?) and modify it as we need and make a new PDF. I don’t have much knowledge of the PDF file format, but looking at Wikipedia, it seems that a stream object type might be able to store the compressed binary data of the .typ file.
Is such a thing possible? Is it a bad idea? Thoughts?