Hacker News Reader

Comment by jabbany | Hacker News Reader

Franky somebody should do like a TinyDF replacement for pdf... That's also a somewhat bloated format at this point.

Ideally though, the tiny* formats should be forward compatible (?) and interpretable as a valid non-Tiny* document. That would be quite nice for wider adoption (like JSON).

View on HN

"somewhat bloated" is the understatement of the century. PDF documents can contain executable code in two different programming languages (JavaScript and PostScript), interactive content, digital signatures, their own form of encryption, and more. Why they thought adding JavaScript to an already infamous attack vector of a format is beyond my understanding.

Javascript was added to PDF in order to implement interactive and dynamic forms, including input validation and normalization and dynamically adding/removing form elements based on user interaction. The contents of forms can also be directly submitted to a server. Basically, Adobe didn’t want HTML forms to steal their show, so they implemented the equivalent with PDF. Not saying that was a good idea, but here you go.

PDF can contain executable PostScript? That was new to me. I know the syntax is still PostScript, but without the programming language. Is there a nice description somewhere?

It seems that my understanding of this was not correct. I should have checked this before posting it, but I have a vivid memory of some PDF that rendered a different maze each time you opened it using PostScript and that override everything else. Most of what I'm reading suggests that the turing-complete aspects were removed in the design of PDF.

Old versions of PDF allowed to embed arbitrary PostScript for some area of the page, but this was always specified not to have any effect except if the PDF gets converted to PostScript. Especially it was specified to be ignored by PDF viewers and non PostScript printers, making it rather useless. In current PDF versions, this is no longer allowed.

PDF is indeed a very bloated format, but documents are inherently complex. I don't think you can create anything that deserves the name "tiny" but can still handle anything you might want to send to a printer (and that's a much smaller use case than "anything you want to display on a screen").

PDF is bloated because it went beyond documents into 3D models and interactivity.

A compressed, paginated and indexed postscript file is not complex.

What about: DjVu [1]?

It's not very well known, but it's so elegant, and also efficient, especially for scanned files.

[1] https://en.wikipedia.org/wiki/DjVu

PDF/A is a standard with a bunch of features excluded.

As someone who has implemented generating PDF/A documents, I can say it is still a very complicated format, and I would be very happy if it was replaced with something a little more straightforward.

For PDF there are OpenXPS and DjVu.

Reply on news.ycombinator.com