HN Explorer

Do we need a new PDF reader?

7 points by milani ↗ HN
I'm not satisfied with PDF readers specially on my linux machine and for academic work. Recently Mendeley and Readcube came to life but they focus on things other than the reader itself (Plus Readcube doesn't support Linux).

Are you satisfied with current readers? What are your needs?

10 comments

[ 0.25 ms ] story [ 65.1 ms ] thread
secoif
what specific needs do you have that aren't met?
milani
Among issues, I can point to two:

Usually I should scroll down to the end to see references and then back to top to continue reading, life could be a lot easier if readers could parse references and show them in the sidebar or split the view so I could scroll down the bottom pane.

I can not repeatedly select a portion of text, move to toolbar, click on Annotation, select Highlight, in order to highlight a sentence! I wish readers had better support for keyboard or at least used modern UI/UX techniques.

arh68
All my 'needs' are met so here are some wants. Rendering pdfs was solved long ago, so I'm not sure what distinction you see between Mendeley and 'the reader itself'. It's been a while since I used Mendeley, but it wasn't perfect.

References would be best done a la Wikipedia, hover to see details, click to follow [the coup de grace here would be taking the pdf footnote text, extract a URL/search query, attempt to download a pdf, auto-parse it, file it, open it, browse back and forward like a browser, maintaining scroll position].

A scroll marker that helps you keep position, a la Readability. Better integration w/ file managers would be good, but replacing features is okay. Managing pdfs is a more specific workflow than managing 'files': a tree-explorer to see all pdfs in which folders (no other files), a tree of recently viewed items (+ nesting for reference-hopping, etc). All the auto-parsing magic (Title, Author..) is great but I would not reinvent it.

milani
It's not about rendering PDFs. It's about a good software that uses pdf rendering engines to deliver great functionality. Your `wants` reflects features missing in current readers.
ig1
Have you seen PDFjs ? (Javascript PDF reader from Moz)
Maximal
My 2-cents: highlighting and extracting my notes is fundamental for me. The best PDF readers that I have found are:

OSX: Skim (with an honourably mention to DevonThink) iPad: PDFExpress

ised
1. PDF's have become a nasty vector for exploits. Thank you Adobe. Need a viewer that can address this problem. My suggestion: a "viewer" that decompiles PDF's to Postscript, deletes any potentially harmful Postscript and then recompiles to a simple, safe early version PDF, all in one shot. I'm working something like this for myself using Ghostscript, but someone with more skill really needs to tackle this problem.

2. Ever tried to read/skim 100's of PDF's in rapid succession? (A common task if you are doing academic research.) It's near impossible. By comparison to reading text documents or viewing images, reading lots of PDF's is SLOW. Need a viewer that can view PDF's as fast as we can view text documents in pagers like less or images in viewers like feh. My suggestion: a viewer that extracts each page as an image and then views the PDF's as a series of images. (Essentially making reading them the job of an image viewer.) The trick is getting the sizing and resolution right, and dealing with hundreds if not thousands of extracted images.

PDF is great for printing. It also looks great on a screen (e.g. for casual, occasional reading, or presentations to an audience). But for reading lots of academic papers or any sort of document in bulk, PDF is absolutely terrible.

Postscript development is really an underappreciated area, I think. How does Apple achieve such lovely text on their displays? I could be wrong but I have always had a hunch their expertise in Postscript plays a role. They were once Postscript pioneers... remember the LaserWriter?

bobisme
The only PDF reader I have installed is Chrome, and I'm pretty happy with it.