Hacker News Reader

Comment by xavixavi | Hacker News Reader

What PDF readers on linux are 'best recommended'? I compared acroread with evince and kpdf in the past and found that acroread was the most parsimonious in use of resources (especially RAM on large documents). Have not repeated the comparison recently, so will be glad to hear about recommendations with more recent experiences with large documents with images.

View on HN

Personally, i prefer "zathura". besides filling out forms it can do everything, with a very sleek ui and outstanding keybindings (vim-esque). I encourage you to try it :)

It's a fantastic piece of software. It can use different backends (mupdf or poppler) and it can also deal with DjVu files.

Just some general info: Evince and Okular (the successor to kpdf) both use poppler under the hood (which is maintained/released by the Okular maintainer), so their performance and behavior will be somewhat similar, though you can make a difference at the app level with things like tiled rendering and smarter prefetching. I'm not on top of how the two compare there, though. Okular also has pluggable backends and there's an in-development mupdf backend around.

Firefox's builtin pdf.js reader has really matured and I switched to using it as my primary reader. It's even good enough for me to read books since it remembers the location I'm on, even if the browser process dies horribly.

I use mupdf. It has a quite minimal user interface but it's fast and has really good rendering. It's written by the ghostscript guys.

I also use Okular. It has recently gotten support for keeping multiple documents in tabs rather than separate windows. However, I find it to be slow when searching big documents and slow to render large graphics in a page.

llpp is lovely: http://repo.or.cz/w/llpp.git

I like apvlv. Very vim like. you can even split with :s and do a lot of cool stuff.

One that's easy to overlook - just open in google chrome.

I absolutely dislike the pdf viewer in Chrome. It fails at even the basic things a pdf reader should support. E.g., it lacks a concept of pages. You can't jump to a certain page. You can't even know the total number of pages. It doesn't support the pdf's TOC. For any pdf which contains more than one page it is therefore useless. I can't understand how the Google folks ever thought this was a good idea.

I prefer Okular or Evince for Desktop viewing and pdf.js for in browser viewing. pdf.js is a bit slow when loading the pdf. But the user interface is quite nice in my opinion.

On the other hand its one of the (or the? at least last I checked) only pdf readers on Linux which supports subpixel rendering of text and other vector content, which makes documents much more easily readable in my opinion.

Reply on news.ycombinator.com