PDF Support on Kindle 2
by brian hefele

Despite its breaking of Kindle homebrew, I decided to install update 2.3 last night. Native PDF support, no matter how thin, is a Good Thing. I immediately ran the viewer through its paces, queueing up a variety of PDFs:

So, what's the verdict? Good, not great. First, the bad - no reflowing, no zoom. These two things mean that much of your type will be very small. Articles like What Parents are Saying… which flow text in two columns will be unreadable in portrait orientation. Fortunately, Kindle 2.3 also supports reorienting, accessed easily from the 'Aa' menu. Reorienting Kindle to landscape makes this two-column article difficult to read rather than impossible. This would depend on lighting circumstances… It's easy to read with decent light, but the reduced contrast makes it tricky under lesser lighting. Reflowing would allow the text size to be controlled, and the two-column format to be reduced to one… But alas, there is no support for the feature.

A document made to be printed on letter paper is 8.5" wide, Kindle's screen is just shy of 5" wide in landscape orientation. That's a noticeable shrink, but it still leaves most documents (especially those set single-column) readable. Some other aesthetic decisions such as any font embedded in the PDF or margins (Just the Facts is a transcript and has a large left-side margin to allow for the speaker's name, for instance) will have more of an impact on readability on the small screen than they would have had in print.

Other notable omissions are highlighting, notes, and definitions. These are pretty bad… One major reason to support PDFs is that of research materials - peer reviewed articles and the like. Not supporting highlights and notes makes Kindle nearly useless for this application. Definitions can be worked around by typing the word in question and looking it up straight from the input box. I can understand the lack of highlighting in scanned PDFs. But not having this ability in proper text-based PDFs is inexcusable. Same goes for notes, which could still work on scanned PDFs on a page-by-page basis.

All is not lost for research work - at least search is still supported in the PDF viewer. Of course, this does not apply to scanned PDFs, there is no on-the-fly OCR engine to search images. Scanned PDFs are also very slow to load, with each page taking quite a few seconds to render. Just getting from page one to page two of Patent 1539895 was a tedious exercise. At least it is doable, however, and these documents can be reoriented just like any other. Even in landscape orientation, Patent 1539895 was extremely difficult to read. Not being the best csan to begin with, and set in two columns, the overall contrast of this document made it something that I would not really like to read through on the Kindle.

The good news is that, as I already stated, most of my documents were perfectly readable. Speed was good too - when you load a PDF, you typically get a message saying 'Loading page 1 of xx,' this takes somewhere between two and three seconds. The message made me worry that every page would take as long to render, but this is not the case (excepting, of course, image-based PDFs). It seems that this is just an initial interpretation of embedded fonts or similar. Time to display was about the same for most documents, from my cartoon (which is a pretty considerable pile of vector art) to my résumé (which again has a few vectors, and several weights of a typeface) to Deloused (which is quite graphics-heavy, being the storybook to go alongside a concept album) to normal articles (which in themselves usually contain a font and some supplemental graphics).

Search is also perfectly snappy, and quite useful so long as you're not in a scanned PDF. Additionally, while notes are not supported, bookmarks are, even for scanned PDFs. This does not make the absence of notes/highlights forgivable, but it is something. Overall, this is better than what we had before, and slightly better than I expected. I never expected reflow, didn't really expect zoom (although both of these would make all readability issues go away, for text- and image-based documents), but was kind of surprised at the absence of highlighting and notes. However, I didn't expect rendering to be nearly as quick and accurate as it is, and in fact most of these articles end up more beautiful than native .mobi documents - so long as the text stays a readable size. It is worth noting that 'readable' for me is pretty small, those who are farsighted may find PDF support much less useful. Hopefully, Amazon will continue to provide updates to make the PDF viewer more useful, and while they're at it, epub support would be nice!

Final thoughts: For pleasant reformatting of epub and html files, as well as reflow of PDF, I use Calibre/any2mobi. I personally use the any2mobi command line tool, which works a charm but has a lot of settings to play with, depending on the needs at hand. Excellent work from the open-source community, though the GUI has a typically lacking open-source appearance.

Kindle 2.3 also includes a new screensaver of Ralph Ellison, which is a welcome addition to the existing author-party.

categories: review, tech, software
date: 2009-11-25 18:23:47
blog comments powered by Disqus