I'd dispute the entire premise of the article - I have not heard anyone say that PDF is a legacy format. It's quite clearly the preferred format for putting physical documents into a digital format.
I've heard plenty of people say that PDF should become a legacy format. Maybe I've been doing it wrong all these years, but PDF is a pain in the ass to work with from a code point of view.
I love PDF. Just last night I synced the revised report on Haskell 98 (as a PDF) over to my iPad and stayed up late reading it in iBooks. Having the format baked into a platform helps, I'm sure (anybody remember the days they called quartz "display PDF"?) so anybody who only knows about Acrobat Reader is forgiven for the bad impression they have of the format.
Indeed. PDF is a great experience on OS X, and a decent one on iOS. (While its main UI is rather cluttered, I feel GoodReader does a better job with PDF than iBooks, solely for the margin cropping and table of contents support.)
I found that switching pages in GoodReader was pretty painful, but this seems to have been fixed in the most recent release. (The main UI is a little weird, too, but it gets the job done.)
iBooks PDF is decent, especially for a 1.0 product. They do need to add TOC support and 2-up would be nice. But my biggest issue with iBooks is organizational, with only 30 books, my collection is hard to manage. (I'd like to keep separate lists of "to read", reference, "already read", etc.)
Ideally, Apple would also add PDF/epub podcast support for magazines. I imagine Saveur (PDF) or a SciFi magazine (epub short stories) automatically feeding into iBooks.
epub is good for novels and the like, but I appreciate PDF for stuff where formatting matters. (e.g. Autumn Omakase http://www.tastingmenu.com/autumnomakase/default.htm ) PDF is also helpful for stuff that iBooks epub can't handle well, like technical documents.
On OSX, I really like Preview - it's a dream to use. My only complaint is that it's a little sluggish with JBIG/JPeg2k PDFs. (Scans of old books.) I think it's possible to render these pdf files as quickly as djvu files, but Apple hasn't yet made the necessary optimizations (rendering jpeg2k at reduced resolution).
The author argues that PDF isn't obsolete, because it's an excellent replacement for paper forms that need to be reproduced exactly. This is true. But I think these paper forms are themselves becoming obsolete as bureaucracies switch to online forms.
We use PDFs for a different purpose. We use them as a intermediate format when processing and archiving documents from other software. We need to support 100+ file formats, and we need to accurately preserve the underlying appearance.
PDF is great for this kind of "visual capture," because it's the most common format that can preserve arbitrary formatting and graphics. And unlike TIFFs or PNGs, PDFs can be searched and highlighted, because they optionally contain the full text of the document.
The problem isn't with PDF itself, but more PDF viewers/creators are such a super-vast-minority behind a single company's product that it is hard to imagine it as anything other than a bad/doomed technology. But hopefully GNU PDF can change that (and Gnash too).
The problem isn't with PDF itself, but more PDF viewers/creators are such a super-vast-minority behind a single company's product [...] But hopefully GNU PDF can change that
I don't see how. They are behind all the other free software PDF implementations. GhostScript even has a new renderer (mupdf) that is ahead of GNU PDF.
In the same way that Internet Explorer's 98% market share was problematic, and that the web is a better place with competition/awareness of browser choice. As it is now, few consumer know of options other than Reader. (And anybody who's had to use FoxIt for professional-PDF use knows it isn't a working solution for designers.)
FWIW, I think a majority of Mac users don't actually install Adobe Reader, since Preview handles them natively and fantastically for nearly all purposes.
Adobe should have kept PostScript (or the clear-text based first versions of PDF.) Shortening PostScript, not to speak of going the binary route, was premature optimization.
