Perhaps a digression, but xpdf isn't a sane mouseless PDF viewer?
One of my favorite features is pressing 'q' causes it to quit, solely because I can quickly (It's blazing fast, at least on Linux) open up a pdf, pageup/pagedown through it, and quit the application quickly, efficiently and easily.
So it's not built with the most attractive toolkit, but it does seem to meet your requirements, no? (Okular is close to converting me from xpdf, but it's those extra several milliseconds during startup and the lack of a single key to quit the app that keeps me with from switching.)
You can easily configure all the key shortcuts you want.
I use 'q' for quitting, 'w' for fit to width, 'p' for fit to page, 'g' for goto-line, etc...
I love its minimal GUI interface (by default, C-m for toggling menu visibility!), highly customizable navigation panel, and its 'trim margins' view option.
I also find it faster than xpdf after startup. The only thing that i don't like right now is the slow chm rendering.
Great stuff! Aren't there licensing issues involved with making arbitrary fonts available over the web though?
I'd talked to a few font foundries in the past, and very few would let me use their fonts with a @font-face (or, in their eyes, distribute their fonts indiscriminately).
Understandable concern. For now let's just say that we have a blog post in the queue (part 3/3 of the HTML5 series) that describes exactly how (and to which degree) we store and (re-)distribute fonts.
It would do a huge service to the web if you can get the blessing of foundries for a method that can be used by anyone for displaying licensed fonts using @font-face.
Even creating derivatives of fonts is often restricted, eg, the Adobe Font Folio EULA considers derivatives under the same license as full fonts and the Linotype EULA states "It must be ensured that the Font Software cannot be fully or partially extracted from said documents" (emphasis mine). So, presumably, it would be potentially problematic even if you are extracting individual characters and generating new fonts.
As I understand it, since fonts can't be copyrighted it's purely a software licensing & "piracy" issue. If you've figured out a way we can all get around that it would be a big deal. Maybe you are in a unique situation because you aren't agreeing to an EULA and, instead, simply extracting characters from fonts embedded in documents uploaded by your users. Hopefully you'll share the details of how your attorneys have advised you about what a website's liability is in this area using different techniques
Honestly, on one hand I'm hesitant to comment about it and give a voice or any encouragement to the foundries, but on the other hand it would be pretty shitty if scribd just skirts these issues and plays fast and loose with the situation while the rest of the web is still stuck using the same old font stacks because they can't afford to risk the legal fees.
Fonts can't be copyrighted. They didn't agree to any EULA. All they need to do is convert the font to some other format, and they are legal.
And BTW, this line is ridiculous: "It must be ensured that the Font Software cannot be fully or partially extracted from said documents".
It's possible to embed a subset of the characters, and that's what most require. But to prohibit even a subset? That means you can not use the fonts for any electronic documents whatsoever unless you convert them to bitmaps, which means they are almost useless for PDFs.
I personally would never buy a font with such a restriction unless all I made was print.
BTW, if you want to be legal, just download (pirate) the font file, and convert it to some other format. You never agreed to a EULA, and fonts can't be copyrighted (only the original file can be).
In theory they could complain about the original pirate act, but they would have zero evidence, and any subsequent usage is legal.
Mechanical conversion from one format to another doesn't defeat copyright protection. If you write a book in English and I translate it into French (a change much more fundamental and much more creative than you're proposing), that doesn't entitle me to publish it without your permission.
The shapes of the glyphs are not protectable. The font file is considered to be a computer program, and is protectable as such. Converting it to a different format doesn't get around copyright any more than compiling a C program does.
As gjm11 noted, the font software is copyrighted (as I understand it), the actual glyph design is what can't be copyrighted. If you are still distributing a portion of the copyrighted work (the font software), that would presumably be a copyright violation.
I don't know the actual extent to which copyright protection covers font software or derivatives. There is at least some related case law, but I've only skimmed it and am not an attorney.
That means you can not use the fonts for any electronic
documents whatsoever unless you convert them to bitmaps,
which means they are almost useless for PDFs.
Font licenses always specifically address PDF embedding.
They could start crawling the web today for PDFs that embed fonts infringingly, but they aren't very sophisticated, and even they aren't that stupid.
The situation on the ground isn't going to change from the way its been for the last 25 years -- normal people are going to continue to use whatever fonts, obtained however, and distributed wherever, to whomever, whenever. %90 of font choices made made will continue to be crap.
For the record, I must say, Matthias, you kick butt. I've used your tools (swftools.org) for the last 7 years and found them to be extremely well built and useful. And the "tech support" has been amazing (you just complain about a bug or make a feature request on his mailing list and you get a fix / enhancement within days). Incredible work. Scribd made a really, really smart move bringing you on board to lead state-of-the-art document rendering. Please keep up the great work.
Rotating text by generating custom font files with pre-rotated characters and putting each character in a separate absolutely positioned div element is more "practical" than CSS transforms? I find that difficult to believe.
IE (including IE6) supports transforms, as you point out yourself. Of course they don't do it in a standards-compliant way, but that's true for their custom font support also. What problems did you run into?
Performance issues, among others. IE's DXImageTransform, as the name suggests, renders an html element to an image and then transforms that entire image.
The VML APIs that've been in IE forever are all pretty damn fast individually, but displaying the results slows the rendering down significantly. It's like how straightforward PHP is damn fast as a templating system, but absurdly slow for implementing a high-level templating system in.
Besides, text selection wouldn't work.
Have y'all managed to avoid implementing a GDocs-style system for hiding text underneath bitmaps for selection?
Rendering an HTML element to an image isn't inherently slow; Safari does the same to hardware accelerate CSS animations via Core Animation and it's quite fast there. Of course, it's entirely possible IE's implementation is terrible.
I'm curious about the other issues you had as well.
I think CSS is the easy way. CSS rotation of text looks just fine in any IE and in Chrome, but I'm afraid that Firefox will have to come up with a new text rendering engine to get the text readable when rotated. It looks terrible in Firefox.
I wonder how the character spacing issues are resolved - for example ff and fi generally should overlap and physical print typography uses different characters for this. This is definitely possible in PDF (though I'm not too sure of the details), but then how does this translate back to HTML?
I've been really impressed with scribd when I saw the html text roll out. That, plus the fact that the website had a copy of out of print textbook that I needed raised my kudos value of scribd.
"[...]we need to store three font files: One for Internet Explorer (.eot) one for embedded devices (.svg) and one for Firefox, Safari, Chrome et al (.ttf)."
