High-quality typography with good justification, kerning, etc. -- although just how high that quality is depends a lot on the workflow of the publisher. Ooops -- I almost forgot ligatures.
Much better contrast ratios than we see on current reader hardware.
Cheap books have better resolution, and high-quality printing (glossy magazines, textbooks, etc.) have incredibly better resolution than current reader hardware. On the order of 300DPI for cheap books, up to ~3000DPI for the high-quality printing vs. 180DPI for current readers.
A wide variety of formatting options (chosen ahead of print time, of course): multi-column, hand-optimized placement of figures, charts, and graphs, Drop-caps, and on and on.
Note that many of the above things could be improved in existing reader software without undue effort. For example, TeX does decent-but-not-outstanding justification and kerning (in part by doing decent-but-not-outstanding automatic hyphenation). Since TeX is free software, it ought to be possible to re-purpose their code (or at least their algorithms) for these tasks. Similarly, Apple has demonstrated that automatic use of available ligatures is reasonable to implement in modern software. That's not open source, however, so either eBook manufacturers or Adobe would have to replicate Apple's results.
Contrast ratio and resolution improvements depend directly on the hardware guys, though.
Xenophon
|