To handle certain East European and other languages (Eg. Maori), we have been forced to use embedded fonts with ADE.We generally use a free licensed font which has a wide set of characters such as Gentium. Rather than embed the whole font, we usually create a specific sub-set of characters to keep the file size to a minimum.
The scary part of iPad announcing no embedded fonts support was how did it handle at least the basic extended Latin sets. On the next page you can see our quick and nasty Entities Test Book images which show iPad coming out well (thankfully). Their basic fonts at least support an extended range of necessary and useful characters.
We are also doing an embedded fonts test to see how iPad handles them, hopefully by just ignoring them. If this works we will still keep embedding font sub-sets for ADE if we want to read anything other than English West Europe languages and Malay, and hopefully iPad will just ignore the embedded fonts. I like comparing these on Lucidor as well as it uses the Gecko rendering engine and has the best presentation options available in any e-pub reader.
You can see the images of the entities test results on the next page.
Click on the images for a full screen view
As expected iPad and ADE do well with this. This file actually contains entity statements rather than UTF8 characters. This was to test the performance on iPad as they specifically state character encoding must be used and there must be no entities. So this works out well. As a best practice it can be a wasteful overhead to use entities, but they are defined and allowed in the XHTML specification.
This was a very pleasant surprise and far from the ADE crash and burn. Very nice for the Maori books we are currently producing, and of course a lot of other languages.
Again. Very exciting. The missing character set in iPad is minimal.
This was very impressive on iPad and quite unexpected in many ways. . OK it is a far from complete list of symbols, but there is enough there to say congratulations to Apple at least for have a competent set of glyphs available.
Of course this still doesn't solve more interesting font issues, but is a step forward. Of course there are more tests to do with various CSS properties to see what else will, or wont happen.