Typography
There’s currently a lack of advanced typography support in both formats and readers. Hyphenation for example should be a basic feature in every reader. Formats based on XML should extend their support for advanced typography, if we want something as good looking as PDF/print, yet reflowable.

Unique ID
Creating a database for unique IDs is a pretty tough work. You also need an easy way for anyone to get these unique ID through web services (REST/SOAP), avoid or group together duplicate entries, a good search system that will recognize the book name even if it’s slightly mispelled etc…
I understand why having a unique ID system widely available would be fantastic, but it’s something that will need a lot of time/work.

epub and the net
Yes, it would be very easy for most browsers to support natively epub and this would help the format.
I would also like an easy way to “widgetize” an epub file. With more and more blogs, social networks and forums, I’m sure that a lot of people would love to embed books or sections of a book to display them on their own webpage.

Strange because we immediately think in terms of an XML structure of the book’s contents and that I would dismiss. Horses for courses, DTBook, DocBook, TEI etc.,. only have XML in common, but little in terms of structure – I believe this is a good thing. Not just the books but the different reading needs suggest very differnt forms of markup.

TYPOGRAPHICAL AND REFERENCE STANDARDS.
However, at the micro-level, of attributes, entities, ids, internal referencing and application of common elements we do need, especially for e-books, tighter standards and consistant application.

Part of this could be resolved by better tools.
For instance: quotes should in my opinion be totally replaced by “q” elements, let the stylesheet look after nesting. We should either have the single closing quote always representing an apostrophe, or use the “apos” entity, likewise with the treatment of hyphens, softhyphens, soft-new-page markers within the text — these details should be made consistant, clear and easy to follow.

If we could get typorgaphical issues sorted out, content at least could be treated consistently, even if the tagging cannot (which I think is best solved at another level – transformation tools).

UNIQUE ID GENERATION
In a couple of days I will be announcing a simple system of generating Unique Identity Codes, this is something each and every publication should include.

A standard here is much needed. For instance, each author should have an ID, each Work, each Publisher, each Translator and every Edition. E-books should always be referenced by their Edition ID for this can reference all the relevant information pertaining to them.

EPUB allows for inclusion of unlimited ID space in the Dublin Core RDF section, only one of course can be the reference to the work in the main. It would be nice to get this right from the begining.

For this to be useful however, we need a world registry, it need no be very complex and any University library could be its home.

This week I also am hopeing to talk to a University Librarian to see if it will be willing to act as respoitory of catalogues of e-book editions. I hope if this fails, that someone else might persuade some other institution to take responsiblity.

Generating world wide unique id codes, anonymously, with the added advantage of being able to generate on a local machine endless extra unique codes (once a net code is obtained), has some importance for e-books and much else. Next week (with the help of a small company) I hope I will have a wiki, the basic codes and site for generating the codes, up and running (it is a free service).

EPUB AND THE NET
The thing that strikes me about epub format is its simplicity. An unzipped form on a website requires nothing else but a browser for online reading – zip it up and it is an e-book. Maybe browsers can handle Zipped directories, and this would simplify things even more, an internal href would then open the book, download the directory and delivers the e-book ready to read.

The only major change I would envisage for the format is the acceptence of any well formed XML as the base file – hence XHMTL 1.1, DTBook, TEI, DocBook could be included.

Publisher standards are to my mind a different thing. How quotes are dealt with, IDs, referencing, when entities should and should not be used, in English and similar languages whether UFT8 should be mandated or not etc., nuts and bolts stuff.

Best practice standards, whether dialogue in novels should have speaker references in their attributes (for audio), and how to do this that is largely consistent to most XML markup schemes, or should such audio prompts be set externally in a style sheet (I tend to favour the latter).

PS

Hadrien, given your excellent work on feedbooks, I would be very interested in your opinions on these suggestions.

Along our support for epub and IDPF, Feedbooks will also study
microformats and other similar ways to create data/service standards.
We already have a standard file format for e-books, the next step
might be a standard markup for e-books published on the Web (something
we could easily turn into an epub file) ?

EPUB is such a clean and clear format, that at this stage the standard has clear advanatages. Whatever happens to ADOBE may only really on effects the fortunes of Digital Editions – I think we are fairly safe in that direction.

However, CSS is so limited typographically, I have long pinned my hopes on dynamically generated PDF through stylesheets.

The promise of reflowable PDF may be a real blessing, but the tenuous preservation of XML markup in PDF also needs to be addressed and ADOBE has not been as good as it should on this. Reflowable PDF may be perfect solution, or a perfect curse for e-books. It is wait and see.

However, alternatives are not limited. EPUB specifies SVG graphics, and these could supply just the typographical control needed to produce nice looking pages on e-paper devices (which I believe is an important key to the growth of e-publishing).

If one door closes, or becomes too hemmed in for practical and open use, perhaps another can be forced to serve the purpose, perhaps even better a better alternative?