I realized that my schpeal last time on reading hypertext sounded more like a complaint than an endorsement of a way of writing, but I don't mean to come off like that. While I think it's true that the advancement of technology: better portable text display, better design practices and standards, better fonts, better industry adoption, and so forth, would help digital text, I don't don't think we can blame it all on the technology.
I should also say, as a disclaimer that I'm not particularly interested in the artistic implications of hypertext. This isn't to say that they aren't there, more that, I don't have any particular expertise or experience aside from, of course, the ongoing experiment of TealArt.com, and my own experience as a writer in several different contexts. Actually I think it's more complicated than that, but I think that would need to be unpacked a bit more in a different context. What drew me to this topic, is more the prospect of paperless publishing and communicating, in concern with discussions about OpenAccess, a nagging interest in ergonomics, and not "oh cool, look at this new way to be creative." Just so you know.
As I've been thinking about it, the discussion of "writing" or producing hypertext is so very closely tied to the publishing of digital, and the discussion is further muddied by
One of the things that people like a great deal about dead tree versions is the the fixty of type on a page. I've learned to remember passages in books by location on the page, and while the ability to search ameliorates this slightly, this cognitive ability is something that I think hypertexters would be wise to work with. While there are ways to break up text: images, column(s) and width, generally I suspect that people will engage the most with your text if they can absorb it in chunks that are about 500-750 words. This works out to be 3-4 good paragraphs, and all the usual edicts and suggestions regarding white-space remain relevant.
I'm also concerned with the issue of citation digital text. Not only is it hard to give precise citations for digital text, but footnotes and what not are equally difficult to produce in an authentic and appropriate format. I think we might be inclined to trace this at least in part to the development of the web (and more importantly) mark up languages (HTML/XML/etc) by people in the sciences where close textual referencing is not particularly common. Amongst bloggers, citation is achieved through linking, often to wikipedia (another debate for another day), and while this model increases the interconnectivity of the 'blog community, but forces curious readers to jump through more hoops than perhaps is necessary, and is downright inconvenient if you're citing/linking to a longer page. I suspect the next version of Markdown, will have some sort of allowance for footnotes, but I think there's still a lot of room to think about how the practice of citing others' work gets translated into digital formats. I like the MultiMarkdown, solution, but have ended up using the PHP Markdown Extra Format and some sort of XML format probably the solution. I'm starting to realize that so many of my problems could be solved with a much greater larger understanding of XML and some basic Perl or Python abilities.
But that's a diversion, and I'll bring us back around to the question of the week: How do we, as digital text creators, produce text (words etc.) that "work" best digitally. I definitely think manageable chunks are the way to go, and the weblog framework definitely pushes us towards a natural serialization, but thats as much cultural as anything. My second series of thoughts relates to an idea of genres. The kinds of things that seem to be the most succesful in digital formats are non-fiction things (manuals, essays on blogs, scientific/academic articles, correspondence etc.) and even though I think blogs might lend themselves particularly well to fiction, there isn't a large community that I'm aware of, though I'd love to be corrected. I think length plays a role here, as does the general reticence to "curl up with a computer" for pleasure reading, but I think there's something larger here. Part of it is without a doubt the fact that there's something about the way we think of novels (ie. the structure and story telling technique) that is very tied to the book, and I also think that in a way the boundary between truth and fiction is somewhat blurry online, because of the way we tend to represent ourselves. Not that the same issues aren't present in non-digital formats, but that we're more sensitive to them in digital formats.
I hope this these fairly random thoughts can provide some inspiration. I'd like to be more coherent, and more focused with regards to some of the other subjects on my list to cover. I know that there are folks out there reading this, and I also know that I'm just a guy with a domain, and an overactive sense of opinion, so I'd love to hear what you think about these ideas/subjects.