I recently took some of my writing time to create a makefile (Novel Makefile) to manage work I hope to be doing on a new novel project. I've started outlining and researching the story in earnest after having spent the past few couple of years talking about it, and I think writing will commence soon. In another post I'd like to write up some thoughts on the tooling and technology of writing non-technical/non-manual long-form.
This post, drawing from the spending some time buried deep in production is about the state of (conceptually) longer form work in digital mediums. Or, at least a brief commentary on same.
The tools that I use to write technical materials do all sorts of cool things, like:
- provide instant cross referencing,
- generate great indexes, and
- automatically generate and link glossaries.
This is not particularly unusual, and in fact Sphinx is somewhat under-featured relative to other documentation generation systems like DocBook. 
And yet people publish ebooks that virtually identical to paper books. Ebooks seem to say "*this electronic book is the best facsimile of a paper book that we can imagine right now,*" while totally ignoring anything more that a *hyper*text rightfully might be.
I enjoy reading ebooks, just as I have enjoyed reading paperbooks, but mostly because ebooks basically are paperbooks. I've written posts in the past challenging myself,and fiction writers in general, to actually do hypertext rather than recapitulating prior modalities in digital form.
At various points I've thought that wikis might be a good model of how to do hypertext, because the form is structurally novel. Any more, I don't think that this is the case: wikis are unstructured and chaotic, and I've come to believe that the secret to hypertext is structure. There are so many possibilities in hypertext, and I think much experimentation in hypertext has attempted to address the chaos of this experience. This does serve to highlight the extent to which "the future is here," but it obscures the fact that structure makes narratives understandable. Think about how much great, new, innovative (and successful!) fiction in the past decade (or so) is not structurally experimental or chaotic. (Answer: there's a lot of it.)
The not-so-secret of hypertext, is (I suspect,) tooling: without really good tools the mechanics of producing a complex, interactive textual experience  is difficult for a single writer, or even a small group of writers. Most tools that manage the publication and writing of text are not suited to helping the production of large-multi-page and mutli-component texts. One potential glimmer of hope is that tools for developing programs (IDEs, build systems, compilers, templating systems, introspection tools, pattern matching, etc.) are well developed and could modified for use in text production.
The second non-so-secret of hypertext is probably that hypertext is an evolution of text production and consumption, not a revolution. Which only seems reasonable. We have the technology now to produce really cool text product. While tooling needs to get better, the literature needs to do some catching up.
Lets start making things!
|||It's not that Sphinx is "bad," but it's clearly designed for a specific kind of documentation project, and if you stray too far outside of those bounds, or need formats that aren't quite supported, then you end up without a lot of recourse. Having said that, the "normal," well supported and most projects--documentation or otherwise--will only very rarely hit upon an actual limitation of Sphinx itself.|
|||To be clear, I'm partial to the argument that today's computer games, particularly role-playing games, are the things that the futurists of the 1960s and 70s (e.g. Theodor Holm Nelson) called "hypertext."|