During the recent Digitial Publishing Toolkit hackathon, the proposal, made in the context of the INC project, for a using a browser-based editing workflow centered on HTML as the core “document” format was challenged in a discussion initiated by Florian Cramer who proposed focusing instead on the Markdown format. The resulting spirited discussion began focussed on questions of ease of use and features of the respective approaches and ended up touching on fundamental principles of document markup in relation to (digital) publishing.
Can editing HTML ever be something possible for “non-experts”? Or, better said, for those whose skills do not include some working knowledge of HTML. While web programmers, technical editors, and designers often come to understand HTML’s particularities through their work, publishers and copy editors tend not to. Florian argued for the clarity of Markdown’s cleaner “structured text” format over the idiosyncrasies of HTML. The particular needs of the INC, whose reader publications are primarily collections of academic essays with elements such as footnotes and references, figures and captions, was a key issue. HTML (in its latest incarnation HTML5) lacks a firm standard for marking up footnotes (though current standards documents do suggest some best practices using HTML’s existing “link” or anchor element). In contrast, markdown offers an explicit and convenient syntax for specifying footnotes (also through a kind of “best practices” extension to the original syntax by the “multimarkdown” project which has now been widely embraced by many other markdown tools such as pandoc). Beyond just the important details related to footnotes, the ongoing discussion revealed a number of more fundamental questions and principles related to markup languages.
The concept of markup exists for decades now, coming primarily from the intersection of engineering and publishing interests to make generalized methods for indicating structure and editorial intentions (through “tags”) in text documents for the purposes of producing technical documentation in a “multi-path” way. In other words, a flexible system where the same input text files can be used to produce documentation in a variety of languages and/or for a variety of output forms and methods of printing. The SGML standard (Standard Generalized Markup Language, 1986) formed the basis on which Tim Berners-Lee’s HTML (Hypertext Markup Language, 1989), design would form, and led to the parallel development of XML (Extensible Markup Language, 1996) to include applications beyond web publishing.
A Markdown-formatted document should be publishable as-is, as plain text, without looking like it’s been marked up with tags or formatting instructions. While Markdown’s syntax has been influenced by several existing text-to-HTML filters — including Setext, atx, Textile, reStructuredText, Grutatext, and EtText — the single biggest source of inspiration for Markdown’s syntax is the format of plain text email.
from the Markdown “Philosophy”
Markdown was born from the online writing practices around blogger John Gruber (2004). In its very naming, Markdown thumbs it’s nose at the generality and extensibility promised by HTML’s markup preferring instead to focus on human read- and write-ability. With Markdown, writers use indentation, blank lines, and bracket text with typewriter symbols like asterisks, brackets and parentheses according to a set of predefined rules to indicate structural elements in a text: paragraphs, headers, and block quotations, as well to give emphasis to text, make links, and place images. Where Markdown lacks a feature, either custom extensions are added (such as the notation for a footnote) or else HTML markup can be included and which then simply “passes through” the Markdown conversion. The Markdown system is published under a free software license so others are free to contribute and re-implement it’s functionality in their own software, which has led to the format being included in a large number of tools and platforms.
Markdown’s less-is-more approach privileges readability over the extensible functionality of markup, and as a result favors writing in a “clean” and inherently “rawer” stage of development. Ideally this allows text writing to be focussed and maximally editable while leaving open a range of possible options in terms of how the text can then be further manipulated (laid out, combined, reformatted) or used by other people with other tools. Despite its initial design for publishing HTML, it’s minimalist approach produces text sources that are inherently “open to interpretation” and thus suitable to a variety of uses. In a sense, where markup provides its flexibility through an explicit layering of tags, markdown “leaves space” for other uses by omission.
Performing the text
During the discussion, we arrived at the useful metaphor of comparing sheet music to a recording of its musical performance. While technically possible to use audio analysis to convert an audio recording into sheet music, the process is inherently prone to error. Any musical performance in fact includes many more nuances (slight rhythmic variations, differences of interpretation, the expressiveness of each performer not to mention the details of the audio registration itself).
In this metaphor, the audio recording is like an HTML document, and could equally be the WYSIWYG output of a word processor like Word, or a PDF document. While technically possible to convert any of these formats into markdown, the process is inherently likely to produce errors. Like the recorded performance, a published representation is rich, loaded with extra information particular to its context (page layout, interactivity, the subtleties of typography and spacing). These documents lack the simpler structure of a markdown source of the same text (where such subtleties simply cannot be expressed). Like the sheet music, the markdown requires the additional skills of the performer to “bring it to life” in an actual performance. The skills of the designer or web programmer are thus engaged to make form-specific adaptations or interpretations of the “source” text to make a particular rendering.
The limits of markdown & the needs for flexible notation
For all the benefits of markdown, there are some important factors to consider as limitations. By enforcing a kind of “content vs presentation” separation Markdown privileges traditional/academic textual writing where the assumption is that visual presentation flows unproblematically and deterministically from the text. Following the metaphor of musical notation, many kinds of music, from improvisational to serial music, would be ill-served when primarily written and expressed in traditional musical staff notation. Particular modes of expression demand particular forms of notation. One can imagine a multitude of “domain specific” structured text or other forms of writing to complement markdown in a publishing workflow.
The use of Markdown can also reinforce traditional professional separations and hierarchies where writers write, then hand off to a designer who “designs” the eventual presentation. The workflow is unidirectional and sequential as editorial & design work can’t be done in parallel or give feedback to the initiating writing process. It remains an urgent and exciting area of work to consider workflows and tools that allow for parallel and cyclic workflows where the textual and the visual can fully interoperate without a return to the problematics of “WYSIWYG”. A strict separation, while beneficial in other ways, also potentially ignores emerging online practices of writing where the loop of writing, designing, and publishing drives new forms of writing, design, and publication.
The importance of standards
Finally, it remains urgent to enforce/encourage the translation of best practices into open standards through engagement with standards bodies like the W3C. The addition, for example, of the audio and video tags to HTML5 has helped stimulate an explosion of tools and applications for publishing audio and video online. The addition of firmer standards for marking up academic references & footnotes would provide a target for which various structured text and other formatting tools could then aim and facilitate the kinds of interoperability of tools that markup standards have long promised.