The history of markup and WSYWYG (or, as a friend said, WYSIAYG - what you see is all you get) is fascinating.

The early markup systems (runoff and its derivatives like troff, nroff, IBM's SCRIPT) focused on manipulation of representation.  Normal, bold, italic, font size, justification and centering, and so on, were the vocabulary of the old systems.  These systems, to me, were assembler language for contemporary phototypesetters.

In the late 1970s and early 1980s we began to get systems that, as Douglas noted, could express the logical structure of documents.  GML and SCRIBE were my first exposures to this way of thinking and they made life much much better for the writer.

The standards work that created SGML went a bit overboard, to my taste.  The only really serious adopters of SGML that I can think of were the US military, but there may have been others.

Along the way were some fascinating attempts at clever hybrids.  Mike Cowlishaw built a markup system for the Oxford University Press back in the early 1980s on secondment from IBM.  It had a rather elegant ability to switch between markup mode and rendering mode so you could peek at how something would look.  I know that it was used by OUP for the humongous task of converting the OED from its old paper-based production framework to the electronic system that they use today, though I have no idea what the current details are.

The hybrid model is not dead, by the way.  The wikimedia system adopts it ... you may edit either in markup mode or in WSYWYG mode, though I find the WSYWYG mode to be frustrating.  Sadly, the markdown stuff used by wikimedia is pretty annoying to work with and the rendering is buggy and sometimes incomprehensible (to me, at least).

Making a strong system that includes inline markup editing AND WSYWYG editing with clean flipping between them would be fascinating.  Sadly, the markup specifications are flimsy and the ease of creating crazy markup like <h1><b>blah blah</i></h2> in edit mode makes for some difficult exception handling problems.

Marc
=====
nygeek.net
mindthegapdialogs.com/home


On Sun, Jun 2, 2024 at 8:40 AM Douglas McIlroy <douglas.mcilroy@dartmouth.edu> wrote:
> Were you surprised when folks settled on word processors in favor of markup?

I'm not sure what you're asking. "Word processor" was a term coming into prominence when Unix was in its infancy. Unix itself was sold to management partly on the promise of using it to make a word processor. All word processors used typewriters and were markup-based. Screens, which eventually enabled WYSIWYG, were not affordable for widespread use.

Perhaps the question you meant to ask was whether we were surprised when WYSIWYG took over word-processing for the masses. No, we weren't, but we weren't attracted to it either, because it sacrificed markup's potential for expressing the logical structure of documents and thus fostering portability of text among distinct physical forms, e.g. man pages on terminals and in book form or  technical papers as TMs and as journal articles. WYSIWYG was also unsuitable for typesetting math. (Microsoft Word clumsily diverts to a separate markup pane for math.) 

Moreover, WYSIWYG was out of sympathy with Unix philosophy, as it kept documents in a form difficult for other tools to process for unanticipated purposes, In this regard, I still regret that Luca Cardelli and Mark Manasse moved on from Bell Labs before they finished their dream of Blue, a WYSIWYG editor for markup documents, I don't know yet whether that blue-sky goal is achievable. (.docx may be seen as a ponderous latter-day attempt. Does anyone know whether it has fostered tool use?)

Doug