[net.text] what WYSIWYG systems are and aren't

reid@glacier.ARPA (Brian Reid) (12/10/85)

Two points that need to be made:

(1) The important thing missing in most WYSIWYG systems is structure.
(2) When most people say that they want WYSIWYG, what they really mean is
    that they want realtime prettyprinting, not realtime proofing.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
(1) Most WYSIWYG systems don't let you get your hands on structure. Consider
    VisiCalc and its descendants. If you have a cell with the number "14",
    where did that 14 come from? Is it a constant 14, or is it the result of
    evaluating some formula? VisiCalc programs have a "show formulas" mode,
    in which the thing that you look at on the screen is not the data, but
    the reason that the data is there.

    In a document processor you need that kind of thing, too. It's not
    enough to show you that the heading is centered in 16 point Optima. You
    have to show WHY the heading is centered in 16 point Optima. The problem
    with having a "mode", like VisiCalc, is that there is no easy way to
    divide up the screen, such as into cells, so that any given area shows
    either content or structure.

    There is no sound theoretical reason why it is not possible to build a
    WYSIWYG editor that shows both content and structure. However, if you do
    build one, then what you will end up with is a screen display that looks
    an awful lot like what I see on my screen when I edit my Scribe
    files--text with embedded markings that tell me about the structure of
    the text. I am not aware of any multiple-font WYSIWYG system that has
    ever been programmed that (a) lets you see what you are going to get,
    (b) lets you see, on demand, the structure behind what you are going to
    get, and (c) lets you do a decent job of things like automatic
    numbering, footnotes, cross reference, index, table of contents, table
    of figures, etc.

(2) People mean very different things when they say they want WYSIWYG. Some
    people want a faithful screen update after every keystroke. Some people
    want a "prettyprinted" rendering of the document, not necessarily the
    way it is going to come out, but looking all nice on the screen. Others
    would like to be able to pull a fast proof--previewing a page on the
    screen before printing it. Yet they all use the term WYSIWYG for it.

    Me, I don't want any of that. There has not yet been built a CRT screen
    that is accurate enough to show me the level of proofreading that I want
    to do, and until there is such a beast, perhaps 4000 by 4000 pixels on
    the screen, I don't want to waste my time and burn my eyeballs looking
    at some fool "what you're going to get" image on the screen. All it's
    really showing me is the line breaks and a crude approximation to
    spatial positioning, anyhow. I've solved this problem for myself by
    buying an Apple LaserWriter of my own, which sits right behind my head
    as I type this; I can get a proof on it in a few seconds, much more
    comfortably than I could get on a screen.
-- 
	Brian Reid	decwrl!glacier!reid
	Stanford	reid@SU-Glacier.ARPA