WYSIWYG and usability (was: Re: Rapidly navigating buffers using search

[Peter Flynn]

bolega wrote:
> On Jul 8, 3:36�am, David Kastrup <d...@gnu.org> wrote:
> 
>> In a way, it is a losing battle. �People expect software to just work
>> without reading manuals. �95% of all Word users, for example, create
>> their documents by mostly visual manipulation of their text without
>> having a clue about underlying structures like references, style sheets
>> and so on. �The result is unmaintainable crap, but they would not know
>> better. �Word tries keeping up in this battle of computer illiteracy by
>> doing things like enumerations, styles and so on "automagically",
>> second-guessing the user, and the user tries second-guessing Word in
>> order to get around that.
>>
>> It is an escalation of mutual cluelessness. �The more userfriendly a
>> piece of software becomes, the more this becomes a problem for
>> _competent_ people willing to learn about their tool. �At least Emacs is
>> at its heart and in most of its modes a WYSIWYG system with regard to
>> the actual file contents: regardless of the crap people do, what ends up
>> on disk is that what they see on their screen.
>>
> 
> Rare pearls of wisdom ... from DK.

Actually, DK posts a lot of wisdom, but not all is as quotable as the
above :-)

> The new interface of office 2007 with tabs instead of pull-down menu
> is a lot better in terms of visual throughput.

I'm not familiar with that term. The ribbon is an experiment in
interface usability (for doing which, Microsoft is to be congratulated,
regardless of the outcome, and regardless of whether you think the
ribbon better or worse than the old-style menus). Unfortunately it is
based on the tyranny of the majority: it's fine for the unthinking user
that David describes above; it is almost certainly suboptimal for his
"competent user".

> A wysiwig editor with a good markup or definition language can go a
> long way to educate the user about the underlying features while at
> the same time providing user-friendly convenience.

The majority of users don't want to be educated about underlying
features, or indeed about anything outside their field. They just want
the system to produce what *they* *believe* to be right, whether it
actually is right or not. As far as they are concerned, if it looks
pretty, it's right. The fact that it may be unusable, obsolete within
days, unreadable, or whatever, remains forever beyond them (oddly, even
when they have to pay to have it put right afterwards). Fortunately,
most of the material concerned is transient, ephemeral, or simply
unimportant.

However, in the middle (between those users and the "competent user")
there will be users willing to learn how to do it right and avoid
mistakes and unnecessary expense. But my own research is showing that
these users do still insist on a synchronous typographical interface
(what they call WYSIWYG, even when it's not, quite). What I am
attempting to measure is how far you can go towards retaining the visual
manipulation of the text before the interaction descends below the bar
of mutual second-guessing that David describes.

> Things are certainly progressing in this direction.

Yes, slowly.

> I have not used LyX but I have heard that it is wysiwig with the
> option of viewing code in various representations.

It's near-WYSIWYG. They describe it as WYSIWYM (Mean; implying conscious
intent). For users in my middle group above, who require Instant Textual
Gratification [tm :-], it's a valuable tool in that it has made it
clearer where some of the boundaries lie.

///Peter
-- 
(Followups set to c.t.t)