Re: [Usability] My second file chooser proposal

[Magnus Bergman <magnusbe algonet se>]

On Mon, 15 Sep 2003 23:01:24 -0400
Sean Middleditch <elanthis awesomeplay com> wrote:

> On Mon, 2003-09-15 at 22:42, Magnus Bergman wrote:
> > On Mon, 15 Sep 2003 20:23:18 +0100 (BST)
> > Alan Horkan <horkana maths tcd ie> wrote:
> 
> > > As far as I knew scan order is entirely cultural and based on
> > > reading.
> > >
> > > Europeans and North Americans and more, read from left to right,
> > > top to bottom.
> > 
> > Are you sure it's entirely based on reading? I've came across
> > theories about the fact that most people are right handed has
> > something to do with it. This might affect we consider right to be
> > forward and left
> 
> Right.  So let's royally screw over the left-handed people, like half
> my family, in all parts of the world, because interface ergonomics are
> based on handedness and the left-handers don't matter.  Just like we
> don't care about the other "handicapped" users in GNOME, right? 
> *that'll* go over well.  If it is handedness that matters most, the
> solution isn't "most people are right handed", it's, "how do we make
> the GUI change order based not on language, but handedness
> selections?".

Gnome already has at least one feature for left handed people and thats
to swap the mouse buttons. I know a couple of people who are left handed
and used this feature (not in gnome but in another popular desktop
environment). But since most computers they used had stopped bothering
and switched the feature off on their own computers too. So left handed
people are already screwed over by the "common practise".

What I meant was that the evolution of a culture does not take
minorities (like left handed people) in consideration in cases like
this. GUIs are (and I think) should be based on the result of the
culture. So it's the culture that screws us over (yes, I'm handicapped
too).

> > being backwards. Another thing that might have had an impact on the
> > western culture is Christianity which had a concept of right being
> > good and left being evil. That was taken quite seriously some
> > hundred years ago and might have had an impact cultures with a right
> > to left locale as
> 
> Do you have a source of this?  It truly sounds like something pulled
> from a behind...

It was first told to me by a priest who guided me through an old church.
He said that the church had one good and one "less good" side. The
origin of this, he said, was that the devil would come from the north
and thats from the left since the church is pointing to the east. He
also mentioned some examples of superstition based on this, but I don't
remember that now (it was like ten years ago).

In some extension this still lives on today. Like that the "left hand
path" (sinistra vivendi) is the evil path to walk down, the path of the
devil. And the word "sin" comes from "sinister" which is the latin
word for left. It has also been suggested that the adjective "right"
comes directly from direction "right" since "right is the right way to
go". I think it would be easy for anyone to check this up.

I get a feeling I'm WAY off topic now..

> > well. Many years of western influences (like video players and comic
> > books) and globalisation in general has for sure influenced other
> 
> Comic books?  All the graphic novels *I* read are right-to-left. 
> Comic books w/ right-to-left orders are only in cultures where the
> language goes right-to-left.  Comic books aren't the influence,
> they're another sympton of the reading order of the language.

Yes, to a small extension Japanese culture (like mangas) has an
influence on us (making us *slightly* more used right-to-left layouts).
But I believe that our culture (including comic books from DC Comics and
Marvel) has bigger influence on other cultures, making them more used to
left-to-right layouts. The reason I mention comic books (as an example
of many things) is that it keeps its layout even after translation.

> As any scientist will tell you, *nothing* about correlation implies
> causation.  You need proof and research, not random guessing or
> convenient fact inventing.

Yes, I see your point. There was a lot of guessing and speculation, but
I wouldn't say random. Well the examples were random, but that kind of
lies in the nature of examples.

I'm just not a science kind of guy and I didn't really intend to
anything. Rather to justify that the following assumption has
not been proved: "People of all cultures who read from right to left
scans visual information from right to left and like prefer the layout
mirrored". It is likely to be true, but not proved AFAIK. (Please tell
me if it actually has, so I can sleep at night instead of thinking about
this).

> > cultures too. And perhaps those things has affected their scan order
> > too. (If anybody know about some scientific research at least
> > related to this, don't hesitate to tell me.) What I really want to
> > say is that it's PERHAPS not safe to assume that it is appropriate
> > to "mirror" the gui if the direction of the text is right to left.
> 
> Unfortunately, it is quite hard to work going in one direction when
> the text moves in the other.  If you don't believe me, try it.  Layout
> a dialog to work right-to-left and keep using English.  It gets a bit
> awkward, doesn't it?

Yes, I know. I have used Windows a few times and dialogs with random and
awkward layouts are quite common there. I begin to suspect that we have
slightly different ideas of what left-to-right layout means. What I mean
is that the widgets should be placed you interact with them. For example
in a FOSD you first select the directory, then the file (it makes no
sense to first select the file then the directory), therefor the widgets
for doing this should be placed in that order. GTK has got this right,
most other GUIs I've used hasn't (Windows 3 to mention one). Also
buttons for going forward (like OK, Yes and Continue) should always be
placed to the right of buttons for going backward (like Cancel, No and
Back). Again GTK has got this right but most other GUIs hasn't (Windows
to mention one). Increasing and decreasing things should correspond to
right and left respectively. I've only seen a couple of designs going
against this luckily.

> If you're going to force users to use an order that isn't the same as
> their language, you'll probably need to force the culture to swap
> their language around as well.  ;-)

No, that was almost the opposite of what I intended. My intention to not
(necessarily) force the order to be the same the language. At least
until it is proved that this is an adequate behaviour in each and every
locale. From a more technical point of view my suggestion was to
eventually separate the layout-order into text-layout-order and
gui-layout-order. I must admit that I wasn't aware of GTKs features of
rearranging the layout than I began this discussion. And by now it has
been narrowed down to a very minor detail which nobody really cares
about (at least not you and me or anybody we know or care about).

> > > I was quite surprised to see that the Asian version of Microsoft
> > > Windows I saw had everthing left to right, but directional text
> > > support is difficult and Western influence is strong.
> > 
> > You don't have some pointers to this? It would be very interesting
> > to look at. It would help to know what their Windows verison look
> > like if I find some Asian people to discuss this subject with.
> 
> One thing to remember is that the Win32 API does *not* make reordering
> of dialogs easy at all.  Where GTK can reorder certain things on the
> fly, Win32 forces manual layout of everything.  I can't say for sure,
> but I'd imagine that hinders reflow of dialogs a *lot*.

I know that GTK is a lot better than the Win32 api from the programmers
point of new, I've tried them both. I don't think we have much to learn
from windows. I'm planning to talk to Asians about the their windows
version (and usability in general). It would be nice to know what
the subject of the discussion looks like.

> > > > I'm not personally aware of any languages that flow bottom to
> > > > top, but there are plenty that go right to left.  If a dialog
> > > > only works well in left-to-right scenarios, it could be less
> > > > than ideal.
> > > 
> > > Japanese goes in columns from top to bottom, right to left.
> > 
> > Both text in columns and rows exist. I think rows are slowly taking
> > over. I have read in several places that humans are faster at
> > scanning rows than columns (and AFAIK that doesn't differ between
> > cultures). This is mentioned in most interface design guidelines, i
> > think.
> 
> Has to do with eye movements and such, I believe, yes.  I know I've
> seen study/research on this before, and believe you're correct.
> 
> > It might be interesting to Microsoft interface design guidelines
> > also says that menus should only contain a few items (since they are
> > relatively difficult to scan). Further it recommends that two level
> > menus should be avoided and that more than two levels must never be
> > used. After coming up to this (which they probably spent a lot of
> > resources on) they design the start button as a central part of
> > their GUI!
> 
> On one hand, they've got some good points about menu usage.  On the
> other hand, I haven't seen a graphical UI that *doesn't* have large,
> deep menues in a couple places at least.  I don't think we can rip on
> Microsoft for doing what we and everyone else do.  ;-)  (We have
> enough*other* reasons to rip on them, anyways ^,^ )

But they started it. And now almost every GUI has that annoying start
menu (luckily it can be removed in Gnome). I can think of a few
applications I use which are more than two levels deep, but in those
cases a consider it to be bad design. The bookmarks menu tend to become
too big, this is somewhat the responsibility of the user but I consider
it to be a slight problem. Can you give me any example that convinces me
that menus sometimes *has* to be large and/or deeper that two levels.

By the way, apparently Microsoft consider this a problem themselves
since they have added a feature which hides infrequently used items from
large menus. Which I think is working around the problem rather than
solving it.