You-Read-It-Here-First Forum Index You-Read-It-Here-First
A collection of textual novelties
 FAQFAQ   SearchSearch   MemberlistMemberlist   UsergroupsUsergroups   RegisterRegister 
 ProfileProfile   Log in to check your private messagesLog in to check your private messages   Log inLog in 
If you want to read the articles here, go ahead, just click on a forum and find a thread that interests need to register! If you want to post something... either new or in response to someone here, then click the Register link above. It's free... and it's fun to write your ideas here. You can even create a "blog" by starting a personal thread in the Daily Life Every Thread A Diary section...

The disintegration of the graphical user-interface

Post new topic   Reply to topic    You-Read-It-Here-First Forum Index -> Ideas about software
View previous topic :: View next topic  
Author Message
Site Admin

Joined: 17 Mar 2006
Posts: 712
Location: Oregon

PostPosted: Fri Oct 19, 2007 8:34 am    Post subject: The disintegration of the graphical user-interface Reply with quote

I mean the title in both ways. As in both the falling apart,
and as the lack of unity.

I find the situation has reached a crisis.

Before describing that crisis, I'd like to say a little
bit about my background, and the events that
triggered me to write this entry.

I've been programming computers for nearly 30 years,
and have been involved in many innovative designs,
especially of user-interfaces. I did my graduate work
concentrating on human-computer interaction across
a multitude of what are called "modalities" (text, GUI,
etc.), specially concentrating on spoken language and
face-to-face interaction, as it happens. More recently,
I've been a reviewer for a conference on software quality.
Also, I use a kind of old computer (Pentium II), with
Windows XP 2002, and a dial-up phone line.

These experiences help provide a vocabulary or maybe
a framework to understand the nature of the crisis,
but the problem itself wasn't clear to me until I recently
helped some friends use some web applications, and I,
myself, used an online application to list a large number
of items on eBay.

As a highly experienced user of software, and even a designer
of user-interfaces, I generally find it easy to navigate through
web pages to accomplish some specific task (fill in a job
application, enter items for listing on eBay, etc.), even when
that particular application is new to me. I might not like
the way the application is designed. I may become
frustrated at the number of steps required to accomplish
even a simple goal, or the number of screens that must be
loaded (at dial-up speeds), but I generally persevere and
complete my task.

But my friends, who are not tech-savvy, were having a lot
of problems, and I realize that a lot of the kinds of problems
they were having were ones that I experienced as well,
but which my expertise and perseverance masked in a way.
What I saw as a problem requiring a work-around was to them
an insurmountable barrier.

Now, I could be complaining about the design of the screens,
or the labeling of buttons, etc., but my criticism of the kinds
of interfaces I've been using and helping my friends with goes
much deeper. Web interfaces have degenerated from (usually)
just being "bad" (or difficult to use) to being unpredictable,
which is even worse. Ironically, this has happened as new
tools for improving interfaces have proliferated, and software
producers have increased their attention on "quality".

Ultimately, the crisis comes down to what are called "haptics".
This term refers to the sense of touch, of feeling. I'll be describing
below, problems ranging from the level of the overall design,
to screens, to labeling of buttons etc., but I find the most serious
problem comes down to what might be considered the lowest
level, the most basic action allowed in a graphical user
interface, the mouse click.

I'm not sure I need to catalog just how bad these applications
are. My experiences with the problems with widely used
programs must be the same ones, or similar to ones you've
encountered. Frankly, though, I can't resist complaining,
and in this case, it puts the argument into context.

Case 1: Applying for employment at Fred Meyers
<excerpted in a separate post>

Case 2: Using TurboLister, Auctiva and eBay
<excerpted in a separate post>

Management thinks that if some is good, then more is better, and
decides they'll ask for and even require the street address of prior
employers, even though they'll most likely never use this
information. Want to avoid talking to a lot of prospective employees,
so then filter them down by asking a personality-type question.
Great! If one is good, then 100 must be even better, right?
And what is the worst thing about this whole mess, is that each
of these failures compound upon the others, making the system
literally a nightmare to use. Okay, so its a figurative nightmare.
Sue me.

The thing is, most of this is not new. Maybe when systems were harder
to build more care was taken in getting them right. Maybe with
so many systems in place, the design talent has gotten stretched
a bit thin. Maybe these web applications are built on top of
hard to adapt legacy systems, so the designer was forced into
some of these compromises. The result is the same. From the
highest level requirements and design of a system, down to the
choice of the name on a button, the failures of the system are not
additive, but multiplicative.

What is new is the mouse and the GUI. And the notion of a mouse
click has become increasingly problematical. The problem I've had
in understanding this problem is that I have been too good at internalizing
all the complexity that has been forced into what has to be the simplest
conceivable interface: pointing.

Modern interfaces both create and require users with an entirely new kind
of skill. I'll call it "mouse-savvy", although part of it is "button-savvy"
and "drop-down savvy", etc. What is the meaning of a click? What counts
as a click? What is the meaning of moving your mouse?

These kinds of questions are not the kinds of questions system designers
have been thinking about very much, but they add an entirely new
layer of problems which, of course, compound on the normal and
predictable kinds of failures of traditional analysis, design, and
implementation that have always been with us.

First, there's a mouse cursor and a text insertion cursor. The mouse cursor
changes its appearance at various points. Sometimes its a little arrow.
Other times, it looks like a text cursor. Other times its an hourglass.
Other times, it can be pretty much anything. Okay. I put the mouse over
a text box and it looks like a text cursor, so I can start typing, right?
Well, I suppose you know that isn't true, but it sure frustrates a friend
of mine as he tries to learn how to send email. You know that you have
to click in the text box before you can start typing. But is that even true?
What if you were working in a different window, and then come to this window
and click in the text box? I know this works well a lot of the time, but
I've encountered places where it didn't work. How about this one: you
select an item from a drop-down list. Can you then move your mouse
to the next field you want to work with? The answer is: sometimes.
Sometimes you are just setting a particular parameter, and other times
you're doing that plus you are triggering a web action. And there is no
way to know in advance.

The default folder is not the right one, and the number of icons to display
is not right either. So, I'll just change the folder and the number of icons
and have a new screen displayed that has the information the way I want
it. That's why those options are there, right? Well, while I'm changing the
second parameter, the window blanks out and re-displays the information
about the right folder, but the wrong number of icons. I change it, and
before I can add a search string to another option, the screen is re-displayed
again. Flash! And what happened to that click I just made while the page
was being transmitted and loaded? I'm sure I don't know. Even as an
"expert", I don't know for sure. How is the novice even going to have a
chance against this kind of unpredictability?

Web accesses imply latency. Even for high-speed modems, delays
occur. As an expert, I've internalized which situations I can keep
doing other work, and which situations I'm better off waiting till the
computer has completed its communication and processing. Sometimes
there will be a clue about this: the cursor turns into an hourglass, or
I'll see some activity at the bottom of my browser window. Maybe I'll
see some blinking lights on my modem. But maybe not. And would
it be enough if the clues were a little better, a little more consistent?
Not much. When is a click not a click? Let me tell you. Lots of times.
How is the novice to know this?

And if that weren't enough, what about right-clicking? Or not clicking?
Now, even resting your mouse cursor on top of a screen region, or passing
it through a screen region is enough to trigger an action, or even worse, a
web access.

I'll grant that part of the problem I encounter is because I use a low-speed
connection. Consider the following scenario, though there are many, many
such situations. I go to a screen I've gone to many times before, say,
entering my name and password to access my email. Between the
processing time of the host computer, and the low bandwidth of my
connection, the time it takes to paint this screen (even though it is
very simple) might be between 1 and 8 seconds. I click on the name
field and type it in, then tab or click to the password section. Meanwhile,
the browser has just finished painting the screen and now, helpfully,
it decides to put my cursor in the name section. Now I'm typing my
password at the end, or maybe middle of my user name, in clear text.
Frustrating, yes, but secure? Not at all. Talk about compounding problems,
my browser keeps track of names that I use to fill in the name section.
next time I click twice, or start typing my name, it, helpfully, shows me
a list of names I've entered in the past, including the one that has my
password stuck on the end or in the middle of my user name. If I catch
that, I'm lucky. I know how to delete this friendly default. But if I miss it
chances are good that someone else can just doubleclick on the name
field and see all the names I've ever used to log in, including the one that
has my password mixed in. How secure is that?

My point is that developers aren't using low-speed connections and so don't
even see the intermediate stages their screens go through. On another
screen, the page is displayed, and then somewhat later, a box is opened
at the top, and a status display is shown. Sometimes this status display
needs to expand to be slightly taller. The link that I was about to click on
below has now shifted down. Where I'm about to click is no longer a link,
best case, and is an entirely different link, worst-case. When is a click
not a click?

And my larger point is that, with the proliferation of new web technologies
The meaning of a click, the most basic building block, the most "primitive"
action one can take in a GUI, is largely unpredictable. Add Java, Flash,
Ruby, etc. to the mix, plus inconsistent design, and variable latency,
and you've got a nearly incomprehensible result from something as simple
and as basic as clicking a mouse button. Its no wonder that my friends
can't navigate around an application that should be straightforward. One
friend became so uncertain and paralyzed that when I told him to double-click
the icon in the center of the screen (the desktop), he was paralyzed into
inaction. Do you mean this one? he asked, pointing with the mouse cursor
to the icon in the center of the screen (the only one nearby). The one in
the center, I say, hoping that his accomplishing this will give him a boost
of confidence. But its too late. If I asked him to draw an X in the center of
a piece of paper, he'd have no hesitation, but he's become so full of uncertainty
that he doesn't trust his own understanding of what to do. I can almost hear
his thoughts. Maybe there's some other center of the screen. Better get

When I add these very fundamental problems with the meaning of mouse clicking
to the myriad of failures at the other levels of system analysis, design,
coding, etc. It is little wonder that my non-techno friends hate computers.
The wonder of it is that I haven't stopped to write a book about the failures
of nearly every web application I've ever used. In some ways, my skill at
working around these problems, my adaptability, has blinded me to the
difficulties that my novice friends encounter. It's hard enough to get work done
in any situation even out in the real world. Add in terrible design and
implementation choices by the developer and his corporation and work becomes
a battlefield, accomplishing a simple task becomes a Herculean ordeal.
Add in latency, and it becomes a frustrating task requiring minute concentration
and attention to clues of failure. Add the fragmentation of the meaning of
mouse manipulations and it becomes a chaotic, confidence-sapping nightmare.

I like to describe situations in terms of metaphors for people who are not
technically inclined. Would my friends have trouble finding and pushing the
right button to start a toaster? Not much. Entering some numbers on a
calculator or adding machine? No.

I'm starting to think of modern graphical user interfaces in terms of a
particular image. It's starting to seem like the engine room of an old
steam ship that hasn't been in for maintenance for a very long time.
The engine crew are old hands at keeping this engine going, wiping
off the moisture from the pressure gauge, and banging on the glass
so the needle doesn't stick. They turn a couple of valves that you
didn't notice before, and check some other gauge to make sure the
pressure is okay. Steam shoots out of a nozzle right about where you
were standing a minute ago. It's Thursday, so they need to shut the
engine down to repair some persistent leaks. In walks the new guy,
all clean with his uniform pressed. Can he persuade or help those
guys get things moving so they can get to port before the storm hits?
Not likely. And too bad about that nice clean uniform.

One thing's for sure: this is no way to run a navy.
Back to top
View user's profile Send private message Visit poster's website
Visit the Instant Postcard Collection @
Looking for postcards of that favorite place? Family origins? Or that perfect vacation, except for the photos?
Researching your dissertation? Serious collector? Just looking for something neat?
You've found the right place to add to your existing collection, or to start a new one.
Time Clock

Joined: 11 Sep 2008
Posts: 47
Location: Vancouver, WA

PostPosted: Thu Sep 11, 2008 7:29 pm    Post subject: Last chance Reply with quote

You might appreciate this:

One of the LAN games I enjoy, "Spaceward Ho", has an Armageddon option where if everyone selects it on the same turn, half the universe super-novas.

There are variations on the dire "last chance to abort" warning dialogue boxes given if a player selects this option, but my favorite one was always:

Are you sure you want to
destroy the universe?

[Unsure] [Not Unsure]
Back to top
View user's profile Send private message Visit poster's website
Display posts from previous:   
Post new topic   Reply to topic    You-Read-It-Here-First Forum Index -> Ideas about software All times are GMT - 8 Hours
Page 1 of 1

Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum

Powered by phpBB © 2001, 2005 phpBB Group