GUIs (was Re: A comment on Slashdot)

GUIs (was Re: A comment on Slashdot)

[cpt.kirk]

There seems to be a lot of mis-understanding about Grphical User
Interfaces (GUIs). The first being that they are purely (or even mostly)
graphical in nature. The next misconception is that command line
interfaces don't have any graphical content to them.

I will point to how our beloved Pine brins in GUI features. At the top of
the screen is a band that has the program name, what I am currently doing,
the folder I am in and how many messages are there. At the bottom are some
control key codes to do various things such as send and/or cancel my
message. Additionally when new mail comes in a message comes up for a
short time letting me know how many peices and the subject and author of
the last peice of email recieved. These are all graphical elements in a
"text based" environment. 

Going to the GUI, I would imediatly point to Word. It is still a word
processor. The point is typing text. While it certainly extends this to
beyond imagination, it still remains a program for typing text. And then
lets look at email. While certainly there are pictures sent, the vast
majority of information is still textual in nature.

What it really comes down to is how the text is presented. A GUI adds
certain elements to the presentation. The shape of the letters will be
varied. The space between letters can be optimizes for best reading, or
just to "look cool." An astounding number of "graphics" on web pages are
in fact text that has been drawn to have a certain look that would not be
possible in using straight HTML.

If you could somehow "see" what all the fuss is, it would probably bother
you even more the amount of effort that has been going into GUIs. But
honestly, if someone had thought it all the way through from the begining,
access to a gui would never have been a problem. With the exception of
real drawings, such as a diagram or schematic, most information istextual
in nature. (Well much is now comming out in audio format.)



Kirk Wood
Cpt.Kirk@1tree.net
------------------

Why can't you be a non-conformist, like everybody else?

Re: A comment on Slashdot that concerns me

[Martin G. McCormick]

	That's quite true.  The only thing you don't have is serial
access during the initial configuration such as installation or during
the boot sequence.  Usually, it isn't all that necessary to have the
boot messages talking, but it is nice to have that capability if
something goes wrong and it is time to start being a diagnostition.

	I also got on to slashdot.org and read a lot of the linux VS
Windows thread.  I am not sure if some of those folks are actually
serious with some of their assumptions or are just jerking our chains
to see what we do.

	Anyway, the one thing I have not heard anybody say is that a
text-based interface is usable by everybody by one means or another.
A so-called graphical interface must be modified for any other form of
access.  When blind people use Windows, it is only because there has
been some progress in making it sort of behave like a command line
interface.  There just is not a good way to directly translate purely
visual information in to anything else that works as well.

	There is an interesting experiment that Bell Labs did in 1951.
I remember the year because it happens to be the year I was born so
things like that kind of stand out.

	What they did was to build a rather clever voice synthesizer
for that day out of a spinning wheel with bands of holes in it.  When
the wheel was spun at a certain rate, the holes raced past at
different frequencies.  A light shown through the holes and struck a photo
cell like the kind used in film projectors to convert the wavy band of
the film sound track to speech and music.  By blocking or unblocking
light from different bands of holes, the scientists could produce lots
of musical tones of different pitches.

	The next neat thing they did was to record a human voice on a
spectrograph which draws varying lines on a strip of film that
correspond to all the constituent frequencies in the sound being
recorded.  In this case, it was a man saying "Never kill a snake with
your bare hands."

	They took the film and used it to block and unblock the beams
of light through the bands of holes on their sound generator.  The
bands of holes corresponded to the center frequencies of all the
octave bands on the spectrograph machine.  The result was a voice that
sounds kind of like a DecTalk saying the recorded sentence very
clearly.

	The spectrogram looks like strange light and dark bands on the
film and means little to the eye except maybe that of an engineer, but
it did cause the generator to produce pretty good speech.

	The scientists then tried to make speech of their own by
manually painting spectrograms in a way that they thought would
produce new words and voices.  It never did anything but make weird
noises.

	My whole point is that the easiest way for a person who is
blind to do complex tasks on a computer is to use text.  It may be
that when tactile displays become dirt cheap and we can put our hands
on a screen and feel shapes, we may actually get closer to using a
GUI, but right now, we only use Windows when it can be bludgeoned in to
behaving like a command line.  Think about it.

	By the way, the only reason I remember the Bell Labs
experiments is because they appeared in a Bell Telephone Hour special
in the late fifties on television and I happened later to read about
them while doing a report in Graduate school.  My memory back to
earlier times is probably no better than anybody else's.  Just thought
I had better throw that in.

Martin McCormick WB5AGZ  Stillwater, OK 
OSU Center for Computing and Information Services Data Communications Group
Tommy Moore writes:
>Serial port accss is allready possible. I used it for months before
>speakup came out. You just have to figure out how to get the serial port
>to let you login off of it so that you can access it from another pc.
>
>
>
>_______________________________________________
>Speakup mailing list
>Speakup@braille.uwo.ca
>http://speech.braille.uwo.ca/mailman/listinfo/speakup
>

Re: GUIs (was Re: A comment on Slashdot)

[cpt.kirk]

Janina,

You put things quite well. One of the things that I found interesting is
that in Windows 3.1 there were quite a few companies that created less
graphical oriented shells. But to be honest, I think that MS has done a
terrible job of opening up Windows. I have a feeling that it was partly to
keep others from replacing portions of the code.

Having been through MS supplied training twice, I can ussure you that MS
changed the way several things worked such that they could no longer be
easily extended. The stated purpose was not to block extension of the OS,
but it was the result. But you are right that if access is had at the
lowest level possible things will be easier to deal with.

As for sending a picture of text, that is done too much for many reasons.
Not only does it cause problems for screen readers, it is also a bandwidth
waste. I also think that those writting books, and programs to creat HTML
should make alt text more prevelant then they do. It is so simple to add
that there is no excuse for it not to be employed. Then again, that is one
more symptom of the laziness that drives many to include a picture of text
in the first place. They are too lazy to work on their design.

Kirk Wood
Cpt.Kirk@1tree.net
------------------

Why can't you be a non-conformist, like everybody else?

Re: GUIs (was Re: A comment on Slashdot)

[cpt.kirk]

There are two uses for gifs as spacers. One is to replace the standard
rule with a graphical one. This takes very little bandwidth and can add to
the overall appearance of the page. I think it would be best to put in two
quotes here. (But that is assuming the person is bothering to put in alt
text.)

Recently there have been some who uise a gif to eat up space on the page
so that the text is made to look the way they want it. This is disgusting
abuse of bandwidth. I class it along with those reprobate companies that
make the first page a huge graphic that requires at least 800X600
resolution to display. They should have their web pages deleted and cattle
prods taken to their ... (Well I guess I might be getting out of hand
here, but probably not.)

Really, part of the problem on the web today is too much emphasis on looks
and not enough on content. The web is not print media, and many firms
can't grasp that fact. Perhaps we could come up with a protocol for
p[laces that make graphics that don't convey anything. It is one thing to
show me a picture of someone. Or to give me a drawing or map. But to have
a stupid logo blown to the size of my screen is bothersome.

Maybe search engines could be modified to not list these sites since they
don't have content.

Kirk Wood
Cpt.Kirk@1tree.net
------------------

Why can't you be a non-conformist, like everybody else?

Re: GUIs (was Re: A comment on Slashdot)

[Janina Sajka]

It would be accurate to say that all information displayed on a computer
monitor is graphical. Obviously, while this is true, it isn't helpful.

The real question is when and how the picsels are rendered. For example:

If text is retrieved from a web page as a file of picsels, we who are
blind are left out. There may as well be no text involved, from our point
of view, because we can't render it it in any way usable by us.

If, rather, the text is transmitted as a series of codes representing the
chars, and then some local agent converts these codes into picsels on the
monitor, we have the opportunity to fork the process and obtain access
through something like braille or synthetic speech.


To put this is in completely no technical terms:

The first instance is a picture of text. You can runa program like
Arkenstone's Open Book to convert it into the second form, and you might
succeed. But it is really a picture and the things you can do with this
kind of text are only the kinds of things you can do with pictures --
crop, rotate, zoom, etc.

The second is a set of codes -- just like the alphabet is a set of codes
irrespective of any particular way of rendering the codes. The kinds of
things one can do with this kind of text are the kinds of things one can
do with codes -- e.g. one can transform them into another set of codes
like braille, one can compare them to a dictionary of words for spell
checking, and one can change the kind of graphical rendering these codes
will get when they're printed on paper or on screen.

With this in mind, I want to share my view on the graphical user
interface. To my mind, the gui is neither bad nor good. Nor is it
intrinsically inaccessible. It's just another way of rendering an
interface. To my mind, the more rigorous and compartmentalized the
rendering, the more accessible a gui can be. If those aspects of the user
interaction which are inherently text in the coded sense are accessible at
that level, the gui should be very accessible. If those levels are
unavailable because they're proprietary--or because they're not regorously
adhered to, the interface becomes less accessible.

In other words, Microsoft Windows isn't intrinsically inaccessible. It's
just that Microsoft hasn't shouldered much of the responsibility for
providing the modesl and data sc reen reader developers need to build
access. Rather, the screen reader developers have had to do this
themselves. And, when they've put that amount of effort into the job,
they, in turn, have been highly proprietary about their solutions. Then, a
new generation of Windows is released that follows different rules at
important points, and the process begins again.


				Janina Sajka, Director
				Information Systems Research & Development
				American Foundation for the Blind (AFB)

janina@afb.net


On Tue, 11 Apr 2000 cpt.kirk@1tree.net wrote:

> There seems to be a lot of mis-understanding about Grphical User
> Interfaces (GUIs). The first being that they are purely (or even mostly)
> graphical in nature. The next misconception is that command line
> interfaces don't have any graphical content to them.
> 
> I will point to how our beloved Pine brins in GUI features. At the top of
> the screen is a band that has the program name, what I am currently doing,
> the folder I am in and how many messages are there. At the bottom are some
> control key codes to do various things such as send and/or cancel my
> message. Additionally when new mail comes in a message comes up for a
> short time letting me know how many peices and the subject and author of
> the last peice of email recieved. These are all graphical elements in a
> "text based" environment. 
> 
> Going to the GUI, I would imediatly point to Word. It is still a word
> processor. The point is typing text. While it certainly extends this to
> beyond imagination, it still remains a program for typing text. And then
> lets look at email. While certainly there are pictures sent, the vast
> majority of information is still textual in nature.
> 
> What it really comes down to is how the text is presented. A GUI adds
> certain elements to the presentation. The shape of the letters will be
> varied. The space between letters can be optimizes for best reading, or
> just to "look cool." An astounding number of "graphics" on web pages are
> in fact text that has been drawn to have a certain look that would not be
> possible in using straight HTML.
> 
> If you could somehow "see" what all the fuss is, it would probably bother
> you even more the amount of effort that has been going into GUIs. But
> honestly, if someone had thought it all the way through from the begining,
> access to a gui would never have been a problem. With the exception of
> real drawings, such as a diagram or schematic, most information istextual
> in nature. (Well much is now comming out in audio format.)
> 
> 
> 
> Kirk Wood
> Cpt.Kirk@1tree.net
> ------------------
> 
> Why can't you be a non-conformist, like everybody else?
> 
> 
> 
> _______________________________________________
> Speakup mailing list
> Speakup@braille.uwo.ca
> http://speech.braille.uwo.ca/mailman/listinfo/speakup
> 

Re: GUIs (was Re: A comment on Slashdot)

[Mike Gorse]

On Tue, 11 Apr 2000 cpt.kirk@1tree.net wrote:

> As for sending a picture of text, that is done too much for many reasons.
> Not only does it cause problems for screen readers, it is also a bandwidth
> waste. I also think that those writting books, and programs to creat HTML
> should make alt text more prevelant then they do. It is so simple to add
> that there is no excuse for it not to be employed. Then again, that is one
> more symptom of the laziness that drives many to include a picture of text
> in the first place. They are too lazy to work on their design.
> 
I agree for the most part, although I don't like seeing text descriptions
for spacer gifs (alt="" would probably be better for them).  But then are
spacer gifs really necessary?  I don't really know one way or the other.

Re: GUIs (was Re: A comment on Slashdot)

[Janina Sajka]

Hi, Kirk:

There's one particularly onorous way in which Microsoft has treated the
Windows assistive technology industry. I, for one, am galled that our AT
people have been willing to accept the monstrous task of divining the
inner workings of Windows, just so that they can develop an off screen
model to build speech and braille interfaces on. Strikes me that the OSM
should be M$'s responsibility.

Would video card manufac turer's put up with divining how to get video
resolutions and color depth out of Windows? Would printer manufacturers
settle for divining how it is that M$ apps can be made to print?

Add to that your observation of how frequently M$ changes its mind and
this situation becomes even more unbelievable, imho.


				Janina Sajka, Director
				Information Systems Research & Development
				American Foundation for the Blind (AFB)

janina@afb.net


On Tue, 11 Apr 2000 cpt.kirk@1tree.net wrote:

> Janina,
> 
> You put things quite well. One of the things that I found interesting is
> that in Windows 3.1 there were quite a few companies that created less
> graphical oriented shells. But to be honest, I think that MS has done a
> terrible job of opening up Windows. I have a feeling that it was partly to
> keep others from replacing portions of the code.
> 
> Having been through MS supplied training twice, I can ussure you that MS
> changed the way several things worked such that they could no longer be
> easily extended. The stated purpose was not to block extension of the OS,
> but it was the result. But you are right that if access is had at the
> lowest level possible things will be easier to deal with.
> 
> As for sending a picture of text, that is done too much for many reasons.
> Not only does it cause problems for screen readers, it is also a bandwidth
> waste. I also think that those writting books, and programs to creat HTML
> should make alt text more prevelant then they do. It is so simple to add
> that there is no excuse for it not to be employed. Then again, that is one
> more symptom of the laziness that drives many to include a picture of text
> in the first place. They are too lazy to work on their design.
> 
> Kirk Wood
> Cpt.Kirk@1tree.net
> ------------------
> 
> Why can't you be a non-conformist, like everybody else?
> 
> 
> 
> _______________________________________________
> Speakup mailing list
> Speakup@braille.uwo.ca
> http://speech.braille.uwo.ca/mailman/listinfo/speakup
> 

Re: A comment on Slashdot that concerns me

[Tommy Moore]

Slackeare can be easily installed over serial port.
Don't know about debian, but I think it can.

Re: GUIs (was Re: A comment on Slashdot)

[cpt.kirk]

Janina,

I agree that M$ should provide the offscreen model. This is especially
true since you are not supposed to decompile the software and all. Now I
won't go into software lisince isssues and what is really legal. But if
one were to abide by the letter of the lisence, then I doubt that a screen
reader could be written.

I am also quite familiar with the broken Active Accessibility. I believe
that M$ wants to appear responsible, but they are not willing to put up
the ante. I know for instance that MSAA was broken and they knew it ahead
of time. There was no surprise when IE4 released. It was known and the
decision made to go ahead.

But I think that when a company decides to close their code off from view
they should shoulder 100% of the load in making the code workable to all
who have a need to interface to it. The truth is, that they can and should
do better. They could make keyboard access a requirement. (For that
matter, they could probably make it so that keyboard access just happens
for all things.) They could also make it so that every peice of textual
data is always available. But they have chosen not too.

Part of the problem is that it has been years since there was a true
coordinated development effort in Windows. The project has grown to the
point where it is no longer coordinated. And in case your wondering, no
there is no documentation on the registry that comes anywhere close to
complete. There are branches that developers impliment because they need
to store settings and the programmer is the only one to know what is being
used for what. In some of those cases the programmer has left the company.

I used to be a staunch supporter if M$. I will not spend any of my money
on a product again if given a choice. In fact, I would ask for a refund
for any bundled apps that came on a computer.

Kirk Wood
Cpt.Kirk@1tree.net
------------------

Why can't you be a non-conformist, like everybody else?

Re: GUIs (was Re: A comment on Slashdot)

[Scott Howell]

man I knew there wa sa reason I liked this guy.



----- Original Message -----
From: <cpt.kirk@1tree.net>
To: <speakup@braille.uwo.ca>
Sent: Tuesday, April 11, 2000 9:32 PM
Subject: Re: GUIs (was Re: A comment on Slashdot)


> Janina,
>
> I agree that M$ should provide the offscreen model. This is especially
> true since you are not supposed to decompile the software and all. Now I
> won't go into software lisince isssues and what is really legal. But if
> one were to abide by the letter of the lisence, then I doubt that a screen

Re: GUIs (was Re: A comment on Slashdot)

[cpt.kirk]

Here is my take on the issue (for what it is worth). Active Accessibility
should be a real part of the OS. It should not be a patched version and
there should not be a special update (except perhaps that if it is updated
it could be offered until such time as the next core update comes out).
Basically, if one has the latest updates to the system, they should have
the latest updates for this.

But to do this would take far more then the small staff they have
"dedicated" to the task. Instead it would take sitting down and figuring
out how things work and planning accordingly. I am not a believer. I don't
believe that there will be a single core for "home" and "business." The
end of the seperate lines for 95 and NT is always pushed back. That is
because neither can be modified easily for the intended use. A full
install of linux OS and X windows is less then half the number of lines of
code that Win2000 has.

I don't think that the M$ solution will ever be anything else then a
catch-up for those looking to make it accessible. The only way that would
change is for the blind people to create the powerful force that the deaf
people have. Even then, I just don't know. Weasle Bill would probably talk
about how they couldn't innovate (attempt to turn other people's ideas
into their private proprietary standard).

Kirk Wood
Cpt.Kirk@1tree.net
------------------

Why can't you be a non-conformist, like everybody else?

Re: GUIs (was Re: A comment on Slashdot)

[Janina Sajka]

Hi, Kirk:

Yeh, M$ has, what? 24 or so staff on accessibility?

As compared to how many on Word alone? Something like 3,000, I believe.On
Tue, 11 Apr 2000
cpt.kirk@1tree.net wrote:

> Janina,
> 
> I agree that M$ should provide the offscreen model. This is especially
> true since you are not supposed to decompile the software and all. Now I
> won't go into software lisince isssues and what is really legal. But if
> one were to abide by the letter of the lisence, then I doubt that a screen
> reader could be written.
> 
> I am also quite familiar with the broken Active Accessibility. I believe
> that M$ wants to appear responsible, but they are not willing to put up
> the ante. I know for instance that MSAA was broken and they knew it ahead
> of time. There was no surprise when IE4 released. It was known and the
> decision made to go ahead.
> 
> But I think that when a company decides to close their code off from view
> they should shoulder 100% of the load in making the code workable to all
> who have a need to interface to it. The truth is, that they can and should
> do better. They could make keyboard access a requirement. (For that
> matter, they could probably make it so that keyboard access just happens
> for all things.) They could also make it so that every peice of textual
> data is always available. But they have chosen not too.
> 
> Part of the problem is that it has been years since there was a true
> coordinated development effort in Windows. The project has grown to the
> point where it is no longer coordinated. And in case your wondering, no
> there is no documentation on the registry that comes anywhere close to
> complete. There are branches that developers impliment because they need
> to store settings and the programmer is the only one to know what is being
> used for what. In some of those cases the programmer has left the company.
> 
> I used to be a staunch supporter if M$. I will not spend any of my money
> on a product again if given a choice. In fact, I would ask for a refund
> for any bundled apps that came on a computer.
> 
> Kirk Wood
> Cpt.Kirk@1tree.net
> ------------------
> 
> Why can't you be a non-conformist, like everybody else?
> 
> 
> 
> 
> _______________________________________________
> Speakup mailing list
> Speakup@braille.uwo.ca
> http://speech.braille.uwo.ca/mailman/listinfo/speakup
> 

-- 

				Janina Sajka, Director
				Information Systems Research & Development
				American Foundation for the Blind (AFB)

janina@afb.net

Re: GUIs (was Re: A comment on Slashdot)

[cpt.kirk]

You are right that at least M$ has done something. But I would also remind
you that Netscape isn't the only group not supporting MSAA. Heck, even the
screen readers weren't jumping on that wagon. Hynter-Joyce bypassed it
when they supported IE4 initially. Don't know the company name, but
Window Eyes also was slow to support it. And Daulphin found that not
paying much attention to M$ whims was benificial as well.

While I think Netscape could have done more, it was also fighting to even
stay alive (and failed). Truthfully, Winblows is so convoluted, that I
think screen reading is always going to be somewhat convoluted. Keep in
mind that though MSAA is supposedly part of the OS, support is not
automatic. In my limited view, the OS should be able to deal with reading
any text it creates. Thus, software manufacturers should not have to worry
about being accessible (other then to include such things as keyboard
shortcuts).

DOS applications never took time to be accessible. They sent text off to
the OS and it was done. Same with the console apps in Linux. I would bet
that few (if any) of the apps you use in Linux ever had blind accesibility
as a part of their design criteria. It should be built in by running on
the accessible OS.

Kirk Wood
Cpt.Kirk@1tree.net
------------------

Why can't you be a non-conformist, like everybody else?

Re: GUIs (was Re: A comment on Slashdot)

[Victor Tsaran]

Actually, WIndows2000 has MSAA as a part of the operating system. It is not
a question of Microsoft not including their own technology, but who supports
it. However much we blame Microsoft, but at least they are trying to do
something. Many apraised companies like Netscape, didn't even consider
supporting MSAA.
regards,
Vic

----- Original Message -----
From: <cpt.kirk@1tree.net>
To: <speakup@braille.uwo.ca>
Sent: Wednesday, April 12, 2000 7:01 PM
Subject: Re: GUIs (was Re: A comment on Slashdot)


> Here is my take on the issue (for what it is worth). Active Accessibility
> should be a real part of the OS. It should not be a patched version and
> there should not be a special update (except perhaps that if it is updated
> it could be offered until such time as the next core update comes out).
> Basically, if one has the latest updates to the system, they should have
> the latest updates for this.
>
> But to do this would take far more then the small staff they have
> "dedicated" to the task. Instead it would take sitting down and figuring
> out how things work and planning accordingly. I am not a believer. I don't
> believe that there will be a single core for "home" and "business." The
> end of the seperate lines for 95 and NT is always pushed back. That is
> because neither can be modified easily for the intended use. A full
> install of linux OS and X windows is less then half the number of lines of
> code that Win2000 has.
>
> I don't think that the M$ solution will ever be anything else then a
> catch-up for those looking to make it accessible. The only way that would
> change is for the blind people to create the powerful force that the deaf
> people have. Even then, I just don't know. Weasle Bill would probably talk
> about how they couldn't innovate (attempt to turn other people's ideas
> into their private proprietary standard).
>
> Kirk Wood
> Cpt.Kirk@1tree.net
> ------------------
>
> Why can't you be a non-conformist, like everybody else?
>
>
> _______________________________________________
> Speakup mailing list
> Speakup@braille.uwo.ca
> http://speech.braille.uwo.ca/mailman/listinfo/speakup

Re: GUIs (was Re: A comment on Slashdot)

[cpt.kirk]

While it could be argued that not all apps are accesible, I don't see any
arguing that this is the fault of the application developers.

The thing is that in Winblows it takes application support to make MSAA
work. M$ Office is a great example of this effect. Because they created
new controls that were not in MSAA definitions, MSAA failed to work with
them. The same thing has happened to other programs. In short, the only
way to program with known accessibility is to limit oneself to the M$
Foundation Library. But this negates the whole point of C++. A developer
is supposed to be able to create a new control. Should it then rest on him
to modify the OS so that speech works?

My feeling is that MS has done a poor job of making their products
accessable. This is a result of several factors. The first being that it
is a late life after thought. Next, there is minimal support given to
making/keeping it accessible. Someone said 12 developers are on the team.
It sounds about right. The lack of commitment is further illustrated in
IE4. It was known that MSAA would break. The product shipped anyway. It
was promised to be right for Win98. That was an empty promise. I saw that
one get cut. I knew one of the guys who was testing that (though not with
a screen reader).

Part of the problem is that Winblows is a cobbled OS. Part of that is due
to market forces. Part is due to the desire to gain ever larger market
share at any cost. I think eventually we will see the downfall of M$
because they can no longer build on what they have. Combine that with a
refusal to start with new code. (Or the inability to do so.) The thing is
that MS suffers from lack of documentation within their own code. There
are segments that nobody there knows what the function is. But when said
segment is pulled strange behaviors appear. Sometimes the fix for the
problem is to turn it into a "feature" that can't be disabled. Such is the
fate of the stupid tooltip that explains what the X in the upper right
hand corner faced. The switch to turn it off broke and they didn't know
why. In the end they just removed that from the config menus.


Kirk Wood
Cpt.Kirk@1tree.net
------------------

Why can't you be a non-conformist, like everybody else?

Re: GUIs (was Re: A comment on Slashdot)

[Victor Tsaran]

Hi, Krik!
This is actually arguable, I mean, the accessibility of all Linux apps.
Let's consider, for instance, Midnight Commander. Very simple app, right?
But if the screen reader doesn't support scrolling correctly, then using MC
would practically be impossible.
Regards,
Vic

Re: GUIs (was Re: A comment on Slashdot)

[Matthew Campbell]

Kirk Wood writes:
 > In short, the only
 > way to program with known accessibility is to limit oneself to the M$
 > Foundation Library. But this negates the whole point of C++. A developer
 > is supposed to be able to create a new control. Should it then rest on him
 > to modify the OS so that speech works?

I thought the whole point of MSAA was that applications could define
their own controls and use the MSAA API to make them accessible by
providing information about them to accessibility aids like screen
readers.  MS has implemented MSAA in Internet Explorer, and to some
degree in Office as well.  Before MSAA was introduced, screen readers
could provide access to applicatiosn that used the standard Windows
controls, but not necessarily to applications that had their own
custom controls.  I thought MSAA was designed to solve that problem.

But anyway, this is a Linux mailing list, so I'd like to present a few
ideas about how access to GUI's under Linux could be achieved.  I
think the main problem in achieving access to Linux GUI's is that
there are many GUI toolkits, each having its own set of widgets
controls.  I'm aware that there's a project called GSpeech which is
trying to add speech to the GTK toolkit, which is used by the GNOME
desktop.  It's a GTK module that's automatically loaded into any
application which uses GTK.  It has its own commands for reading
things on the screen, and it currently supports the Festival software
synthesizer.  I'm glad the author of GSpeech is putting the time and
effort into this area, but I don't think he's using the right
approach.  GSpeech only provides access to GTK and GNOME, and though
GHONE may be your primary desktop, you're bound to run an application
sometime that doesn't use GTK as its toolkit.  I suppose that every
toolkit and window manager could be speech-enabled in a similar way,
but then you'd have to use different screen review commands for each
toolkit, and keep track of which one is used by each application you
run.

I think a better approach is to have a screen reader which runs
outside of all your applications, and then modify or extend the GUI
toolkits so that they expose information about the GUI's to the screen
reader in a consistent way.  And this screen reader should be able to
at least read the text in any application window, regardless of what
GUI toolkit is being used, so that you have some level of access to
all X applications.

-- 
Matt Campbell <mattcamp@crosswinds.net>
Web site:  http://www.crosswinds.net/~mattcamp/
ICQ #:  33005941

Re: GUIs (was Re: A comment on Slashdot)

[Janina Sajka]

Isn't MSAA de facto evidence of the failure of MS operating and
application environments for access. Put the other way, if Windows and
Windows applications could directly support alternate interface
technologies, would there still be a need for MSAA?

My answer, just in case someone cares, is that MSAA proves that M$ is the
wrong environment for people with disabilities.

On Fri, 14 Apr 2000 cpt.kirk@1tree.net wrote:

> You are right that at least M$ has done something. But I would also remind
> you that Netscape isn't the only group not supporting MSAA. Heck, even the
> screen readers weren't jumping on that wagon. Hynter-Joyce bypassed it
> when they supported IE4 initially. Don't know the company name, but
> Window Eyes also was slow to support it. And Daulphin found that not
> paying much attention to M$ whims was benificial as well.
> 
> While I think Netscape could have done more, it was also fighting to even
> stay alive (and failed). Truthfully, Winblows is so convoluted, that I
> think screen reading is always going to be somewhat convoluted. Keep in
> mind that though MSAA is supposedly part of the OS, support is not
> automatic. In my limited view, the OS should be able to deal with reading
> any text it creates. Thus, software manufacturers should not have to worry
> about being accessible (other then to include such things as keyboard
> shortcuts).
> 
> DOS applications never took time to be accessible. They sent text off to
> the OS and it was done. Same with the console apps in Linux. I would bet
> that few (if any) of the apps you use in Linux ever had blind accesibility
> as a part of their design criteria. It should be built in by running on
> the accessible OS.
> 
> Kirk Wood
> Cpt.Kirk@1tree.net
> ------------------
> 
> Why can't you be a non-conformist, like everybody else?
> 
> 
> _______________________________________________
> Speakup mailing list
> Speakup@braille.uwo.ca
> http://speech.braille.uwo.ca/mailman/listinfo/speakup
> 

-- 

				Janina Sajka, Director
				Information Systems Research & Development
				American Foundation for the Blind (AFB)

janina@afb.net

Re: GUIs (was Re: A comment on Slashdot)

[Christopher Moore]

Martin,
Nice job on exposing the accessability issues.  Midnight Commander (MC) is
a file manager.  The released version appears to use highlighting rather
than following the cursor. A simpler file manager is pilot.  If you run:
pilot -gjv
You get a single column display with cursor following the filenames.  
Even as you illustrated, when probrams don't behave nicely out of the box,
Linux has plenty of weapons to handle the output.  In addition, in the GPL
world, you can get your hands on the source code and either make mods
yourself or get advice from the original developer.


73, Chris w1gm@sdf.lonestar.org
SDF Public Access UNIX System - http://sdf.lonestar.org

Re: GUIs (was Re: A comment on Slashdot)

[Martin G. McCormick]

	I am glad to read this thread as I see I am not totally crazy.

	Someone mentioned Midnight Commander.  I am not sure what that
application is, but the poster made an important point
that I will belabor a bit.  There are accessibility problems in UNIX
even without X windows, but they are a somewhat different grade of
beast than the Microsoft variety of problems.  There are applications
in UNIX that try to create a certain screen effect that causes our
linear method of reading to produce results that are hard to follow,
but the big difference is that we may be able to easily fix it if our
screen reader can be controlled in the way it reads to us.  That is an
important difference.  It is possible that we may not know how to make
the screen reader behave or that the screen reader is not capable of
doing what we need it to do, but it could.  I know that all sounds
contradictory, but what I am saying is that the data are present.  We
may not like the order they are presented in, but they are there.  A
little example is in order.

	There is a command in UNIX called df which system
administrators used to see how full a file system has gotten.  If I
just type

df -k /home

from the shel prompt, I hear

Filesystem         1024-blocks  Used Available Capacity Mounted on
/dev/dsk/c0t0d0s0    1893030 1167390   668850     64%   /

	This, by the way is from a Sun Sparc, so your mil age may vary,
but what all that was is two lines that form a little table.  The top
line is the name of the column headers and the bottom line consists of
all the values for those columns.  I must listen kind of carefully to
the stream of speech until I hear the very last element which, in this
case is 64%.  A better way to do that is to use nawk as a filter and
give a command like

df -k /home |nawk '{print $5}'

	That tosses out everything but the fifth column.  I hear

Capacity
64%

which is a little easier on the ears because you don't have to try to
sort out all the digits in the different columns.

	My point is that the data were there and I could doctor them a
bit to make them quicker and easier to read.  In the Microsoft-style
accessibility problem, the data are simply not anywhere to be gotten
using any kind of standard rules or functions.  Several years ago, I
seriously thought about diving in to Windows and trying to come up
with an open-source screen reader as my contribution to society.  To
be honest, I gave up right after I found out that each application is
its own little world and may or may not use standard function calls.
It is the ultimate nightmare scenario of the old DOS practice of
writing directly to the video buffer instead of using BIOS and DOS
function calls.  Microsoft tried to fix this with MSAA, but it was
like asking everybody to always say, "please" and "thank you."  It has
been an utter failure as far as I can tell.  The commercial screen
readers that are so expensive are so partly because one almost has to
practice Voodoo to tease out the data from various programs that each
do something different.

	I say flat out that MSWindows, NT, and 2K are all loosing
propositions as far as access goes as long as there is no sacred
method for handling I/O.  If this is ever fixed, it won't fix the
existing software base, but new software may work with some
applications having the second-level problem I previously described.
At least that would be a step in the right direction.

Martin McCormick


Janina Sajka writes:
>Isn't MSAA de facto evidence of the failure of MS operating and
>application environments for access. Put the other way, if Windows and
>Windows applications could directly support alternate interface
>technologies, would there still be a need for MSAA?