Imagine: A GUI standard.

[Walter Roberson]

Not much call for a GUI on refrigerators,

The Korean company "LG" markets several refrigerators with
LCD displays and a GUI. In some of them, the LCD display acts as a TV.

On some refrigerators, the display allows changing settings
on the refrigerator (e.g., temperature controls.)

Some refrigerators have bar-code scanners and links to UPC databases
and keep track of when various food was added or removed from the
unit, and can be queried for the current inventory or for the
list of items that are getting old.

train track monitors and
nuclear installations.

The search for WMD in Iraq would have been so much simpler if the
installations had good GUIs ;-)

 

[Keith Thompson]

There are many...


Note that a GUI isn't a C feature, its an OS feature.

The same could be said for file I/O.

It happens, by historical accident, that the state of file I/O is such
that it can sensibly be standardized in a language (though some
systems have features that the language standard doesn't capture), but
GUIs probably cannot. It needn't have been that way; if graphics
systems had converged more than they did in real life, we might have
<stdgraph.h> in C.

 

[Dann Corbit]

The same could be said for file I/O.

It happens, by historical accident, that the state of file I/O is such
that it can sensibly be standardized in a language (though some
systems have features that the language standard doesn't capture), but
GUIs probably cannot. It needn't have been that way; if graphics
systems had converged more than they did in real life, we might have
<stdgraph.h> in C.

Java has a GUI standard.
HTML is a sort of GUI standard.
PHIGS is a graphics standard.
GKS is a graphics standard.
ODA is a graphics standard.
CGM is a graphics standard.

All things are possible if you want to do them bad enough.

Of course, a GUI standard probably does not make a lot of sense for toaster
ICs.

I guess that an add-on GUI standard for programming would have to be a step
above, sort of like the difference between a hosted and non-hosted C
implementation.

It is also possible to standardize the interface to operating systems. That
is what ACE does:
http://www.cs.wustl.edu/~schmidt/ACE.html

Standards are nothing but agreements. "Hey, everybody... Let's do it like
this!"

 

[Chris Hills]

Not much call for a GUI on refrigerators, train track monitors and
nuclear installations.

Wrong on all three counts.
At the risk of advertising see www.segger.com
that graphics library (not GUI) has been used on all three things you
mention above.

I am sure most of the other graphical library systems can say the same.

 

[Chris Hills]

Java has a GUI standard.
HTML is a sort of GUI standard.
PHIGS is a graphics standard.
GKS is a graphics standard.
ODA is a graphics standard.
CGM is a graphics standard.

All things are possible if you want to do them bad enough.

Absolutely and that is the point.

The industry does NOT want a standard GUI. There are a lot of graphics
libraries out there and GUIs al of whom would campaign against a
standard unless it fitted their model exactly.

It the work was started by ISO all the interested parties would slow it
to a standstill until it died.

Of course, a GUI standard probably does not make a lot of sense for toaster
ICs.

I am sure at least one toaster has an LCD and graphical display.

I guess that an add-on GUI standard for programming would have to be a step
above, sort of like the difference between a hosted and non-hosted C
implementation.

Are we talking about a GUI or a graphics library? A subtle different to
my mind.

One looks like Windows/Gem/OSX/ Solaris and the other is a library to
put graphics on screen like a control panel, coffee machine etc. It one
tends to have "windows" not (Windows TM) and the other tends not to.

Standards are nothing but agreements. "Hey, everybody... Let's do it like
this!"

There are far to many companies out there with their own commercial
system to not want a standard.