Re: GUI INTERFACE

"James J. Weinkam" <jjw@xxxxxxxxx> writes:
> Peter Flass wrote:
> > I'm not an authority, but my guess would be this:
> >
> > With a pseudo-conversational program (or conversational, for that
> > matter) the user interaction is controlled by the program. This gets
> > somewhat more complicated with full-screen stuff, but basically the
> > program sends something to the user, and the user responds. The user
> > interactions are limited. An HTML form is an updated version of
> > pseudo-conversational. The difference between C and P-C is not visible
> > to the user, but is a function of whether the process sticks around
> > waiting for the user to reply, or goes away and is re-started by
> > "something" when he replies.
> >
> > With event-driven programs in general, the interaction is not controlled
> > by the program but by external events, which may or may not be random or
> > occur at random times. The program recieves notification that an event
> > occurs, does whatever is necessary to handle it, and goes back to
> > waiting. The original event-driven programs were process-control, where
> > the events were generated by external equipment, and the action
> > generated some feedback to the equipment.
> >
> > In *some* ways a GUI program could probably be considered event-driven,
> > since the user can interact in many non-sequential ways, and the program
> > also receives notificatio from the system of other events shch as system
> > shutdown. While I'm typing input to this message, for example, I can
> > click on a menu that might call up an independent dialog such as
> > "print", "help", etc that may run in parallel with the typing task.
>
> Yeah. If you study Computing Science, you will learn that there are a number of
> different models of interprocess communication. All are logically equivalent in
> the sense that each can simulate all the others. A lot (most, in fact nearly
> all) system designers make the mistake of assuming that since the various
> methods are logically equivalent they are equally efficacious. Nothing could be
> further from the truth. In my estimation, the message passing paradigm is the
> most fragile.

> Consider the desire to kill an errant process. In a system based
> on the message passing paradigm, the only way to do this is to send the process
> a message ordering it to commit suicide. The lunacy of this approach should be
> obvious to the meanest intelligence. The process is *HUNG*. Is isn't responding
> to any messages. Any and all messages go into the bit bucket. Worse, actually,
> they accumulate in the process's message queue, eating up valuable storage. The
> only way to clean up the mess is to reboot.

> What is needed is a *CONTROL*
> program that can look at the data structures that describe all the processes in
> the system and the resources that they hold and remove the errant process and
> return its resources to the available pool. That ain't gonna happen with
> message passing.

???

Just why can't a *CONTROL* program/process be produced
which can do the things described? If the OS tracks
the resources used by a process, why can't it kill
the hung process and release/retrieve the resources?
.

Relevant Pages

  • The cost of just saying no
    ... control their behavior in the face of temptation. ... Once these resources are exhausted, ... another task that requires a significant amount of self-control. ... performed worse on the Stroop task, indicating that they had used up ...
    (uk.philosophy.humanism)
  • Re: [ckrm-tech] Re: [Lse-tech] [PATCH] cpusets - big numa cpu and memory placement
    ... provide a fairly rich degree of control over what share of resources ... on the same hardware resources, the tuning of the critical application ... not physical isolation in the sense of unplugging the ... The above is not the only use of cpusets - there's also providing ...
    (Linux-Kernel)
  • Re: Trouble with strongly-typed Image resources at design-time
    ... problem using Image resources from the Resource Manager. ... one for the control and one for the test ... and all works fine at runtime and design-time. ... disposing the Image objects incorrectly because it works fine when I ...
    (microsoft.public.dotnet.framework.windowsforms)
  • Re: ATI Catalyst Control Center with .NET v2
    ... the Microsoft Error Report pop-up opened "ATI EXTERNAL EVENT UTILITY.EXE ... > conclusion that it's just not worth the system resources it takes up. ... > I have spent more time trying to get the cat control ctr running right than ...
    (microsoft.public.dotnet.general)
Loading