[IxDA Discuss] Fwd: Latency

[Claude Knaus]

---------- Forwarded message ----------
From: Claude Knaus <clyde7 at gmail.com>
Date: Oct 1, 2007 6:22 PM
Subject: Re: [IxDA Discuss] Latency
To: Jens Meiert <jens.meiert at erde3.com>


On 10/1/07, Jens Meiert <jens.meiert at erde3.com> wrote:
> > > > However, if the expected waiting time is short, like latency on a
> > > > mouse click, it is a problem if there is no prompt feedback. It is a
> > > > shame that today's operating systems did not adopt real-time
> > > > scheduling from operating systems like the Amiga OS. In my opinion,
> > > > user interaction imposes real-time constraints on feedback as a
> > > > requirement.
> > >
> > > True, but rather should /avoiding/ latency be a critical software
> > > design principle.
> >
> > Both are important. What you are saying is best effort practice. What
> > I was mentioning was that unbounded waiting (like when my machine
> > starts to swap in/out memory after a mouse click) is a terrible
> > experience. At least the window system should continue to remain
> > responsive.
>
> Of course, I didn't want to disagree but rather emphasize the importance to design software so that it induces as few latency as possible.

So, using a real-time system could be a solution for both designer and
engineer. The designer could specify a maximum tolerable latency. I am
only disappointed that real-time technology is not commonly used for
user interfaces.

-- Claude