🚀 go-pugleaf

In article <4j4n5j$o7c@twizzler.callamer.com> mgallo@slonet.org (M. Gallo)  
writes:
> In article <4iugmv$i9d@news1.delphi.com>,
> John Strohm <john.r.strohm@BIX.com> wrote:
> >Bengt Kleberg wrote:
> >>
> >>Last time I looked Oberon did not have preemptive multitasking.
> >>Has that been added?
> >
> >Not that I know of.
> >
> >Observe that the Oberon System works just fine WITHOUT preemptive
> >multitasking.  Observe that Microsquishy Windoze works just fine WITHOUT
> >preemptive multitasking.  Observe that the early Macintosh operating system
> >worked just fine WITHOUT preemptive multitasking.  Observe that there are a
> >rather large number of embedded systems that work just fine WITHOUT
> >preemptive multitasking. 
> 
> That sort of depends on what your expectations are, doesn't it?
> 
> >So why do you require preemptive multitasking?
> 
> How about for doing long-running calculations in the background
> (without having to decompose them into smaller problems and save
> their state)?
> 
> How about for responding to high-priority system events that *must*
> be addressed immediately (without having to rewrite drivers to
> respond to particular interupts)?
> 
> That said, I think Wirth & Gutknecht have shown how effective
> cooperative multitasking can be and that it is sufficient for a
> great many applications.
> -- 

Multitasking has been useful to me on a few occasions on an Amiga,
NeXT and on NT . 
Its nice to be able to put builds and large calculations etc. into the
background. Running a test case while editing source code can be useful too.
In the frantic hours before deadlines your head will sometimes seriously 
multitask at the level that you are interacting with the computer. When
that happens then much smaller tasks  get kicked off in windows in
significant numbers until the machine starts to 
crawl. It is a different style of use that isn't common or always useful
but when you need it, it is a great thing. Making workstations do it in
a way that is really responsive and that the user interface 
accounts for in a nice way is a good trick. If you can make the kernal
reentrant like on the Amiga then you can get really nice behavior.
Lack of memory protection and memory scrambling were problems.
On the NeXT the DisplayPostScript server is not reentrant or something and
it can cause me real headaches when I'm in a hurry. NT's defaults make the
responsiveness course sometimes but in gnereal it seems to be ok. Since
NT does not have a significant display server it does not have 
the problems that I have with NeXTStep but it has the problem 
that it doesn't do what NeXTStep does.

Most times, all I need is  cooperative multitasking. 

richard