To: <[EMAIL PROTECTED]>
>Date: Sun, 10 Oct 1999 19:36:21 -0400 (EDT)
>From: Ish Rattan <[EMAIL PROTECTED]>
>To: [EMAIL PROTECTED], [EMAIL PROTECTED]
>Subject: Re: [rtl] curiosity
>
>> From: Sy Wong <[EMAIL PROTECTED]>
>> To: [EMAIL PROTECTED]
>> Subject: [rtl] curiosity
>>
>> I am curious to know if Alstair's teachers are experienced designers of
>> (cut off-->) hard real-time systems.
>
>You are being too harsh on the teachers, they may have no bearing on
>the contents of the post..
>
>-ishwar
I must be unclear that I was NOT reflecting on Alstair and especially his
teachers. A TV program decrying Oakland CA shools using teachers
qualified in certain subject matters as interchangeable components to
teach classes they have no background. The program was not a reflection
against the teachers at all. Computer science and HOLs grew up in the
era of economic necessity to time-share the use of expensive mainframes.
These seems to be implied in the word "scheduling" that prompted my first
email. I was merely asking the teacher's backgrounds to check my
perception of the current CS community. The original computers were
designed by mathematicians and electronic engineers. The establishment
of CS departments in universities came later.
Perhaps Alstair's interest to understand rtlinux code is a good step to
bridge the hareware and CS cultures. In fact, I almost wanted to do like
he does in trying to understand the code to check the possibility of
using rtlinux alone without Linux. Inputs from a very knowledgeable
source pointd out that RTLinux depends on the full Linux. I got a wrong
impression of RTLinux based on a magazine article I read. This saved me
from looking at RTLinux source codes unless somebody can tell me that the
part RTLinux needs from Linux can be pried out and grafted to make
RTLinux independent, for a system without any multi-tasking implications.
I re-emphasize that I asked my questions because of my ignorance of
rtlinux and not a reflection on anybody. Now I am slightly wiser after
being enlightened. I am also grateful to Paul Koning:
>Date: Mon, 11 Oct 1999 13:49:31 -0400
>From: Paul Koning <[EMAIL PROTECTED]>
>To: [EMAIL PROTECTED]
>Cc: [EMAIL PROTECTED]
>Subject: Re: [rtl] curiosity
>....
>I don't know where you got that impression. The word "scheduler" is
>not at all limited in the way you suggest.
>ANY system that has more than one thread (task, process, pick your
>favorite name) has a scheduler. It may not be called by that name,
>and it may not look like much.
That is very true. I will specifically include primitive systems with
tasks initiated directly from clock interrupts sampling some time-varying
inputs and drops to some background tasks without any hint of operating
systems or HOL constructs.
>For example, in classic Forth
>(explicit release of control, non-preemptive, single priority)
>systems, the scheduler is pretty small and doesn't necessarily show up
>as a block of code you can easily point to. But, functionally, it is
>there.
>Similarly, real time systems have schedulers. Some are simple, others
>not so. Timesharing systems also have schedulers (typically rather
>primitive ones). Clearly, since the system requirements are different,
>real time systems have schedulers with different properties than
>timesharing systems.
Perhaps my use of time-sharing was unfortunate. To CS persons, TS may
imply time-slicing like that in the first Dartmouth Basic system put out
by GE where each user gets a fixed slice of time. When the load is
heavy, you wait longer like the centralized phone system of HMOs. My use
of time-sharing means sharing a single computer among several weakly
dependent tasks under some visible operating system such as NT, OS/2,
Unix or Linux. I quote the expert that enlightened me.
>It is correct to say that effectively Linux is a background service, but
>it still can't (easily) be thrown away, as real time Linux depends on it
>for some services:
>
>1/ Initialization of the system (hardware/processor/interrupts etc.)
>
>2/ Some system services such as module insertion, kmalloc (during
>initialization) for the real time task's memory.
The RTlinux apparently fits my definition of time-shared use of a
processor among more than one (major) tasks and not fit my image of an
embedded systems. My impression is that the program memory is
permanently residing inside the processor with statically pre-allocated
ram space addressed directly. The address allocation are done at code
generation time during development. Would you need RTLinux if each task
is assigned to a separate processor where scheduling is implied by system
wiring? DOS also fits the "ANY sytem" definition, ignoring my possibly
incorrect terminologies. DOS provides the services listed. Is it
practicable to modify rtlinux to run DOS in lieu of Linux as low level
task, without and making it a major project? Could DOS be considered an
RT task as described in some RTLinux descriptions?
SY Wong, 11.10.99
--- [rtl] ---
To unsubscribe:
echo "unsubscribe rtl" | mail [EMAIL PROTECTED] OR
echo "unsubscribe rtl <Your_email>" | mail [EMAIL PROTECTED]
----
For more information on Real-Time Linux see:
http://www.rtlinux.org/~rtlinux/
