Re: MISC-d Digest V98 #28

["GARY B. LAWRENCE" <garyl@xxxxxxxxxxx>]

MISC-d-request@pisa.rockefeller.edu wrote:
> 
> Subject:
> 
> MISC-d Digest                           Volume 98 : Issue 28
> 
> Today's Topics:
>          MISC news
> 
>     ---------------------------------------------------------------
> 
> Subject: MISC news
> Date: Fri, 10 Jul 1998 17:10:01 -0300
> From: jfox@dnai.com (Jeff Fox)
> To: MISC
> 
> Dear MISC readers:
> 
> We tried to make the July 8 .8u HP run at Mosis with a new F21
> prototype this week.  We didn't quite make it and plan to submit
> a chip to Mosis on Aug 19.  This will bring these chips back in November.
 I'm glad to see a fabrication run being done. I for one still would 
 like to see an F21 chip. The use of a Forth chip with display and 
network in the internal hardware seem to me to be very important
features in areas such as local factory networks, home control networks,
and mutiprocessor systems. 
     I would like to get some feedback from this group on a subject that
interests me at the moment. I would like to see F21's be used in a 
multiprocessor version of a multitasking operating system. Although
initially this would seem harder to do than the current open source 
system I consider as the model - the Linux system on an X86 single
processor - it seems that by placing system tasks in separate F21's
would result in a more modular high level design. 
    One area of the operating system that might easier to break out than
others might be the file system. Since file commands have delay in
response from the hard drives, sending file commands over a system
network to a different processor would not slow down the overall
response much and would allow the use of a separate processor and memory
for operations that are coming more complex with each new version of
an operating system. Things like high performance file clustering,
automatic tape backup, file system redundancy and redundant disk arrays
show an increase in the complexity of this part of the operating system.
Using a separate Forth processor to handle the standard file
manipulation commands and control the drives would make a more modular
design. The drive control code could be done at a higher level than the
 x86 machine code drivers used now in systems such as Linux and the
rebuild of an experimental system should be faster. Also since each F21
has it's own display, a monitor program could be added to the tasks of
the different system F21's and a person could switch a screen to the
different F21's using a video switch circuit to see the status of
various processes on each F21 chip. Monitor programs on Linux need to
add processes to the system and would seem to me to be more complex to 
 write. 
  Use of F21 boards might make it easier to add new tools to an
operating system. For instance an F21 board that is not running a
current process could be used to build "browse" type links of command
words used in a currently running program. This may make it easier for a
user to flip to a display that will show use of commands from scource
code or comment code. Unix systems use a "man" page for this but the 
Forth "browse" structure may do this in a simpler fashion. 
  A Forth system that does not try to totally copy Unix programs and
commands should have a much simpler structure since Forth itself acts as
the shell and interpretative Unix languages like pearl could be done by
using Forth extensions in interpretative mode. I am hoping for some
feedback from this group on these ideas. 
			best regards Gary Lawrence