Intro ...

Brian Scott Haskett haskett at cs.utexas.edu
Thu May 5 17:43:03 GMT 1994



On Thu, 5 May 1994, Jeffrey S Armfield wrote:

> Control algorithms are no problem. I need to know more about real time
> operating systems and some of the more esoteric hardware stuff.

OK- I have been thinking about this the past few weeks.  I think that the 
best way to implement a *cheap* real-time data acquisition/control 
application is to do so on a DOS machine.  I don't think anyone wants to 
hear the nitty-gritty explanation of real time, but a hard real-time 
system is just not economical or practical.  To achieve real-time, disk 
i/o and user i/o (or basically anything that takes a long time) cannot be 
used.  This rules out anything with virtual memory.  So....  a nice, 
stripped down version of DOS with little or no TSR's would be a good 
platform to implement a soft real-time system, in my opinion.  Just go to 
the local used computer store, pick up a cheap 286 laptop, and go work.  
The application would have to store all of its data in RAM, until it can 
exit real-time mode and store to / read from the disk.  

Of course, this is just the thinking that I have been banging around the 
past few weeks, and is certainly open for discussion.

-Brian



More information about the Diy_efi mailing list