Re: Indexer LPT and granularity

Fitch:

Anything is possible. :-) However, I don't think this is a reasonable thing
to do.

I'm marginally familiar with VMware and, with my limited knowledge, it would
not work for this. It sets up a virtual machine under linux (or windows)
which makes available the host machine's resources. So an output to a
parallel port by windows will pass through the linux I/O routine, causing
double the timing problems. It may well be possible to "give" a hardware port
to the windows application though which would solve this first problem. Then
it is like having two machines sitting in one box. Windows is in a
subservient window running under linux (this may be one reason I like it :-)
Communication between the two "machines" is though a network using smb and
samba. The network may be internal to the box but appears to the two OSs as a
real network. As you say one kludgy mess. It looks the same as having two
computers on a local network, linking file systems together using smb or
ntfs, and running one as an X-client on the other.

I'm not familiar with any of the other imulators. They may work entirely
different.

I think linux and windows aren't good examples here. Both have very good GUI
interfaces but neither is designed for real time applications. (EMC works
only due to a real time kernel patch). Is there a good, relatively
inexpensive, real time operating system available?

Hugh Currin
Klamath Falls, OR

On Saturday 26 January 2002 08:23 am, you wrote:
>
> Is the following possible:
>
> The controller package, like the one from NIST is running on the
> computer in Linux with all the real time benefits that bestows on it.
> On the same machine (say a 1.5GHz machine (or at least one powerful
> enough that computer capacity isn't an issue) is simultaneously running
> in windows the operator interface with all the nice graphics that makes
> available - the interface is of course running in a virtual machine
> created by a SW environment that allows running some form of windows on
> top of Linux.
>
> Can the operator interface running in the windows environment
> communicate with the NIST package running native Linux in a way that
> would make it possible to get the best of both worlds?
>