(was slashes, now NFS devices)

[Frank Peters]

: On 6 Mar 91 14:52:21 GMT, peter at ficc.ferranti.com (Peter da Silva) said:

In article <GLX9W24 at xds13.ferranti.com> peter at ficc.ferranti.com (Peter da Silva) writes:

> In article <FWP1.91Feb23160240 at Jester.CC.MsState.Edu> fwp1 at CC.MsState.Edu (Frank Peters) writes:

>> 1. NFS is designed to be operating system independant while RFS
>>    assumes UNIX on both ends of the connection.

> OpenNET lets you do things like accessing a named pipe from DOS. If
> anything operating system independence would seem to imply that file
> semantics should be interpreted in the server.

Nope.  The date in a /dev file is just that...data.  If the client is
a UNIX box with the sense to interpret that data as a "device pointer"
(for want of a better term) then they get to use it.  Thats just what
a diskless workstation does.  If the client doesn't know how to
interpret that data then it doesn't interpret it...just like it
doesn't interpret SPARC data as a program.

I won't claim this is the way it must be done.  I will claim though
that it is a reasonable thing to do.


>> 2. This is probably most important.  NFS is stateless.

> This is the killer. I think it would be reasonable to relax statelessness
> for devices, though.

If you are going to relax statelessness for devices then you might as
well relax it for everything.  In which case you might as well scrap
NFS.  The stateless philosophy, right or wrong, is ingrained into NFS.

If you want remote device access I really think one of the other
systems would be a better start.

Frank
--
Frank Peters   Internet:  fwp1 at CC.MsState.Edu         Bitnet:  FWP1 at MsState
               Phone:     (601)325-2942               FAX:     (601)325-8921