Date   : Sun, 20 Jan 1985 13:20:37 CST
From   : mknox <mknox@ut-ngp.ARPA>
Subject: Reply to: cp/m standards

One thing that I have long argued with the system folk at DRI is for the 
DRI supported definition of a large number of additional OPTIONAL CP/M
calls.  I wanted them to define a few new BDOS calls, and a ton of BIOS
calls, all of which would be optional.  For an implementor to bring up
CP/M on a new machine would require no more than it does now.  But if
he wanted to add, for instance, support for a real-time clock, then there
would be a standard call for it, complete with a description of what to
return.  

I use this example because most all new designs support clocks, extra
ports, and the like.  But they all do it differently, even though all
are running CP/M.  

Key things would be:

  o Support calls for a large number of potential hardware devices, 
    including I/O ports, clocks, interrupts, and screens.

  o A 'configuration' call to return the system configuration (in some
    standard format).

  o A standard mechanism for returning a 'fault' or 'not-available'
    status.

So far some of the DRI people say they have been arguing for the same
thing.  Others are afraid a long list of calls (even optional) would
scare off developers.