Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!crdgw1!ge-dab!tarpit!bilver!bill
From: bill@bilver.uucp (Bill Vermillion)
Newsgroups: comp.sys.tandy
Subject: Re: Advanced Systems
Message-ID: <1991Apr20.173759.13891@bilver.uucp>
Date: 20 Apr 91 17:37:59 GMT
References: <1991Mar29.162359.2428@pdn.paradyne.com> <1991Apr14.164352.28893@bilver.uucp> <3986@batman.moravian.EDU>
Organization: W. J. Vermillion - Winter Park, FL
Lines: 61

In article <3986@batman.moravian.EDU> halkoD@batman.moravian.EDU (David Halko) writes:
>In article <1991Apr14.164352.28893@bilver.uucp>, bill@bilver.uucp (Bill Vermillion) writes:
>> In article <3763@batman.moravian.EDU> halkoD@batman.moravian.EDU (David Halko) writes:
   
>> If you want to find out what the MCA bus is all about in a fairly easy to
>> understand way, pick up the book "The Microchannel Architecture Handbook"
>> published by Brady. About 300 pages and $30.00.
 
>If you read the literature on the IBM 6000 series (I had read initial
>literature which was released about a year ago... not sure about their
>newer workstations, if they have any out now), you would see that the
>MCA bus which are in those workstations do not use the same MCA bus, but
>rather, a derivative of the same bus- although they still call it the
>MCA, there are differences.
....
>Now, is this the PC based MCA bus or the 6000 based MCA bus??? Most likely
>this is the 6000 MCA bus, since I doubt IBM had changed their PC line of
>bus's after releasing the original MCA since they are trying to gain 
>widespread acceptance of their new "standard"...

Suggest you read the book on the MCA architecture.

It is a bus spec and is processor independant.

The full MCA spec has yet to be implemented in any machine (that I am aware
of).

The 6000's use a derivative of the bus specs, as do all the other machines.

Even the model 95's have a more extensive implementation of the bus than
the orginal 70's and 80's, and the Tandy 5000. (refering to you comment on
doubting that IBM has changed the busses).

The whole design of the bus is structured around this concept.  The system
will "match" the cards.    A card engineered for the original
implementation on the 80's (for example) will work in a higher class of
machine.   All the features of the upper machine will not be implemented
but the original will still work.  Same goes for speeds.  They will match
each other.  We don't/won't have the problem of bus speed matching that is
inherent in the ISA architecture.

Conversely, a board designed to the specs, and implementing features not in
the orignal machines will still run in the original but without any of it's
added features.  

Yes the bus implementations are different among the 80's the 95's and
6000's but they adhere to the specs.

If any of the boards  require drivers you may be hindered by not having
drivers for the available OS.   The '386 implementation support DOS, OS/2,
Xenix, Unix, and AIX, and possibly some more that I have overlooked.
If you have, for example, a multiport board from a secondary source you 
will have to have software drivers to handle the boards, but the hardware
should just drop right in.

Further discussions should probably take place under comp.arch.


-- 
Bill Vermillion - UUCP: uunet!tarpit!bilver!bill
                      : bill@bilver.UUCP