[comp.sys.ibm.pc] EISA: Warren Report

pat@sys1.TANDY.COM (09/18/88)

CLONEPAQ BUS INFO
From Carl Warren of Warren Group Labs
Posted for general interest/discussion

 
Date:     Fri Sep 16, 1988  8:45 pm  CDT
From:     Carl Warren / MCI ID: 310-9380
Subject:  more on EISA the clone clone bus
 
It has now been 3 days since Compaq announced EISA.  Every day a
new company announces they are conforming to the specification. 
The question arose could you get a spec.
 
The answer is yes and yes it is more than a functional
specificatin as I once thought.
 
The document appears to be a very well thought out technical
docuement that describes a future Compaq product.
 
It is 144 pages and although I only caught a glimpse appears to
be very complete in its treatment of the AT and extension bus.
 
But completeness doesn't necessarily mean that it is a good
approach, or technically sound.
 
My recommendation is multifold:
 
        1. obtain the document it will have some value
        
        2. Lobby Compaq to present their document to IEEE
           P996 for possible standardization
 
        3. IBM should do the same with Micro Channel and
           lift license restrictions--closed architectures
           tend to be non acceptable; ask Apple, they havee
           limited their growthh
 
This industry requires that companies explore multiple methods of
improving the product.  It would be a deraliction of duty on the
part of any company not to explore EISA.  
 
I believe that four strategies are emerging in the short term:
 
                o       An MCA clone fully licensed
 
Clearly IBM is being forced into a corner an will have to relax
license requirements for MCA.  I suggest they open the
architecture as soon as possible.
 
                o       An MCA functional equivalent
 
This is the EISA, or at least appears to be so.  I think this bus
should be restriction free--I don't believe that it is, based on
the agreements vendors have to sign.
 
The EISA must be viewed as a transition or bridge technology.  it
isn't a long range solution and will not support future
microprocessors.
 
I would like to recommend to Compaq that they put the bus up for
public scrutiny.  Why wasn't a prebriefing made to key
publications: EDN for example?  Why weren't copies of the
specification made to the press on Tuesday?  Why  does it cost
$2,500 and come from a Law firm?
 
                o       An AT/Nubus combination system.
 
This to me seems like an ideal solution; not only does it provide
a transition phase, but provides a firm foundation for vendors of
both cards and has the potential of expanding the market.
 
Ast Research has taken this approach, only they have no idea how
to market it.
 
 
                o       A Nubus only system
 
This is the best of the non-IBM bunch.  It offers a solid upgrade
path to match existing and future technologies.  Moreover it
isn't processor restricted.
 
Unfortunately, Nubus is associated with Apple Macintosh implying
proprietary architectures.  The Macintosh is closed only at the
software end not the hardware.
 
Further, companeis like AST, Compaq, DEC, SUN and TI are
preparing powerful Nubus designs that use CISC and RISC cpus such
as Motorola's 8800, and MIPs processor and Sun's SPARC.  Advanced
CISC processors like the 68040 and 80486 also take advantage of
the Nubus.