[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Can a machine test itself? + other issues


You raise some interesting theoretical issues, but my sense of it is that you have passed the bounds of the practical. The economic basis of computers was and is that they may run more than one program. It does not have to be this way; you could design and build a dedicated, one-purpose machine with hardware and software all custom crafted. Quite expensive, really, since all development costs must be amortized over one specialized application. The profit potential is limited, given the costs of development, so no private party would want to do this. Are you sure that the political realities are such that the U.S. Government in a time of deficit would choose to spend money on elections rather than tax cuts? I might agree with you on the desired approach to election equipment, but I am not very influential regarding the federal government's spending plans.

The fact of the matter is that in the real world we have MS Windows and Linux from which to choose, generic PC firmware and hardware, specialized and general purpose printers, proprietary hardware, and either open source or proprietary specialized voting software. From this mix, all US voting systems will certainly emerge. We do have the option of specifiying Open Test to supplement the existing proprietary, private, and inscrutible test organizations currently providing certification. Our standards must provide a measure of how well an individual voting system is likely to function, given our understood criteria.

Otherwise, although interesting, speculation about total custom voting systems (Open Source and Open Hardware Designs?) does not seem helpful except in order to highlight what is needed in the real world.

This is my personal and professional opinion, and it does not necessarily reflect the opinions of my employer.

-- Dick

Richard C. Johnson, Ph.D.
Applications Architect
Oracle Corporation

Stanley A. Klein wrote:

"BIOS" is just the acronym for the name IBM chose around 1980 for their
firmware interface when they built the PC.

The issue you seem to raise relates to the presence of an "operating
system" that is a general program providing services and control to
application programs.  In your embedded systems, the operating system is
likely to be a stripped-down program more tightly coupled, as you say,
to the applications.

The issue is one of generality versus tailoring.  Generality puts stuff
in the firmware interface and the operating system that isn't needed by
every application.  A voting system (or a pacemaker, for that matter)
isn't also going to be used to word process business letters.  And all
that extra stuff that isn't needed for what the system actually does is
a source of concerns for integrity, safety, and reliability.

The bottom line is that it is cheaper to throw in the kitchen sink, even
if you don't need it, but if you do your integrity, safety, and
reliability may go down the drain.

Stan Klein

On Tue, 2004-12-07 at 22:38, Vincent J. Lipsio wrote:

Every computer system needs some kind of firmware to interface the
operating system to the physical hardware. The BIOS just happens to be
the PC name for that function.

Point well taken.  However, my perception is that the use of the term "BIOS"
suggests an assumption that the system is, in fact, a PC, and not hardware
designed for the purpose, with consequent presence of all sorts of stuff not
related to voting, both hardware and software, and, therefore, many more
chances for hidden problems.  I am accustomed to designing systems with
absolutely no unused components, hardware or software, to reduce the cost
of validation and minimize the risk of some wayward component adversely
affecting the system.

Also, the bootstrap and ROM monitor (which, i believe, is what a PC's BIOS
is) are now often linked with the application code on embedded systems.

And, a BIOS is generally perceived as a black box that can be substituted
with any other BIOS designed for the same generation of hardware; hence, this
thread's original author's comment, "I have been inquiring about a foolproof
way to verify a BIOS.  I'm not sure that one exists."

Vince Lipsio

Every program has at least one bug and can be shortened by at least one
instruction -- from which, by induction, one can deduce that every
program can be reduced to one instruction which doesn't work.