War with the Solarian League?

On 2007-10-11, pyotr filipivich <phamp.TakeThisOut@mindspring.com>
allegedly proclaimed to alt.books.david-weber:
> [...] There is an
> anthropomorphism in our view of computer systems [...]

And computers really hate it.

Gym "You supply the set-up..." Quirk

--
Capt. Gym Z. Quirk (Known to some as Taki Kogoma) quirk @ swcp.com
Just an article detector on the Information Supercollider.
 >> Stay informed about: War with the Solarian League? 

On Oct 10, 11:58 pm, pyotr filipivich <ph... DeleteThis @mindspring.com> wrote:
> After a Computer crash and the demise of civilization, it was learned
> Don Sample <dsam... DeleteThis @synapse.net> wrote on Mon, 17 Sep 2007 16:12:01
> -0400 in alt.books.david-weber :
>
>
>
> >The security to prevent such things happening will be more along the
> >lines of limiting access to such areas to authorized personnel only than
> >in building in more safeguards. Ever safeguard you build in, is one
> >more way for something to break, too. Even if all the safeguards fail
> >safe, the fail safe mode will be along the lines of the thing just not
> >working, which can itself be deadly in a combat situation.
>
> Harkness had Tremaine override the "failsafe". There is an
> anthropomorphism in our view of computer systems, they do not "decide"
> to do something, they follow the programming. E.g. the computers which
> monitored the ready armour in the morgue didn't "lobotomize" them, or
> agreed to lobotomized them, they ran a standard procedure for testing
> the connections. All according to the program - which sent 220 volts AC
> down a line intended (by the designers) for 3 volts DC (or what ever).
> The computers didn't know, or care, that the voltage was "wrong" only
> that it was "in spec".

This brings up another issue: Why is a line intended for only 3VDC
connectable to a line carrying 220VAC via a soft switch?
One thing technicians are quickly taught about panel design and
assembly is that the power lines and control lines are not run in the
same cable bundles and the terminal blocks are kept seperate (USB is
allowed to mix them as the voltages are identical). A hard lesson
learned in aircraft wiring harness design is to make it impossible to
make the wrong connection without actually removing the connectors and
fashioning rat tails-- anything less leads to misconnected cables
(aircraft have been lost, in the bad old days, when it was possible to
cross connect the roll damper or engine fire extinguisher squibs).
Engineers are a conservative lot, not overly prone to feature creep.
It would take some doing to make one incorporate a feature with
potential dire consequences that offers no useful functionality.

The mayhem caused by opening valves and pumping chemical rocket fuel
is also negated by the paranoid design-- hardware interlocks that
prevent the valve from opening without a hose connected to the
fitting. You could still get around it, but you would have to
physically defeat it from within the boat bay.

It is not impossible to design a bug-free computer control system, if
you are willing to accept the consequences (which are mostly of the
form of no dynamic memory allocation). The reactor safety system at
the Darlington Nuclear Generating Station is a case in point. It was
eighteen months behind schedule and much more expensive than the
electromechanical systems of earlier nuke plants, but they crafted a
specification that could be both mathematically proven to be bug free
and proven to be implementablebug free. Finally, for the purposes of
getting the operating permit, they had to verify that actual
implementation matched the specification.

Generally speaking, the work you put into ensuring that a safety
feature works is proportional to severity of what happens when it does
not. Star Trek TNG is really bad for this. The Emergency Core
Ejection System should not fail in any circumstance that has not
already destroyed the ship. Safety critical stuff will reject unsafe
commands and the controlled hardware will probably be unable to
execute them, anyways.

The pinnace wedge is another kettle of fish, altogether. There is no
infallible way to prevent a system operating exactly as intended, even
when such operation is inconvenient, that does not threaten to prevent
the system from operating when you need it.
 >> Stay informed about: War with the Solarian League? 

After a Computer crash and the demise of civilization, it was learned
"rlbell.nsuid@gmail.com" <rlbell.nsuid.TakeThisOut@gmail.com> wrote on Sat, 13 Oct
2007 04:44:46 -0000 in alt.books.david-weber :
>On Oct 10, 11:58 pm, pyotr filipivich <ph....TakeThisOut@mindspring.com> wrote:
>> >The security to prevent such things happening will be more along the
>> >lines of limiting access to such areas to authorized personnel only than
>> >in building in more safeguards. Ever safeguard you build in, is one
>> >more way for something to break, too. Even if all the safeguards fail
>> >safe, the fail safe mode will be along the lines of the thing just not
>> >working, which can itself be deadly in a combat situation.
>>
>> Harkness had Tremaine override the "failsafe". There is an
>> anthropomorphism in our view of computer systems, they do not "decide"
>> to do something, they follow the programming. E.g. the computers which
>> monitored the ready armour in the morgue didn't "lobotomize" them, or
>> agreed to lobotomized them, they ran a standard procedure for testing
>> the connections. All according to the program - which sent 220 volts AC
>> down a line intended (by the designers) for 3 volts DC (or what ever).
>> The computers didn't know, or care, that the voltage was "wrong" only
>> that it was "in spec".
>
>This brings up another issue: Why is a line intended for only 3VDC
>connectable to a line carrying 220VAC via a soft switch?

Good question. Ummm ... Haven went for a standardization of
connectors due to the egalitarian nature of the People's Republic?
Ah, seeing as how it is a well known fact, that electronic
components run on magic smoke, then it is obvious that some of the smoke
and mirrors involved in pulling this off was redirected into the morgue
computers. It was this sudden surplus of magic smoke, which was
redirected by local fans vertically to the seats of the combat suits,
all of which were manufactured in the eight month. This is known as
blowing smoke up their august ass.

>One thing technicians are quickly taught about panel design and
>assembly is that the power lines and control lines are not run in the
>same cable bundles and the terminal blocks are kept seperate (USB is
>allowed to mix them as the voltages are identical). A hard lesson
>learned in aircraft wiring harness design is to make it impossible to
>make the wrong connection without actually removing the connectors and
>fashioning rat tails-- anything less leads to misconnected cables
>(aircraft have been lost, in the bad old days, when it was possible to
>cross connect the roll damper or engine fire extinguisher squibs).
>Engineers are a conservative lot, not overly prone to feature creep.
>It would take some doing to make one incorporate a feature with
>potential dire consequences that offers no useful functionality.
>
>The mayhem caused by opening valves and pumping chemical rocket fuel
>is also negated by the paranoid design-- hardware interlocks that
>prevent the valve from opening without a hose connected to the
>fitting. You could still get around it, but you would have to
>physically defeat it from within the boat bay.
>
>It is not impossible to design a bug-free computer control system, if
>you are willing to accept the consequences (which are mostly of the
>form of no dynamic memory allocation). The reactor safety system at
>the Darlington Nuclear Generating Station is a case in point. It was
>eighteen months behind schedule and much more expensive than the
>electromechanical systems of earlier nuke plants, but they crafted a
>specification that could be both mathematically proven to be bug free
>and proven to be implementablebug free. Finally, for the purposes of
>getting the operating permit, they had to verify that actual
>implementation matched the specification.
>
>Generally speaking, the work you put into ensuring that a safety
>feature works is proportional to severity of what happens when it does
>not. Star Trek TNG is really bad for this. The Emergency Core
>Ejection System should not fail in any circumstance that has not
>already destroyed the ship. Safety critical stuff will reject unsafe
>commands and the controlled hardware will probably be unable to
>execute them, anyways.

Yeah, but it isn't as dramatic as "Warp Core Breach immanent."

OTOH, the equipment on Star Trek (and Seaview. And others) seems to
have been farmed out by the lowest bidder. At the least minor damage,
sparks fly out of the control panels.
>
>The pinnace wedge is another kettle of fish, altogether. There is no
>infallible way to prevent a system operating exactly as intended, even
>when such operation is inconvenient, that does not threaten to prevent
>the system from operating when you need it.

Yep.

You can't make systems idiot proof, idiots can be so ingenious.
--
pyotr filipivich
The two oldest cliches in the book are "The Good Old Days were
better." and "After all, these are Modern TImes."
 >> Stay informed about: War with the Solarian League? 

When you spoke of "Anthropomorphizing Computers", I thought that
perhaps you were referring to Merlin from "The Excalibur Alternative".

John Savard
 >> Stay informed about: War with the Solarian League? 

rlbell.nsuid.TakeThisOut@gmail.com wrote:
> This brings up another issue: Why is a line intended for only 3VDC
> connectable to a line carrying 220VAC via a soft switch?

That is a good question. But there is a reasonable answer, too.

Basically, one imagines a computer system which is expected to work as
specified, but which is capable of connecting pretty much anything to
anything else. Which is not a problem, unless the computer system is
tampered with.

Presumably, Haven viewed that as being just as impossible as, say,
hypnotizing the crew and getting *them* to connect the 220 VAC to the
3 volt circuits in the suits.

Putting "everything" under the computer's control reduces the number
of times when you have to get a crew member to physically connect a
wire from here to there. Also, note that educational standards on
Haven were lacking - and the members of the secret police would
naturally want things to be done for them.

If one had multiple *small* computer systems that controlled specific
individual systems, indeed one could validate the software for each
one, and not give any of them the power to do something silly. But if
you have one big computer system, and your focus is on making the user
interface into "Hey, computer, would you do this for me", well...
you're taking the chance that just the sort of thing will happen as
the novel illustrated. Haven was sloppy and lazy - the books have been
telling us that all along.

John Savard
 >> Stay informed about: War with the Solarian League? 

pyotr filipivich wrote:
> Yeah, but it isn't as dramatic as "Warp Core Breach immanent."

I didn't know that each of us had a Warp Core Breach beside us, every
day of our lives - that the Warp Core Breach surrounds us, and in it
we live, move, and have our being.

Even the eschaton will become imminent before it becomes immanent!

John Savard
 >> Stay informed about: War with the Solarian League? 

After a Computer crash and the demise of civilization, it was learned
Quadibloc <jsavard.TakeThisOut@ecn.ab.ca> wrote on Sun, 14 Oct 2007 16:35:12 -0700
in alt.books.david-weber :
>pyotr filipivich wrote:
>> Yeah, but it isn't as dramatic as "Warp Core Breach immanent."
>
>I didn't know that each of us had a Warp Core Breach beside us, every
>day of our lives - that the Warp Core Breach surrounds us, and in it
>we live, move, and have our being.
>
>Even the eschaton will become imminent before it becomes immanent!

Winner, best typo flame of 2007.

(Oops, to fast on the "ignore" and "send" commands.)

pyotr
--
pyotr filipivich
With a tight organization with a clear doctrine, coupled with
strong party discipline - I think we could bring about complete
anarchy!
 >> Stay informed about: War with the Solarian League? 

On Oct 14, 5:32 pm, Quadibloc <jsav....RemoveThis@ecn.ab.ca> wrote:
> rlbell.ns....RemoveThis@gmail.com wrote:
> > This brings up another issue: Why is a line intended for only 3VDC
> > connectable to a line carrying 220VAC via a soft switch?
>
> That is a good question. But there is a reasonable answer, too.
>
> Basically, one imagines a computer system which is expected to work as
> specified, but which is capable of connecting pretty much anything to
> anything else. Which is not a problem, unless the computer system is
> tampered with.

This actually isn't a reasonable answer, as it is also an issue if the
computer code is buggy. Dynamically assigning what each conductor in
a cable is for at connect time is well outside the set of programs for
which the halting problem is solvable, so the absence of any bugs is
impossible to prove. You can expect random lobotomization of
equipment to be a weekly, if not daily, occurence. It also makes
interface glitches that much harder to debug, as you do not know if
the problem is in the cable, the communications protocol, or a
disagreement between the two entities about what the pinout is
supposed to be.

Unless the computer is a partly lobotomised unit that cannot execute
any program outside of the set of programs for which the halting
problem is solved (we call them industrial grade PLC's), for the
purposes of failure mode analysis, computers are never expected to
work as intended.

Given there is no foreknowledge of how the connection ports are
configured ahead of time, any system connectable to a soft
configurable connector will not have any lines that cannot the maximum
likely voltage (So if Harkness can cause 220VAC to go down a line
configured as a data line, the machine on the end of the cable will
display an error code, but will be undamaged).
>
> Presumably, Haven viewed that as being just as impossible as, say,
> hypnotizing the crew and getting *them* to connect the 220 VAC to the
> 3 volt circuits in the suits.
>
> Putting "everything" under the computer's control reduces the number
> of times when you have to get a crew member to physically connect a
> wire from here to there. Also, note that educational standards on
> Haven were lacking - and the members of the secret police would
> naturally want things to be done for them.

They have to use cables to connect things together, anyways. You may
as well force them to use the right cable. The PRH may not have the
same education base as the Star Kingdom, but they are intelligent and
educated enough to produce interstellar warships, so we can assume
that they know how to do it properly.
>
> If one had multiple *small* computer systems that controlled specific
> individual systems, indeed one could validate the software for each
> one, and not give any of them the power to do something silly. But if
> you have one big computer system, and your focus is on making the user
> interface into "Hey, computer, would you do this for me", well...
> you're taking the chance that just the sort of thing will happen as
> the novel illustrated. Haven was sloppy and lazy - the books have been
> telling us that all along.
>

Warships do not have one big computer system. They have a number of
independent systems with an overall supervisory system that collects
and distributes data to prevent either a single hit crippling the
whole ship, or to prevent a glitch in an APU controller taking down
the combat information systems (USS Yorktown?).
 >> Stay informed about: War with the Solarian League? 

pyotr filipivich wrote:
> After a Computer crash and the demise of civilization, it was learned
> Quadibloc <jsavard DeleteThis @ecn.ab.ca> wrote on Sun, 14 Oct 2007 16:35:12 -0700
> in alt.books.david-weber :
> >pyotr filipivich wrote:
> >> Yeah, but it isn't as dramatic as "Warp Core Breach immanent."
> >
> >I didn't know that each of us had a Warp Core Breach beside us, every
> >day of our lives - that the Warp Core Breach surrounds us, and in it
> >we live, move, and have our being.
> >
> >Even the eschaton will become imminent before it becomes immanent!
>
> Winner, best typo flame of 2007.
>
> (Oops, to fast on the "ignore" and "send" commands.)

I thought it was just a good-natured ribbing, not a "flame". I did not
intend to cause offence.

John Savard
 >> Stay informed about: War with the Solarian League? 

After a Computer crash and the demise of civilization, it was learned
Quadibloc <jsavard RemoveThis @ecn.ab.ca> wrote on Mon, 15 Oct 2007 16:22:47 -0700
in alt.books.david-weber :
>pyotr filipivich wrote:
>> After a Computer crash and the demise of civilization, it was learned
>> Quadibloc <jsavard RemoveThis @ecn.ab.ca> wrote on Sun, 14 Oct 2007 16:35:12 -0700
>> in alt.books.david-weber :
>> >pyotr filipivich wrote:
>> >> Yeah, but it isn't as dramatic as "Warp Core Breach immanent."
>> >
>> >I didn't know that each of us had a Warp Core Breach beside us, every
>> >day of our lives - that the Warp Core Breach surrounds us, and in it
>> >we live, move, and have our being.
>> >
>> >Even the eschaton will become imminent before it becomes immanent!
>>
>> Winner, best typo flame of 2007.
>>
>> (Oops, to fast on the "ignore" and "send" commands.)
>
>I thought it was just a good-natured ribbing, not a "flame". I did not
>intend to cause offence.

No offense intended, none was taken. I actually thought it very
clever. "Its a keeper!"


thanks
pyotr
--
pyotr filipivich
The two oldest cliches in the book are "The Good Old Days were
better." and "After all, these are Modern TImes."
 >> Stay informed about: War with the Solarian League? 

After a Computer crash and the demise of civilization, it was learned
"rlbell.nsuid@gmail.com" <rlbell.nsuid.DeleteThis@gmail.com> wrote on Mon, 15 Oct
2007 04:22:04 -0000 in alt.books.david-weber :
>On Oct 14, 5:32 pm, Quadibloc <jsav....DeleteThis@ecn.ab.ca> wrote:
>> rlbell.ns....DeleteThis@gmail.com wrote:
>> > This brings up another issue: Why is a line intended for only 3VDC
>> > connectable to a line carrying 220VAC via a soft switch?
>>
>> That is a good question. But there is a reasonable answer, too.
>>
>> Basically, one imagines a computer system which is expected to work as
>> specified, but which is capable of connecting pretty much anything to
>> anything else. Which is not a problem, unless the computer system is
>> tampered with.
>
>This actually isn't a reasonable answer, as it is also an issue if the
>computer code is buggy. Dynamically assigning what each conductor in
>a cable is for at connect time is well outside the set of programs for
>which the halting problem is solvable, so the absence of any bugs is
>impossible to prove. You can expect random lobotomization of
>equipment to be a weekly, if not daily, occurence. It also makes
>interface glitches that much harder to debug, as you do not know if
>the problem is in the cable, the communications protocol, or a
>disagreement between the two entities about what the pinout is
>supposed to be.
>
>Unless the computer is a partly lobotomised unit that cannot execute
>any program outside of the set of programs for which the halting
>problem is solved (we call them industrial grade PLC's), for the
>purposes of failure mode analysis, computers are never expected to
>work as intended.
>
>Given there is no foreknowledge of how the connection ports are
>configured ahead of time, any system connectable to a soft
>configurable connector will not have any lines that cannot the maximum
>likely voltage (So if Harkness can cause 220VAC to go down a line
>configured as a data line, the machine on the end of the cable will
>display an error code, but will be undamaged).

I suspect, Harkness would not have directed more down the line than
the line would allow. I pulled numbers from my ... head on the relative
voltages. The same effect could be accomplished running 5 volts DC down
circuits built for 3VDC.

That said, there are other industrial design issues that Haven has
had to deal with. Starting with the problem of finding trained and
competent designers, let alone enough of them to fix everything. For
every guy who can translate a cocktail napkin into a technical drawing,
there are dozens more who can't.
I suspect that one of the constraints of Haven military design and
construction is "the Kalishnikov effect"; the designers know they have
to build systems which must function even when maintained by "peasants".
So no doubt they are trying to make things idiot proof, knowing that as
they do so, The People Ministry of Education is perfecting better and
more clever idiots. And some of them are in design.

So we have the conflicting intent of making things fool proof, which
requires more components to prevent mismatched connections; and at the
same time trying to keep the number of components to a minimum, because
you know some idiot is going to hooked the wrong ones up.


--
pyotr filipivich
The two oldest cliches in the book are "The Good Old Days were
better." and "After all, these are Modern TImes."
 >> Stay informed about: War with the Solarian League? 

On Oct 16, 5:58 am, pyotr filipivich <ph....DeleteThis@mindspring.com> wrote:
> After a Computer crash and the demise of civilization, it was learned
> "rlbell.ns...@gmail.com" <rlbell.ns....DeleteThis@gmail.com> wrote on Mon, 15 Oct
> 2007 04:22:04 -0000 in alt.books.david-weber :
>
>
>
> >On Oct 14, 5:32 pm, Quadibloc <jsav....DeleteThis@ecn.ab.ca> wrote:
> >> rlbell.ns....DeleteThis@gmail.com wrote:
> >> > This brings up another issue: Why is a line intended for only 3VDC
> >> > connectable to a line carrying 220VAC via a soft switch?
>
> >> That is a good question. But there is a reasonable answer, too.
>
> >> Basically, one imagines a computer system which is expected to work as
> >> specified, but which is capable of connecting pretty much anything to
> >> anything else. Which is not a problem, unless the computer system is
> >> tampered with.
>
> >This actually isn't a reasonable answer, as it is also an issue if the
> >computer code is buggy. Dynamically assigning what each conductor in
> >a cable is for at connect time is well outside the set of programs for
> >which the halting problem is solvable, so the absence of any bugs is
> >impossible to prove. You can expect random lobotomization of
> >equipment to be a weekly, if not daily, occurence. It also makes
> >interface glitches that much harder to debug, as you do not know if
> >the problem is in the cable, the communications protocol, or a
> >disagreement between the two entities about what the pinout is
> >supposed to be.
>
> >Unless the computer is a partly lobotomised unit that cannot execute
> >any program outside of the set of programs for which the halting
> >problem is solved (we call them industrial grade PLC's), for the
> >purposes of failure mode analysis, computers are never expected to
> >work as intended.
>
> >Given there is no foreknowledge of how the connection ports are
> >configured ahead of time, any system connectable to a soft
> >configurable connector will not have any lines that cannot the maximum
> >likely voltage (So if Harkness can cause 220VAC to go down a line
> >configured as a data line, the machine on the end of the cable will
> >display an error code, but will be undamaged).
>
> I suspect, Harkness would not have directed more down the line than
> the line would allow. I pulled numbers from my ... head on the relative
> voltages. The same effect could be accomplished running 5 volts DC down
> circuits built for 3VDC.
>
> That said, there are other industrial design issues that Haven has
> had to deal with. Starting with the problem of finding trained and
> competent designers, let alone enough of them to fix everything. For
> every guy who can translate a cocktail napkin into a technical drawing,
> there are dozens more who can't.
> I suspect that one of the constraints of Haven military design and
> construction is "the Kalishnikov effect"; the designers know they have
> to build systems which must function even when maintained by "peasants".
> So no doubt they are trying to make things idiot proof, knowing that as
> they do so, The People Ministry of Education is perfecting better and
> more clever idiots. And some of them are in design.

The one thing that you are forgetting is that the Havenite navy,
thanks to the necessities of the Duquesne Plan, is not as badly off as
the Republic as a whole. The naval academies, before the coup, were
where someone with ambition to do more than collect the BLS could get
an education as useful as that made available to the legislaturists.
In a real sense, it will be very much like the Former Soviet Navy--
three year conscripts doing all of the labor, supervised by career
officers who actually know what to do. Maintenance is done by pulling
the defective module, replacing it with a known good module from
stores, and sending the defective unit to a repair depot, as
opportunities present. The really skilled people (aside from combat
leaders) are husbanded away where they will do the most good-- the
depots.

Planning for the systems being operated by people that do not
understand them makes some of Harkness's exploits even less likely.
To prevent the problem of the rating having the wrong cable for the
job, they will not have the system able to dynamically reconfigure the
pinout of a connector, as they cannot expect the technician to
understand how to take advantage of it. Instead, they will pick an
interface that allows the transmission of power and data, say
USB(appropriate number here), and that will be the only cable needed
for all equipment in the entire fleet (if all of the connectors are
the same, the cables may as well be identical, too). If you expect
everyone on the vessel to be mere users, except for the Intelligence
officer, the EW officer, and a few members of their staff, everyone
else will be locked out from doing anything except running canned
applications. Anyone else who tries to do any modifications to the
program either is an idiot in need of severe punishment, or a spy. It
is possible that only the guys at the depots have administrator access
to shipboard systems.
 >> Stay informed about: War with the Solarian League? 

After a Computer crash and the demise of civilization, it was learned
"rlbell.nsuid@gmail.com" <rlbell.nsuid.DeleteThis@gmail.com> wrote on Thu, 18 Oct
2007 01:19:46 -0000 in alt.books.david-weber :
>On Oct 16, 5:58 am, pyotr filipivich <ph....DeleteThis@mindspring.com> wrote:
>> After a Computer crash and the demise of civilization, it was learned
>> "rlbell.ns...@gmail.com" <rlbell.ns....DeleteThis@gmail.com> wrote on Mon, 15 Oct
>> 2007 04:22:04 -0000 in alt.books.david-weber :
>>
>>
>>
>> >On Oct 14, 5:32 pm, Quadibloc <jsav....DeleteThis@ecn.ab.ca> wrote:
>> >> rlbell.ns....DeleteThis@gmail.com wrote:
>> >> > This brings up another issue: Why is a line intended for only 3VDC
>> >> > connectable to a line carrying 220VAC via a soft switch?
>>
>> >> That is a good question. But there is a reasonable answer, too.
>>
>> >> Basically, one imagines a computer system which is expected to work as
>> >> specified, but which is capable of connecting pretty much anything to
>> >> anything else. Which is not a problem, unless the computer system is
>> >> tampered with.
>>
>> >This actually isn't a reasonable answer, as it is also an issue if the
>> >computer code is buggy. Dynamically assigning what each conductor in
>> >a cable is for at connect time is well outside the set of programs for
>> >which the halting problem is solvable, so the absence of any bugs is
>> >impossible to prove. You can expect random lobotomization of
>> >equipment to be a weekly, if not daily, occurence. It also makes
>> >interface glitches that much harder to debug, as you do not know if
>> >the problem is in the cable, the communications protocol, or a
>> >disagreement between the two entities about what the pinout is
>> >supposed to be.
>>
>> >Unless the computer is a partly lobotomised unit that cannot execute
>> >any program outside of the set of programs for which the halting
>> >problem is solved (we call them industrial grade PLC's), for the
>> >purposes of failure mode analysis, computers are never expected to
>> >work as intended.
>>
>> >Given there is no foreknowledge of how the connection ports are
>> >configured ahead of time, any system connectable to a soft
>> >configurable connector will not have any lines that cannot the maximum
>> >likely voltage (So if Harkness can cause 220VAC to go down a line
>> >configured as a data line, the machine on the end of the cable will
>> >display an error code, but will be undamaged).
>>
>> I suspect, Harkness would not have directed more down the line than
>> the line would allow. I pulled numbers from my ... head on the relative
>> voltages. The same effect could be accomplished running 5 volts DC down
>> circuits built for 3VDC.
>>
>> That said, there are other industrial design issues that Haven has
>> had to deal with. Starting with the problem of finding trained and
>> competent designers, let alone enough of them to fix everything. For
>> every guy who can translate a cocktail napkin into a technical drawing,
>> there are dozens more who can't.
>> I suspect that one of the constraints of Haven military design and
>> construction is "the Kalishnikov effect"; the designers know they have
>> to build systems which must function even when maintained by "peasants".
>> So no doubt they are trying to make things idiot proof, knowing that as
>> they do so, The People Ministry of Education is perfecting better and
>> more clever idiots. And some of them are in design.
>
>The one thing that you are forgetting is that the Havenite navy,
>thanks to the necessities of the Duquesne Plan, is not as badly off as
>the Republic as a whole. The naval academies, before the coup, were
>where someone with ambition to do more than collect the BLS could get
>an education as useful as that made available to the legislaturists.
>In a real sense, it will be very much like the Former Soviet Navy--
>three year conscripts doing all of the labor, supervised by career
>officers who actually know what to do. Maintenance is done by pulling
>the defective module, replacing it with a known good module from
>stores, and sending the defective unit to a repair depot, as
>opportunities present. The really skilled people (aside from combat
>leaders) are husbanded away where they will do the most good-- the
>depots.
>
>Planning for the systems being operated by people that do not
>understand them makes some of Harkness's exploits even less likely.
>To prevent the problem of the rating having the wrong cable for the
>job, they will not have the system able to dynamically reconfigure the
>pinout of a connector, as they cannot expect the technician to
>understand how to take advantage of it. Instead, they will pick an
>interface that allows the transmission of power and data, say
>USB(appropriate number here), and that will be the only cable needed
>for all equipment in the entire fleet (if all of the connectors are
>the same, the cables may as well be identical, too). If you expect
>everyone on the vessel to be mere users, except for the Intelligence
>officer, the EW officer, and a few members of their staff, everyone
>else will be locked out from doing anything except running canned
>applications. Anyone else who tries to do any modifications to the
>program either is an idiot in need of severe punishment, or a spy. It
>is possible that only the guys at the depots have administrator access
>to shipboard systems.

All of which goes to show Harkness's prowess! He managed to subvert
Peep computer systems which were too dumb to fool!

Hmmm, I suspect that Operation Beatrice was just his cleaver way to
cover his tracks as he "disappeared" to their version of Brazil, or the
Swiss banking planet.

pyotr

--
pyotr filipivich
The two oldest cliches in the book are "The Good Old Days were
better." and "After all, these are Modern TImes."
 >> Stay informed about: War with the Solarian League?