From ISOL to NASA, and beyond
John H Meyers
to port TRI1 (a triangle solver) from HP48 to HP49.
Here is a very simplified program, illustrating an essential
part of the operation of this UserRPL triangle solver program,
which works perfectly well in an HP48, in any angle mode:
\<< DEG 0. 'A' STO 60. 'B' STO 90. 'C' STO
@ now we are going to solve for 'A' and store the new answer:
'A+B+C=ACOS(-1.)' 'A' ISOL EQ\-> EVAL SWAP STO \>>
Result on HP48:
ISOL gives the elementary and obvious result 'A=180-C-B'
(insert a HALT after ISOL to see this working),
and the final answer (A=30) gets stored into 'A'
Result on HP49 [1.19-4 or 1.19-5]:
"Const -> Value mode off?"
"Radian mode on?" [ugh!]
ISOL Error: No solution found
It's a misfortune if we can only solve a triangle
if all angles are given in radians, but it's catastrophic
that the calculator can't even isolate 'A' at all,
given 'A+B+C=K' (a correct answer is 'A=K-C-B',
as any HP48 will give you).
Somebody will probably say that ISOL would work on the HP49
if you first purged the variable that you want to isolate
(which you never had to do on an HP48),
but then suppose, in a more general kind of solver,
that you want first to try to isolate a variable, and if you can
do so, then you want to directly calculate its value, but if ISOL
fails, then you want to try the numeric root-finder, starting with
a "guess" that had already been stored in the variable (this
is exactly how the HP18C/19B/17B solvers work, for example).
Unfortunately, before you can even try ISOL on the 49, you must
first *purge* the variable, but then if ISOL still fails, you must
somehow *unpurge* the variable (to get its stored value back,
to use as a starting guess for the numeric root-finder);
besides all these contortions, if the problem involves
angles, the only angle mode that you now can use is radians.
Although the HP49 excels in many ways, the fact that the some of the
simplest and most elementary things, already perfected even before
the HP48 (even in the HP18C/19B/17B financial calcs, for example),
become awkward and difficult on the HP49 (or can't be done at all)
subtracts from the traditional elegance of the HP product line,
and seems to be noticeable by many people.
Sometimes these things are discovered after something has already
been built; for example, after the Citicorp Center office tower
in New York City was completed, the architect came to discover
that its wind resistance had been calculated per code requirements
to resist winds coming straight onto the face of the structure,
but if instead a wind came straight towards a corner, then
the building could very well break apart in a strong wind.
Well, even though it would not be possible to start building
Citicorp Center again from the ground up, this engineer did
find a way to "fix" the building, which in this case was
to weld reinforcements onto the main supporting columns
(which weren't exactly made to original specs, it turns out),
and since Citicorp Center is still standing today,
this intelligent post-construction fix
appears to have saved it from an ignoble end.
I don't know whether any post-construction fixes can be applied
to those things which have stirred distress in HP49 users,
but many great things have already been accomplished thus far,
since the HP49 first came on the market, and if one or two
more things can be dealt with equally as cleverly,
then there would be some further admiration for it
(and maybe even more sales, before the sun finally sets
on the empire of great HP calculators).
Citicorp (now Citigroup) Center:
http://www.pbs.org/wgbh/buildingbig/wonder/structure/citicorp.html
The following lecture makes very interesting reading in its entirety,
particularly about making "small changes" to existing designs,
but you can search for "Ethics" to read only the Citicorp story:
http://www.me.gatech.edu/me/publicat/AugTranscript.htm
Start here (and follow succeeding parts)
for a case history in moral courage and engineering ethics:
http://www.onlineethics.org/moral/LeMessurier/2.html
The same case is repeated here (in case links become broken):
http://www.luisaponte.com/fiftyninestory_crisis.htm
"The Fifty-Nine-Story Crisis", Joe Morgenstern,
The New Yorker, vol. LXXI, no. 14, May 29, 1995, pp. 45-53
"If they built buildings they way we build software...."
http://www.infowar.com/iwftp/risks/Risks-17/risks-17.16.txt
More pics, specs, and history of Citicorp Center:
http://www.thecityreview.com/citicorp.html
http://www.arch.usyd.edu.au/kcdc/caut/html/Citi/CC.htm
"Structural Engineering is the art of
modelling materials we do not wholly understand,
into shapes we cannot precisely analyse,
so as to withstand forces we cannot properly assess,
in such a way that the public has no reason to suspect
the extent of our ignorance."
- Dr .A.R. Dykes
British Institute of Structural Engineers, 1976.
Another case, which did not end happily:
http://www.onlineethics.org/moral/boisjoly/RB-intro.html
"NASA'S very nature since early space flight was to force
contractors and themselves to prove that it was safe to fly...
[but on this occasion] NASA placed MTI [Morton Thiokol Inc.]
in the position of proving that it was not safe to fly
instead of proving that it was safe to fly."
The Feynman report:
http://www.uky.edu/~holler/msc/roles/feynrept.html
Yesterday was the 15th anniversary of the explosion of the Space
Shuttle Challenger on January 28, 1986, which killed all aboard.
How about applying the same standards of responsibility
to governments, as, under similar pressure from corporations
to make money, or to merely to garner votes to gain power,
they engineer the future through tinkering with the economy,
with the environment, or with food (no need to label GMOs?),
and with bombing as a substitute for influencing societies
to act more righteously, with no basis whatsoever to assume
that these do not imperil the safety and future of the world.
"The means by which we live
has outdistanced the ends for which we live.
Our scientific power has outrun our spiritual power.
We have guided missiles and misguided men."
- Martin Luther King Jr
Strength to Love, 1963
-----------------------------------------------------------
With best wishes from: John H Meyers http://www.mum.edu
Sent via Deja.com
http://www.deja.com/
antonio_...@my-deja.com
defaulted to Radians Mode. I was trying to use SOLVE with the equation
R = __Vo^2*sin(2*x)__ for x in degrees.
g
Prof. Parisse answered back and said that Radians mode was the most
logical mode to solve in. Other people in this newsgroup have argued
with me why I complain about such things, compared to the TI89. Now it
appears I have some company. I see no reason why SOLVE couldn't easily
be used in either Radians or Degree mode. And the HP49's inability to
isolate the variable in your example below is beyond belief.
Antonio Delmonico
In article <955lok$bf7$1...@nnrp1.deja.com>,
John H Meyers
of changing things from 48 to 49 (and in the CAS) were
completely followed through, to perceive all their further
repercussions (which is the same theme illustrated in the
remaining previous digressions into other areas of engineering).
For example, there isn't much problem when doing isolated
manual operations, but when those same commands are embedded
in programs, then it starts to get problematic.
Somehow the CAS got "stuck" with having to change these modes;
I don't know if there could have been another way -- could
some CAS commands save and restore the modes which they
require changing? Would that change now break anything else?
You see how some "fixes" almost routinely generate a new bug;
that's the price of the thin margin of safety in the whole OS,
which is "lean, but fragile" -- maybe too complex, who knows.
The ISOL matter is not beyond belief; many related problems came
with changing some HP48 fundamentals, in a manner that seemed
to make it easier to add Algebraic mode to the HP49,
and this was one more such minor casualty of the primary
metamorphosis of the HP48, into something which is
supposed to have evolved in the direction of the TI89,
while retaining something of its old self at the same time
(i.e. a schizophrenic product :)
Could it have been averted by making a *separate* CAS command
for the new version of ISOL, and retaining the old HP48 version?
Probably not, since certain internal elements of old HP48 functions
were dropped, so that old HP48 ISOL might not work any longer anyway.
Can "retrofits" be devised and applied?
Maybe; maybe not.
Maybe it's of little importance, except to old HP48 fanatics :)
Maybe it's of little importance, if the market for these relics
has faded away, anyway, and they've become throw-away hand-downs
for much younger folk, just to take to school, not out into space
with NASA, for astronauts to have to rely on in a pinch, in case
on-board computers failed, to navigate their way back home :)
-----------------------------------------------------------
With best wishes from: John H Meyers <jhme...@mum.edu>
Jorge Alberto
Yes, Mr. Meyers, I wholeheartedly agree with you, although you're way
too diplomatic for my taste, if I may say so. You see, Monsiure Parisse
seems to care mostly about his own students' concerns first and foremost
as he has rudely retorted many time in this NG. Let's just hope none of
his students finds himself or herself stuck in between Earth and Mars
with no other means of navigation besides and old HP49G (hopefully with
ROM 1.10 loaded in it!) he he he he ....
snip...
--
The absolute does not exist... I am absolutely right on this!
Parisse Bernard
> too diplomatic for my taste, if I may say so. You see, Monsiure Parisse
> seems to care mostly about his own students' concerns first and foremost
> as he has rudely retorted many time in this NG. Let's just hope none of
> his students finds himself or herself stuck in between Earth and Mars
> with no other means of navigation besides and old HP49G (hopefully with
> ROM 1.10 loaded in it!) he he he he ....
Yes I have rudely responded many times, but I was never the first one
to shoot.
The main problem is that many people here are used to the 48 and
expect exactly the same behavior even in limit situations. For
example here isolating a variable that is assigned. You must be
aware that we had extremely few time to make the 49 and we could not
isolate old 48 CAS commands from the change we made in the CAS and
this has boundary effects. But I think I have proved, I have never
refused to examine request from user when they are politely asked.
But please remind that I make improvements on my spare time, and
for the moment I do not have time to do some changes that might
look easy to do but require a lot of changes.