INTEGRALS on HP48
Bill Wright
It seems to be a simple enough problem, (answer is -x*cos(x)+sin(x), of
course) so why is the HP unable to handle it?
Does anyone know of more powerful programs which I can use on the 48 to
integrate symbolically?
sergi
Wright) wrote:
> Why is the HP48G unable to find the indefinite integral of x*sin(x)?
by a ti92! it is too difficult for your hp
> It seems to be a simple enough problem, (answer is -x*cos(x)+sin(x), of
> course) so why is the HP unable to handle it?
> Does anyone know of more powerful programs which I can use on the 48 to
> integrate symbolically?
no! i don't think so
--
\|/
o o
-------------------------------------------------ooo-O-ooO----
Sergi-Serda
e-mail : se...@worldnet.net
Home page : http://www.worldnet.net/~serda/sergi.html
--------------------------------------------------------------
okrabbe
>> Does anyone know of more powerful programs which I can use on the 48 to
>> integrate symbolically?
>no! i don't think so
integration program. There are many incredible programs out there and
I figured someone ought to be able to make a stupid symbolic
integrator.
omid
Gabriel B. Gonzalez
wrote:
>I was wondering why this is. Why has no one written a better
>integration program. There are many incredible programs out there and
>I figured someone ought to be able to make a stupid symbolic
>integrator.
Mainly because the 48 was not made for that. It is seriously
underpowered on the CPU side, and table-matching programs take too
long to do their thing...
gg
------------------------------------------------------------------------
Gabriel B. Gonzalez Information and Computer Science & Biology
gab...@uci.edu University of California, Irvine
10311...@CompuServe.COM
------------------------------------------------------------------------
Mark Wilson
>
> >> Does anyone know of more powerful programs which I can use on the 48 to
> >> integrate symbolically?
> >no! i don't think so
> integration program. There are many incredible programs out there and
> I figured someone ought to be able to make a stupid symbolic
> integrator.
>
You pretty much answered your own question there. "Stupid" symbolic
integrator is exactly what such a thing would be on the HP. Remember,
HPs are made as tools to aid in learning, not replace it. Who really
needs symbolic integration (that fits in a pocket; hmmmm....)?
Basically, a student who doesn't want to bother with actually learning
the processes themselves. A professional engineer or technical person
sitting behind some workstation with hardcore dedicated symbolic math
packages would whip out the HP to do the work (they probably wouldn't use
the mainframe either; thats why we have reference manuals).
Thats why I have such a problem, philosophically speaking, with the
TI-92. With that sitting on your desk, what would you possibly learn in
Calculus? How to enter an equation, and......o.k., thats about it.
Besides, the number of analytically solvable integrations is relatively
small; whats the big deal?
Mark Wilson
--
Beware of Shadows...
They move when you're not watching them...
Bill Wright
(sergi) wrote:
> In article <4ge0pm$i...@reader2.ix.netcom.com>, okr...@ix.netcom.com
> (okrabbe) wrote:
>
> > >> Does anyone know of more powerful programs which I can use on the 48 to
> > >> integrate symbolically?
> > >no! i don't think so
> > I was wondering why this is. Why has no one written a better
> > integration program. There are many incredible programs out there and
> > I figured someone ought to be able to make a stupid symbolic
> > integrator.
> >
> > omid
> i know it is sad, but they prefer to make new games!!! !!!!!!!!!!
>
> --------------------------------------------------------
With the TI-92 out there, someone really should make an effort to program
a really good symbolic intergration program rather than all the (albeit
terrific) games. There's no reason that the HP shouldn't be able to
integrate these two (which the TI-92 does with ease):
x*sin(x) or 1/(x*sq(1-x^2))
Source of any good integration programs welcome....
sergi
> >> Does anyone know of more powerful programs which I can use on the 48 to
> >> integrate symbolically?
> >no! i don't think so
> I was wondering why this is. Why has no one written a better
> integration program. There are many incredible programs out there and
> I figured someone ought to be able to make a stupid symbolic
> integrator.
>
> omid
i know it is sad, but they prefer to make new games!!! !!!!!!!!!!
--
Mark Friedgan
hard to write you take a calculus book take the look up tables in
the back and write a simple look-up and recombine the equations to
make them look different for example the hp can't expand cos(x)^2 to
1/2+1/2(cos(2x)) why cause no one bothered to write the program so
instead of using the hp for calculus most people sit there and play
stupid albeit good looking games.
The Isaac Sublett
> In article <4gg8sk$4...@news.acns.nwu.edu>, hub...@prairienet.org (Mark
> Friedgan) wrote:
> I think it would probably be harder than just looking up tables due to all
> the hairy issues of u substitution, but likely I am wrong. If it was this
> easy, I can't see why no one would make a good integration program yet.
> Are the programmers out there really not good enough with the HP to make a
> good symbolic integration program?
>
> I'd love for someone to prove me wrong, I think it can't be done on the HP...
>
> -bill
I'm certain that it can be done, but the limited resources of the saturn
and the available memory make it un-feasible. In order to even get a
reasonably comprehensive table of integrals it would take at least 20k of
compressed data (look in your calc book - there is a lot of stuff
there). Then, in order to decompress the table and analyze the desired
integral against it would take up at least another 10-15k of memory.
With that right there, all s/sx & g users are left out in the cold. The
program would also run pitifully slow, unless sone initial conditions
were set (command line options as it were). To define a user-friendly
interface for these vital flags/settings would furthur add to the memory
requirements.
I suggest learn the math, and wait for some future calc that will do all
this more efficiently.
Isaac
- "Do not take council of your fears" - Gen. George Patton
************************ sub...@nag.cs.colorado.edu ************************
******University of Colorado, Boulder ** Department of Computer Science******
WWW: http://rtt.colorado.edu/~sublett/tis.html
Bill Wright
Greg Humphreys
: Thats why I have such a problem, philosophically speaking, with the
: TI-92. With that sitting on your desk, what would you possibly learn in
: Calculus? How to enter an equation, and......o.k., thats about it.
What a completely bogus argument. If you ever need to evaluate an integral in
real-world applications, you can whip out your table of integrals. I've
always been of the opinion that all math tests should be open book -- we
should concentrate on applications of learned formulas, not memorization of
the chain rule.
Sorry for the non-hp48 content, but I couldn't pass this one up.
--
The artist formerly known as HUMPER
http://humper.student.princeton.edu/~humper
"Lime and limpid green, a second scene
A fight between the blue you once knew."
Richard Paschal
: Mark Wilson (mwi...@afit.af.mil) wrote:
: : Thats why I have such a problem, philosophically speaking, with the
: : TI-92. With that sitting on your desk, what would you possibly learn in
: : Calculus? How to enter an equation, and......o.k., thats about it.
: What a completely bogus argument. If you ever need to evaluate an integral in
: real-world applications, you can whip out your table of integrals. I've
: always been of the opinion that all math tests should be open book -- we
: should concentrate on applications of learned formulas, not memorization of
: the chain rule.
That's the student response and it proves the preceeding point. When you
get out in the real world, you encounter integrals which aren't included
in any table. You usually have to do a lot of work and then you might
have to hire an $$expert$$ mathematician. Even that strategy is known to
fail. What do you do then? The HP48 just wasn't designed to be the
answer to everybody's wishes.
Richard Paschal
--
---------------------
rpas...@primenet.com
Bill Wright
The Isaac Sublett <sub...@nag.cs.colorado.edu> wrote:
> I suggest learn the math, and wait for some future calc that will do all
> this more efficiently.
Well, I guess the "future is here" in a sense, as the TI-92 does an
incredible amount of integrals...I haven't found many integrals which will
trip it up yet. That's not to say it is perfect, though--but it does do
*a lot* more than the HP.
-bill
Mark Friedgan
would do at least the basic 100 integration lookups and it need not
look very nice and all just so it shows me the either the solution
or a way at the solution so it maybe shows the steps as it tries
them.
For example if i was to ask it to find S(x^2)*(e^x)dx it would show
me that it's using the method of integration by parts.
and say
1: x^2*e^x+....
2: x^2*e^x-2x*e^x+...
3: x^2*e^x-2x*e^x+2e^x-...
4: x^2*e^x-2x*e^x+2e^x-0
5: Done
that's just one of the five or six methods that HP devs forgot to
include in the internal routines for the 48. that one is actually
pretty simple to do i made a program that does just that method but
i need something thats able to perform mutations on functions so
they can be applied to different forms and then put into forms that
are known!
For example if a question has a sin^m*cos^n i want it to be able to
expand and solve that. That is actually pretty simple but i don't
wish to have a list that i have to go looking through everytime i
need an integral solved because that takes to much time if you are
attempting a more complex problem (i.e. physics) that requires you
to solve a large system of equations but you are stuck there sitting
searching for thew right integral form. (just my two cents but worth
considering)
M M FFFF
M m m M F
M m M FFFF They all know the world but do they know the
M M F universe!
M M F
Mark Wilson
>
> Mark Wilson (mwi...@afit.af.mil) wrote:
>
> : Thats why I have such a problem, philosophically speaking, with the
> : TI-92. With that sitting on your desk, what would you possibly learn in
> : Calculus? How to enter an equation, and......o.k., thats about it.
>
> What a completely bogus argument. If you ever need to evaluate an integral in
> real-world applications, you can whip out your table of integrals. I've
> always been of the opinion that all math tests should be open book -- we
> should concentrate on applications of learned formulas, not memorization of
> the chain rule.
>
Whoa there big fella. Missed the point of my original post completely.
As I said, the idea of symbolic integrals on a handheld of any ilk was
dumb. In the real world you whip out Mathcad, or the ever handy integral
tables. Check my original post again, and right above the part you
included is exactly where I said that.
As to your particular beef, is there any difference between looking up
some vague "u" defined integral form in a table, and actually applying it
yourself, and typing a function in then hitting the solver button? YES!!
The TI-92 owner only needs to be able to enter an equation and press the
magic button. I'm pretty sure I could give my daughter in first grade
that power in about fifteen minutes. I am also sure that it will be
another ten years of math fundamentals before she could utilize tables.
I agree that tables are useful and imply a great deal of understanding of
the mathematics involved; with tables you would actually learn a process.
And if the TI-92 was only an electronic reference, then I wouldn't care.
But I stand by the fact that anyone with half a brain could use a TI-92
to solve the same thing and have absolutely no understanding or
appreciation of what just happened
Thats what I said the first time.....
_A_K_
<sub...@nag.cs.colorado.edu> wrote:
>I'm certain that it can be done, but the limited resources of the saturn
>and the available memory make it un-feasible. In order to even get a
>reasonably comprehensive table of integrals it would take at least 20k of
>compressed data (look in your calc book - there is a lot of stuff
>there). Then, in order to decompress the table and analyze the desired
>integral against it would take up at least another 10-15k of memory.
>With that right there, all s/sx & g users are left out in the cold. The
>program would also run pitifully slow, unless sone initial conditions
>were set (command line options as it were). To define a user-friendly
>interface for these vital flags/settings would furthur add to the memory
>requirements.
How about just a program for storing integral tables ??
It would be nice to be able to view an integral formula in the nice
equation write view, but I haven't found any way of doing that.
I have been using DBASE, a really cool library, for storing tables.
It works pretty well, but I still cant get the equation writer
view....
Any one have any suggestions?
Ian S. Nelson
>Mark Wilson (mwi...@afit.af.mil) wrote:
>: Thats why I have such a problem, philosophically speaking, with the
>: TI-92. With that sitting on your desk, what would you possibly learn in
>: Calculus? How to enter an equation, and......o.k., thats about it.
>What a completely bogus argument. If you ever need to evaluate an integral in
>real-world applications, you can whip out your table of integrals. I've
>always been of the opinion that all math tests should be open book -- we
>should concentrate on applications of learned formulas, not memorization of
>the chain rule.
It depends on what kind of math test it is. If you are some sort of engineer
then go ahead and plug-n-chug. If you are taking a math test, then you should
know the material and understand it.
Bernard Parisse
features of the HP48, I think I should give my own version of
symbolic integration.
1/ A function expressed as combination of elementary functions has
(almost everywher) no primitive expressed as combination of
elementary functions. The Risch algorithm gives an answer to the
question. But it is *not* implemented in Derive -> if your TI92
can't find the primitive of a function, it does not prove that
this function has no primitive expressed as combination of elementary
function.
2/ We know algorithm to integrate some classes of functions: e.g.
rational fractions, rational trigonometric fractions, fractions
with one radical of a 2nd order polynomial, polynomial*exponential.
3/ We may use integration by part and change of variables to
rewrite the function and apply 2/
The TI-92 implements probably points 2/ and 3/. For example, if you
have an inverse function (like ln or atan, ..), you only need to
isolate it and integrate by part. All the work is done automatically,
you only have to press the button. But if it doesn't work, you don't
know why.
ALGB implements point 2/ excluding polynomial*exponential and
partially change of variables. I have not yet found a good interface
for integration by part. You have to know a little bit how it works
to use it. For example, if you want to integrate
'sin(x)^4*cos(x)^4'
you shoud not use INT but LIN.
I see the usage of calcs as a tool which computes without errors,
not as a replacement of your brain. Even if you use maple, mathematica,
axiom or whatever, you should know how it works. Last month, I solved
a double integration using maple and I had to choose right the
order of integration because it worked only with one choice.
That's why you need to understand how it works.
Bernard Parisse.
Mayer Goldberg
> I think it would probably be harder than just looking up tables due to all
> the hairy issues of u substitution, but likely I am wrong. If it was this
> easy, I can't see why no one would make a good integration program yet.
> Are the programmers out there really not good enough with the HP to make a
> good symbolic integration program?
>
> I'd love for someone to prove me wrong, I think it can't be done on the HP...
General symbolic integration is a very hard problem: It's not always
possible to detect which forms are integrable and which are not. That
problem was reduced to the Halting Problem a long time ago (in the
50's or 60's).
This said, it is fairly simple to write very good integration programs
for large classes of functions. It has been done time and time again
in MACSYMA, Maple, and lots of other symbolic mathematics packages. It
can OF COURSE be done for the HP.
This said, I'd need a hell of a lot better development environment
than what I currently have, before I attempted to actually write some
code. The Scheme/LISP code is straightforward. If there were a
Scheme/LISP compiler for the HP48, then the problem would be quite
doable. I'm sure HP itself has a better development environment than
just writing RPL/Saturn in an editor, downloading to a 48 and testing
on the 48.
I'm pretty confident HP will improve their symbolic integrator in
future models.
Mayer
Mark Wilson
(Big, heavy hearted sigh...)
You are close to what I've been trying to say, but there is one important
point you are missing: when engineers don't understand EVEN ONE of the
tools they are using, i.e. math, physics, how their pocket monster came
up with root Pi in the denominator, there is a simple consequence; people
usually die.
The HP-48G series was originally looked at by some (myself included) as
too much, especially when you looked at the equation library. That chunk
of ROM is dedicated to most of the basic principles of all engineering
disciplines. But a closer look revealed that the equation library is
more of an electronic reference; it is most useful when you forget the
volume of a sphere, or want to track units automatically. In reality its
pretty nice, but pretty limited, only capable of simple algebraic
rearranging and solving. And that is what makes it so nice, it simply
presents the forms for you, but you still need to do the work.
The Ti-92 provides no understanding to anyone. If a little knowledge is
dangerous, then what is no knowledge?
My final .02 on this...;)
Justin Dossey
> get out in the real world, you encounter integrals which aren't included
> in any table. You usually have to do a lot of work and then you might
> have to hire an $$expert$$ mathematician. Even that strategy is known to
> fail. What do you do then? The HP48 just wasn't designed to be the
> answer to everybody's wishes.
>
The real problem seems to be that there isn't any algorithm to solve integrals
symbolically out there. That is, the derivative function on the HP48 doesn't
seem to rely on lookup tables; it symplifies until it gets dy/dx*x, then does its
thing. If there were an algorithm to reduce x*sin(x) into _|~ (that's an
integral sign) x dx etc, lookup tables would be irrelevant. Any bites? This is
the future of mathematics. . .
Lookup tables will eventually (i hope) go the way of the slide rule.
--
{
justin dossey
houston, tx u.s.a.
dos...@flex.net
http://www.flex.net/users/dossey
}
okrabbe
The arguements that a symbolic integrator should not exist because
people should understand integration is ridiculus. What is the point
of all those games? Why not get a gameboy or some other piece of
junk? Why do they have mathematica and Derive? This is because
sometimes there are more important ways to be spending your time than
integrating some huge problem. And how can you let that piece of crap
Ti-92 beat and hp?
Omid
Mark Friedgan
Actually you can use and editor side by side with an emulator and
that works very well. I suggest using it in win95 so you can use a
compiler side by side with both editor and emulator and you have a
full fledged development environment!
Andy
>
> You know that just pisses me off. A better integrator is not that
> hard to write you take a calculus book take the look up tables in
> the back and write a simple look-up and recombine the equations to
> make them look different for example the hp can't expand cos(x)^2 to
> 1/2+1/2(cos(2x)) why cause no one bothered to write the program so
> instead of using the hp for calculus most people sit there and play
> stupid albeit good looking games.
The comma is located to the right of the M key and the period to the
right of that. That's a pretty impressive run-on sentence though.
Mark Friedgan
Hey Man don't knock my run-on sentences, half the time they are very
much correct. If you would like discussing the intricacies of
sentences this is not the newsgroup to do it in.
Robert Worne
>Why is the HP48G unable to find the indefinite integral of x*sin(x)?
>
>course) so why is the HP unable to handle it?
>
>integrate symbolically?
There is a simple ROM application that does the job quite nicely for this
problem, it does:
Direct Integration, Integration Tables, Definite integral, Partial Fractions,
Substitution, Integration by parts, and Improper integrals.
All with a nice user interface, one of the catches are that you SHOULD
know which method to apply.
It's the Bringslid Math Pac I, available at Educalc.
As an example, to do an integration by parts of
S x^2 sin(x)dx (I gotta use 'S' for integral sign)
you would:
Enter independent variable, 'X'
Answer Y/N if you want it to show all the work, 'Y'
put in 'u' 'X^2'
put in 'v' 'sin(X)'
and it comes out with:
Sx^2*sin(x)dX=
x^2*-cos(x)-
S-(cos(x)*(2*x)
)dX =
x^2*-cos(x)-
2*x*-sin(x)+
S-(sin(x)*2)d
X =
x^2*-cos(x)-
2*x*-sin(x)+
2*cos(x)
In about 10-15 seconds for most problems, up to 45 seconds for more
complex ones.
--
Robert Worne NeXT-OS/2-MacOS
Starving CS Undergrad...Sorry, I don't *do* Windows! I'd rather starve...
-----BEGIN GEEK CODE BLOCK-----
Version: 3.1
GCS/P/S d-?>pu s+:+> a- C++$>++++ UX++++>$ P+>+++ L+ !E !W++ N+++ !o--
!K w--- O++$ M+ V PS>--- !PE+ Y+ !PGP- t@ 5++ X+++ R- tv b+>++ DI !D
G e>+++ h--- r++ y+++**
------END GEEK CODE BLOCK------
Erek Hutto
>
> How about just a program for storing integral tables ??
> It would be nice to be able to view an integral formula in the nice
> equation write view, but I haven't found any way of doing that.
> I have been using DBASE, a really cool library, for storing tables.
> It works pretty well, but I still cant get the equation writer
> view....
>
> Any one have any suggestions?
My suggestion is to make totally sure that HP betters their integration
routines at the ROM level in the future. I find myself often defending
HP against brainless TI users when it comes to integrals in my Calculus
class that the HP can't answer. Certainly every TI (save 92) can't do
it. My question, however is along those lines: How good is TI at
solving integrals? More importantly: Is it better, and by how much?
Erek Hutto
Alfred A. Arduengo
>
> My suggestion is to make totally sure that HP betters their integration
> routines at the ROM level in the future. I find myself often defending
> HP against brainless TI users when it comes to integrals in my Calculus
> class that the HP can't answer. Certainly every TI (save 92) can't do
> it. My question, however is along those lines: How good is TI at
> solving integrals? More importantly: Is it better, and by how much?
>
> Erek Hutto
>
>
Actually, the 92 is quite good at solving integrals. Numerically it is
acceptably fast and symbolically it so far has only failed on about 2% of
the integrals I have tested it with. Frankly, I love its math
capabilities and use it a lot in Mechanics and Circuit theory classes.
But, I still consider myself an HP user because the TI lacks in offering
the user control of the calc. And I miss the RPN when I use it.
Erek Hutto
> What bugs me is that alot of people are saying that you shouldn't
> NEED the integration abilities of the TI..
These people are probably related to the same people who just couldn't give
up their trusty slide rule.
> that you should study study
> study. So why does the HP48 have symbolic functions at all? Hell.. why
> not just use a $10 casio calculator? :) Obviously some people want more
> power than others. I don't know about you but once in awhile I'd like to
> be able to check if the solution to an integral problem I got by solving
> by hand is right or even partially correct.. at those times a decent
> symbolic integration function in the HP48gx would be invaluable..
Good call! Our policy in my AP Calculus class is that we can use any
calculator we can get our hands on, but we must show all of our work. What's
more, many of our tests are designed to where a calculator is sometimes
useful and sometimes not. This means that on integrals, it's a good tool to
check my work. It even helps me find my absurdly ridiculous errors. . .if
it can handle the blasted integral, but integrals are nebulous and it's
impossible to make something to calculate EVERY integral EVER. HP still
needs to better their integration.
Erek Hutto
Erek Hutto
>WHo gets the better grade? I do. Who knows more? You.
> Who gets shafted? Both--I don't learn anything, and you don't get
> credit.I doubt very seriously there would be much of a grade difference. I tear
through integrals. What's more, AP tests will eat through a 92 user.
Especially on the calc-free portion. I get the credit and the knowledge.
That's the great thing about learning something with proficiency, you get the
credit. I'm still not sure that the qwerty board would be allowed on the AP
test. I've been on many tests where they specified no QWERTY. Also, the way
my class works out, we are in Cal 2 material right now. All we do is figure
integrals. Physics II AP is where I'm applying it. I still think that
future HP's need to better they're integration, though.
> In a perfect world, our nice little TI-92 boys would sit at home, and use
> the box to help them learn, having it demonstratem the integration steps
> as they do the homework. In the real world, these bozos will know little
> more than function entering, and they will blow your socks off grade
> wise.Again, I seriously doubt it. One of the kids in our class never had a
calculator and is always in HEAVY competition with me, even though I have
an HP. In the job world YES THE WOULD, only if I didn't know function
entering either. The problem is that school doesn't prepare anyone for the
real world. No one solves problems by themselves in an engineering field.
Likewise, there is no restrictions like no calculators, or even computers.
> As a student, in direct competition with these slackers, you should be
> most concerned with anything that doesn't give you a level playing field.HA HA HA!!!!! There is no such thing as a level playing field (Marx was
wrong). Was that a Clinton '92 quote? Anyone can make it. Give me a break.
I could survive easily if I had to destroy my HP. Again I site the guy that
is my competition in grades.
Erek Hutto
Oleg Rodionov
> know which method to apply.
Another catch, you forgot to mention, is how much this thing costs!!!
Mark Wilson
>
> JEEjohn wrote:
>
> > Interesting discussion.
> > Now from the employer's standpoint: They want workers who
> > can do their work effectively, competently, and ****fast****.
> > If fast means using a TI92 (or whatever) - GREAT ! - as long as it's
> > right.
> > SO - the colleges should educate people how to solve problems
> > correctly and quickly by whatever tools available (whether a slide rule
> > or a Cray 3 or whatever) & an understanding of how the tool solves
> > the problem. Unfortunately, many educators find this
> > somewhat difficult to do & thus take the easy way out.
>EXACTLY!!! YES!!! And, that is what I as a student have to keep in mind.
> I'm not going to be sitting in my desk answering integral problems by myself
> with my pencil as a future engineer.
>
> > Just my 2 cents worth2 extremely valuable cents.
>
> HP is still a better product than HP.
> Erek Hutto
Erek, I am an engineer. I do sit at my desk all by myself and do my own
work. I am also expected to do it right the first time without hand
holding. Don't confuse the "team" concept of engineering with
babysitting. If you become an engineer, you will be a valuable asset to
your employer who will not pay another $30-50k to have someone else redo
your work. And there will be no limit on how to do it. But consider
this; MATHCAD, MATHEMATICA, MATLAB, Fortran IMSL libraries, HP-48 & TI-92
all have known and unknown bugs. If a student (and I'm not talking about
you or your smart friends; I'm talking about the not so bright crowd)
only knows how to use a TI-92 to solve a problem, but not how to do it
themselves, then how will they know when the program's answer makes no
sense?
If you know what you are doing, then go for it. If you are learning,
then this (TI-92) is too much
As for the "do the work fast" comment above, Engineers are pretty much
either using workstations or reference books. Does anyone honestly think
the Beoing 777 engineers did that all on handheld calculators? How about
any of it? As the first paperless jet, I know they didn't. And heaven
help us if they screwed up somewhere.
Been there, doing it
Bill Wright
right to use symbolic integration, I think we've moved away from the
thread's original purpose.
WHY is it not possible to create a better symbolic integrator on the HP if
the 92 can? We all know the HP is a better calculator, so what's the
problem? And if it is possible, why is it that no one has created one
yet? Too fixated on creating new, better, faster games?
Is it simply a matter of lack of memory for storing tables, or is there a
more important issue at hand?
-bill
JEEjohn
writes:
>(snip)
>If a student (and I'm not talking about
>you or your smart friends; I'm talking about the not so bright crowd)
>only knows how to use a TI-92 to solve a problem, but not how to do it
>themselves, then how will they know when the program's answer makes no
>sense?
>
>If you know what you are doing, then go for it. If you are learning,
>then this (TI-92) is too much
>
>As for the "do the work fast" comment above, Engineers are pretty much
>either using workstations or reference books. Does anyone honestly think
>the Beoing 777 engineers did that all on handheld calculators? How about
>any of it? As the first paperless jet, I know they didn't. And heaven
>help us if they screwed up somewhere.
Hey Mark! I'm an engineer too! I work for the Army on tanks & trucks.
I get to work on projects where I can say that 1" steel is ok to use!
Very few engineers get to work on programs like the Boeing 777.
Even then, analysis of safety critical parts are likely triple checked,
and then triple checked again to be sure. Point is that very few
incompetent engineers get near to working on any thing important,
because the managers are aware that such people exist. God help
the company that doesn't do this (& they wouldn't have my sympathy).
Maybe we should be finding someway to complain to the schools
about turning out such "idiots". We have where I work a fella with
a Masters degree who I swear can't add 2+2 !!!
Myself, I use a lot of reference books that I'm trying to copy over
equations & database onto my HP48GX. Lots of work though.
Then hopefully I'll be able to solve my problems without using
the reference books.
Second point is: the HP48GX (with a 1M card) can do A LOT
if you think about it & with a little bit of work (OK - a lot of work).
I suspect that we're both talking about the same thing
but from different perspectives.
Old enough to NOT know everything AND to know that the
old engineering whizs & gurus weren't infallible either !!
John Edry
JEE...@aol.com
Mark Wilson
> Of course it is possible, and no, it won't take up all that much
> memory. It's not the storing of tables that's so memory
> consuming. It's the temporary garbage created when various
> possibilities are explores (for example, integration by parts,
> etc.). As to why these were not implemented in the current hp48
> ... your guess is as good as mine.
>
> Mayer
Could you please post a simple demonstration with a small table of this
program? You obviously have some idea on how to implement the object of
everyones desire, and perhaps with some other input, it could be worked
out.
Brian Downing
Gabriel B. Gonzalez <gab...@uci.edu> wrote:
>Really? If I key in '4*6+(5/7)', it takes more keystrokes on the HP
>than on my $15 "cheap" calc. Of course the difference in coming out
>with the answer is negligible. If you try to enter it using RPN, I
>can almost guarantee it'll take longer.
Huh? Some of us Think and Live RPN. I always have to go back and
change parentheses when entering stuff infix on a TI. Anyway:
4 * 6 + ( 5 / 7 ) =/enter 9 keystrokes
4 sp 6 * 5 sp 7 / + 9 keystrokes
It's all a matter of what you are used to. I would definatly say that
RPN has more potential speed than algebraic.
-bcd
--
*** Brian Downing / bdow...@somat.com
*** http://www.somat.com/~bdowning/
*** "It does not do to leave a live dragon out of your
*** calculations, if you live near him." --- J.R.R. Tolkien
Brian Downing
> They've avoided being banned on most tests (by limiting the IR transfer)
> and kept the RPN method, by far the fastest method.
Even if they hadn't limited the IR range, I think most sane test-givers
would allow the "black electrical tape" solution...
Oh yeah. This from the '96 AP "Bulletin for students and parents":
"Calculators with typewriter-style (QWERTY) keyboards will not be
permitted, and calculators with infrared capabilities will be permitted
only if the serial port is covered with electrical tape."
Huh? The _SERIAL PORT_?!!! Boy, am I gonna have fun walking in to the
test with my GX's serial port covered. hehe
Erek Hutto
>
> In article <4hd4dj$g...@news.service.uci.edu>,
> Gabriel B. Gonzalez <gab...@uci.edu> wrote:
> >Really? If I key in '4*6+(5/7)', it takes more keystrokes on the HP
> >than on my $15 "cheap" calc. Of course the difference in coming out
> >with the answer is negligible. If you try to enter it using RPN, I
> >can almost guarantee it'll take longer.
>
> Huh? Some of us Think and Live RPN. I always have to go back and
> change parentheses when entering stuff infix on a TI. Anyway:
>
> 4 * 6 + ( 5 / 7 ) =/enter 9 keystrokes
> 4 sp 6 * 5 sp 7 / + 9 keystrokes
>
> It's all a matter of what you are used to. I would definatly say that
> RPN has more potential speed than algebraic.
It depends on which equation you use, but RPN keystrokes are always < or =
to Algebraic. Me (and two other friends) walked in without practicing AT ALL
to a Calculator contest. Guess what? 1st, 2nd, and 3rd. It seems everyone
else had a Casio or TI. I heard many complaining afterwards:
"How come I practiced so much and didn't even come close."
We were 60 points higher than the nearest competitor (a TIer). That shows
you right there. The sponser of the Calculator contest has said that NO ONE
has ever won at State without a RPN (more specifically HP). There are the
facts. RPN coupled with Stack beats Algebraic with Queuy(sp?) 9 out of 10
equations. It's a given fact.
The '4*6+(5/7)' is an algebraic on something intended for mainly RPN use. Of
course it's going to have more key strokes than some garbage algebraic calc.
Reverse Polish Notation is THE method. Hands down.
Erek Hutto
Jacob G. Schwartz
> In article <4hd4dj$g...@news.service.uci.edu>,
> Gabriel B. Gonzalez <gab...@uci.edu> wrote:
>> Really? If I key in '4*6+(5/7)', it takes more keystrokes on the HP
>> than on my $15 "cheap" calc. Of course the difference in coming out
>> with the answer is negligible. If you try to enter it using RPN, I
>> can almost guarantee it'll take longer.
> Huh? Some of us Think and Live RPN. I always have to go back and
> change parentheses when entering stuff infix on a TI. Anyway:
> 4 * 6 + ( 5 / 7 ) =/enter 9 keystrokes
> 4 sp 6 * 5 sp 7 / + 9 keystrokes
> It's all a matter of what you are used to. I would definatly say that
> RPN has more potential speed than algebraic.
How about 5 sp 7 / 6 + 4 * for a total of 8 keystrokes? It dosn't take much forethought to
do it "inside out" like the original HP RPN manuals always advocated....
Jake Schwartz
Charlie Watts
>> 4 sp 6 * 5 sp 7 / + 9 keystrokes
'4*6+5/7' 8 keystrokes
even with
algeabraic 's
these things DO use proper precedence, after all
Charlie Watts
>> 4 sp 6 * 5 sp 7 / + 9 keystrokes
'4*6+5/7' eval 8 keystrokes
even with
algeabraic 's
these things DO use proper precedence, after all
>5 sp 7 / 6 + 4 *
this is wrong... try 5 sp 7 / 4 sp 6 * + still 9
JEEjohn
>JEEjohn wrote:
>
>> Maybe we should be finding someway to complain to the schools
>> about turning out such "idiots". We have where I work a fella with
>> a Masters degree who I swear can't add 2+2 !!!
>
>I know a guy that came out of Texas A&M and tried to ground some
electricity
>with PVC pipe! What an idiot! I'm a High School senior and know better
than
>that! I do admit that I don't know enough about the engineering field,
yet,
>to state how Engineers do the job, but I still feel that they do things
>totally different than the artificial world created by school. I maybe
wrong
>by previously saying that engineers don't sit all by themselves, but then
>again I maybe right, I just don't have enough experience, yet. Sorry if
I
>offended anyone. . .
>
>Erek Hutto
>
>
You're correct in many ways, in that we very rarely solve a problem
from scratch, we almost always use reference manuals for previously
solved solutions. **But**, an engineer is supposed to know if that
solution is valid, i.e. are the assumptions correct? Your "friend"
obviously "assumed" that the pipe would ground some electricity.
If you need to discuss engineering work more, let me know
& we can start e-mailing each other, since the subject not
strictly related to the HP48G's.
John Edry
Mel Tsai
>hum...@Princeton.EDU (Greg Humphreys) writes:
>
>>Mark Wilson (mwi...@afit.af.mil) wrote:
>
>>: Thats why I have such a problem, philosophically speaking, with the
>>: TI-92. With that sitting on your desk, what would you possibly learn in
>>: Calculus? How to enter an equation, and......o.k., thats about it.
>
>>What a completely bogus argument. If you ever need to evaluate an integral in
>>real-world applications, you can whip out your table of integrals. I've
>>always been of the opinion that all math tests should be open book -- we
>>should concentrate on applications of learned formulas, not memorization of
>>the chain rule.
>
>It depends on what kind of math test it is. If you are some sort of engineer
>then go ahead and plug-n-chug. If you are taking a math test, then you should
>know the material and understand it.
I own the HP-48GX, the TI-85 and the TI-92. I agree that the typical
math student should only use these calculators as tools, not as "grade
getters," and this is the way I use them. Sorry HP users, but if I
had a choice between my TI-92 and my HP on a test or any other
application, I would NEVER use the HP. This is not because I want to
"cheat" on a calc test or whatever, but I believe it is simply a more
user-friendly and definitely more powerful calculator. If the hp had
better symbolic integration already, no one would be complaining about
"calculator ethics" and this discussion wouldn't be taking place.
Granted, the TI-92 has its drawbacks like limited expandability and
less programming power, along with some other things. However, the
latter could soon change with MShell (I'm a zshell programmer and
can't wait to hack my TI-92). Otherwise, I believe the TI is better
in every way.
As an hp user I have gotten used to the interface and rather like its
features, but that's just the point. You have to get "used" to it.
It's simply not an intuitive, easy-to-use calculator, and considering
that the TI-92 is much more powerful in many ways, it doesn't make
much sense to use an HP when a TI is available, no matter what the
application for it is. In fact, if the TI-92 didn't have a querty
keyboard, I'm sure it would be allowed on most of the tests the hp is
allowed on. Try this experiment: give your hp to someone who
normally doesn't use one, and then have them add two numbers together
or try to graph an equation. Then take a picture of the blank
expression on their face after they hand it back to you! That's what
I mean when I say that it's not intuitive.
Sooner or later, the world must "wake up" and realize that technology
will always advance, whether they like it or not. The moral
implications of "only knowing how to use the calculator, not the math
itself" are inevitably thrown out the door, and It's their choice
whether they want to jump onto the bandwagon.
Mark Wilson
Note: Lots O' snips...
>
> I own the HP-48GX, the TI-85 and the TI-92. ... if I
> had a choice between my TI-92 and my HP on a test or any other
> application, I would NEVER use the HP. I believe it (TI-92) is simply a more
> user-friendly and definitely more powerful calculator.
>
> Granted, the TI-92 has its drawbacks like limited expandability and
> less programming power, along with some other things. However, the
> latter could soon change with MShell (I'm a zshell programmer and
> can't wait to hack my TI-92). Otherwise, I believe the TI is better
> in every way.
> Sooner or later, the world must "wake up" and realize that technology
> will always advance, whether they like it or not. The moral
> implications of "only knowing how to use the calculator, not the math
> itself" are inevitably thrown out the door, and It's their choice
> whether they want to jump onto the bandwagon.
You mention several times how the TI-92 is far superior to the HP 48.
SInce this topic comes up frequently, maybe you could highlight some
specifics for those who are undecided out there. First, lets get a
definition of power. Is power speed in certain applications? Is it a
friendly interface? Is it more built in functions? Is it
programmability? Is it the ability to play the first generation arcade
games? I can see the HP and the TI both outpacing each other, depending
upon your definition of power. To me, power is as much speed as
versatility. The real problem is we can't compare like a 286 to a 486 or
pentium; what we have between HP and TI are apples and oranges.
As an undergrad, the TI has certain features that make it attractive. As
a grad student, and a practicing engineer, the HP is a very flexible
tool.
Maybe we should try to compare on specifics instead of general
statements; I would like to know what the TI is capable of. I would be
willing to scare you all with real world, practical things that I've done
on El HP that I still don't know could be done on other platforms without
writing your own code.
Waiting to hear from you,
MArk Wilson
G. Allen Morris III
Note: Lots O' snips...
>
> I own the HP-48GX, the TI-85 and the TI-92. ... if I
> had a choice between my TI-92 and my HP on a test or any other
> application, I would NEVER use the HP. I believe it (TI-92) is simply a more
> user-friendly and definitely more powerful calculator.
>
> Granted, the TI-92 has its drawbacks like limited expandability and
> less programming power, along with some other things. However, the
> latter could soon change with MShell (I'm a zshell programmer and
> can't wait to hack my TI-92). Otherwise, I believe the TI is better
> in every way.
Yes but I think that I would rather have hp-200lx running derive.
Or better yet a nice notebook wich Mathimatica. This would even
be better if I had an cellular connection to the WWW.
The point I'm tring to make is that TI-92 is not really a calculator.
The HP-48 in my opinion is right on the edge of being a calculator,
I would like it to be smaller, and anything bigger and you might
as well get a palm top with a hard disk.
--
G. Allen Morris III http://www-inst.EECS.Berkeley.EDU/~gam3
University of California, Berkeley ga...@cory.EECS.Berkeley.EDU
Junior, Department of Electrical Engineering and Computer Science
Mel Tsai
>You also mentioned that TI
>had *SOME* problems. Yes it does...It is a totally inferior product. That
>is not bias. That's a fact.
Why do you say this? I believe this statement is highly debatable. I
never meant that the TI had "problems", but rather that it could have
been made better in some ways. Remember, I have both of them (and
have used both extensively), and I believe the TI beats it in almost
every way (Why and how is a totally different thread. Find someone
with a TI and see for yourself). Besides, there is nothing more
intimidating to other students than whipping out my TI-92 in the
middle of class! Add a pocket protector and you're pretty much
labeled a genius!
>When people know about RPN, and it's uses, they realize how
>weird and ridiculous algebraic is. Algebraic for me is now NOT INTUITIVE.
>I can't even hardly figure it out anymore. It is the most NON INTUITIVE
>thing in the entire world. Queue(sp?) is almost as bad.
>
>Erek Hutto
The problem is that every single math book and every single math class
is taught in an "algebraic" style! This is how math began. TI and
pretty much every other calculator manufacturer since the start has
capitolized on this, and it seems to work. This fact alone gives HP
users a disadvantage, especially if you find it non-intuitive.
I realize that most dedicated HP users find their calculators to be
superior to most other products, simply because they find it an
efficient and effective way to solve problems. That's fine with me!
But this does not mean that it's the calculator to end all calculators
(I'm not saying the TI-92 is either), nor does it mean that RPN is the
"best" interface for a calculator. Having a closed mind will hurt you
more than it helps you.
Mark Wilson
Snip..
....Remember, I have both of them (and
> have used both extensively), and I believe the TI beats it in almost
> every way (Why and how is a totally different thread. Find someone
> more than it helps you.
to that effect, yet you have not backed up your statement with any data.
If graphing, split screens, and symbolic integration are the only
features of the TI you use, then I can see your point. But what about
number crunching routines? Matrix algebra (i.e. size capable of
handling, what capabilities does it have etc?)? Does the TI do
differential equations? Fourrier transforms? Does it have a stack? I
don't have access to TIs (I haven't even seen the TI-92 on sale around
here yet), so your help is appreciated.
Mark Wilson
arlenk
>Hi folks !
>
>
>> How about 5 sp 7 / 6 + 4 * for a total of 8 keystrokes? It dosn't take
>
>You won a calc contest by getting the wrong answer fast ?!? ;-)
>
>Using algebraics it requires 9 on the HP, using RPN I got 9, too.
>All of this results in 24.7142857143
>The RPN version given above results in 26.8571428572
>
>Are there faster ways which are correct ?
>
Well if you just put the + in the right place:
5 sp 7 / 6 4 * +
It is still 8 places and gives you the correct answer.
Besidees,
'6*4+7/8'
Is only 8 key strokes!
Gabriel B. Gonzalez
Morris III) wrote:
>The point I'm tring to make is that TI-92 is not really a calculator.
>The HP-48 in my opinion is right on the edge of being a calculator,
>I would like it to be smaller,
So you differentiate "calculator" from "computer" by size alone. They
HAVE to be teaching you more than that at Berkeley...
gg
------------------------------------------------------------------------
Gabriel B. Gonzalez Information and Computer Science & Biology
gab...@uci.edu University of California, Irvine
10311...@CompuServe.COM
------------------------------------------------------------------------
Erek Hutto
> And this was Erek Hutto:
> >Of course...That's how I won my Calc contest!!!!
>
> You won a calc contest by getting the wrong answer fast ?!? ;-)
My bad, I must not have been thinking that day...sorry...of course, the
simpler algebraics are harder to find a difference, but the bigger the
algebraic, more efficient RPN will be. I'm too lazy to show an example,
though.
Erek
Charlie Watts
>5 sp 7 / 6 4 * +
>It is still 8 places and gives you the correct answer.
>Besidees,
no, it isn't:
5 sp 7 / 6 sp 4 * +
press 5 1
press sp 2
press 7 3
press / 4
press 6 5
press sp 6
press 4 7
press * 8
press + 9 keystrokes
>'6*4+7/8'
>Is only 8 key strokes!
correct - guys, as much as I prefer RPN, in
this case it just isn't _smaller_. I would
still say its better, just because it makes
much more sense.
charlie
Mark Wilson
> Mega snip...
Thanks for the info. Sorry for the snip, but I can't repost the whole
thing.
Some questions I still have:
Concerning the interface:
How many keystrokes does it take to get the natural log of a number? I
understand that its not a keybaord function. From what I can see of the
keyboard form a grainy bitmap, there aren't a lot of functions
immediately available.
The catalog idea is nice--I miss that from the HP-42S.
When you mention the equation "pretty print" feature, what happens if you
get some outrageously long polynomial on the stack? Does it do a word
wrap type action? Does the font vary to account for large equations?
What is a function stack, and how much memory is allocated for it? Is it
dynamically allocated? Is it like the HP stack, or do you specifically
have to tell it what to include? Is it only for functions?
If the 92 didn't do algebraic expansion easily, could you program it to?
Does it have the summation function built in?
Also, please check on the matrix size--on the HP, a 100x100 identity
matrix doesn't fit in 46k of RAM; how could the TI possibly do one that
big with only ~60k?
What are Z-shell and M-shell? How will they be superior to the HP? If
you've played the arcade games (columns, tetris, etc) on the HP, and the
equivalent on the TI-85, which was better (i.e. faster, smoother etc)?
Finally, you mentioned 98% of the integreals you tried on the 92 were
unitegratable? Is it that bad? Please clarify.
Again, thanks for the info.
Mark Wilson
Gabriel B. Gonzalez
Erek Hutto <shu...@vonl.com> wrote:
>It is a totally inferior product. That is not bias. That's a fact.
Since it is not a bias, but rather a fact, you surely are ready to
back up your claim with hard facts, you know, being the sociologist
and great computer scientist and engineer you are...
>When people know about RPN, and it's uses, they realize how
>weird and ridiculous algebraic is.
Algebraic ridiculous? Weird? Not me buddy. Which uses are there for
RPN that are not HP48 related? (actually, there are several
applications for it, but I bet you don't know them... )
>Algebraic for me is now NOT INTUITIVE.
Try and tell everyone you're not detached from reality...
>I can't even hardly figure it out anymore. It is the most NON INTUITIVE
>thing in the entire world.
Perhaps your problem is with math itself. If you knew the principles,
you could figure it out in prefix, infix, and postfix notation.
>Queuy(sp?) is almost as bad.
HAHAHAHAHA!!! What do poor old queues have to do with this? You're
not saying "kill them too," are you? Now you're a computer scientist
too!
Alfred A. Arduengo
>
> In article <4hstji$b...@agate.berkeley.edu> G. Allen Morris III wrote:
> > Yes but I think that I would rather have hp-200lx running derive.
> >Or better yet a nice notebook wich Mathimatica.
>
> Where does one find Mathimatica or Derive? Are they on the Net?
>
> Thanks,
> Frank Marsh
> FMa...@gnn.com
>
>
>
Sorry, these are commercial programs. You gotta PAY ;-)
justin dossey
> Where does one find Mathimatica or Derive? Are they on the Net?
Sorry, no. (not that I know of, or at least, not legally.)
They cost money; you can get them at the store.
the latest version I have of derive was written WAY back when-- it's
under 360K and text-based; lightning-fast. I wish all software were as
pretty as my old version of derive.
--
{
justin dossey
houston, tx u.s.a.
dos...@flex.net
http://www.flex.net/users/dossey
}
G. Allen Morris III
[snip]
| Algebraic Power:
| TI went all-out when designing the algebraic manipulation capabilities
| in the TI. People who see the way that it manipulates equations and
I think that the software that is doing all this good stuff is a program
called Derive. You can buy this software from Edu-Calc and run it on
a HP-200lx.
Bernard Parisse
need 10 minutes to get the right manipulation (and 10 minutes more
to be sure that you get the right answer). Reducing the number
of necessary operations is only of interest if you write a program.
Now, about the debate of algebraic vs RPN notation:
1/ I can't see why TI and HP are compared. The HP is clearly
superior since you can choose algebraic or RPN. Moreover, you
can use EquationWriter to enter your expressions. Of course, if I
compare the TI92 and the HP48, I choose a HP48GX and I use EQSTK
(or JAVA) to display the stack of math objects!
2/ for input of a non trivial expression, EquationWriter is for me
the best. Algebraic has a big drawback: tracing all this parenthesis.
(Even on a workstation with maple, I find that algebraic is not
suitable to enter complex expressions). RPN is very usefull to
manipulate existing expressions on the stack, but is not intuitive
to enter new expressions.
Bernard Parisse.
Mel Tsai
>Mel Tsai wrote:
>> Mega snip...
>
>Thanks for the info. Sorry for the snip, but I can't repost the whole
>thing.
>
>Some questions I still have:
>
>Concerning the interface:
>
>How many keystrokes does it take to get the natural log of a number? I
>understand that its not a keybaord function. From what I can see of the
>keyboard form a grainy bitmap, there aren't a lot of functions
>immediately available.
>
Actually, all the functions are immediately available, you just have
to type the function name in with the keyboard. Also, the natural
log button is a dedicated key, keeping relatively the same keys as the
TI-85. Most advanced functions, however, are easily accessible right
through the pulldown menus via the function keys.
To get the natural log, type "LN", "[number]", ")", "ENTER".
Another awesome feature (that I forgot to mention) is that it will
always give you an EXACT value for a function, not a decimal
approximation.. For instance, integrating tan(x) from "1" to "pi"
returns ln(cos(1)), not -.615626470386 as other calculators would (you
can tell it to give you a decimal approximation, though).
>The catalog idea is nice--I miss that from the HP-42S.
>
>When you mention the equation "pretty print" feature, what happens if you
>get some outrageously long polynomial on the stack? Does it do a word
>wrap type action? Does the font vary to account for large equations?
>
Large equations have a little "arrow" pointing to the right, and to
view the rest of the equation you cursor up to the equation and press
the right cursor to see more of it. Like I said, the screen is quite
large and the vast majority of functions fit right on the screen (font
sizes really don't change on the screen, rather strategically
positioned).
>What is a function stack, and how much memory is allocated for it? Is it
>dynamically allocated? Is it like the HP stack, or do you specifically
>have to tell it what to include? Is it only for functions?
>
It's more like a repository for previously evaluated functions. You
can tell the TI-92 how many entries it will hold into memory, up to
99. I usually keep it at 10 though, because I think it dynamically
allocates memory for the "stack", although it's not really a "stack"
in RPN's terminology. You can scroll up and select either the function
(equations flush left on the screen) or the answer (flush right),
pressing enter will paste it right onto the command line. It keeps
anything and everything, including functions and error messages right
on the screen.
>If the 92 didn't do algebraic expansion easily, could you program it to?
That would be a complicated program, and it would probably be
impossible to do with TI-BASIC anyways. It does have some provisions
for custom functions, though. For instance, on the command line you
can define "cot(x)" as "1/tan(x)" and from then on use cot(x) in any
integral or derivative you please. It's structure is similar to C++
in this respect, allowing you to return command-line values to
programs and functions.
> Does it have the summation function built in?
>
Yes, and so does the '82 and '85.
>Also, please check on the matrix size--on the HP, a 100x100 identity
>matrix doesn't fit in 46k of RAM; how could the TI possibly do one that
>big with only ~60k?
>
That's just what the manual says. In terms of reality I don't know.
Once you get past 100 by 100, what the hell are you doing anyways!?!
By the way, it's 70k. Since the TI-BASIC language is interpreted,
programs are small. Tetris (the biggest program I've seen) is about
4k! Things like graphics and animations take up the most memory,
though, and this is where the 70k gets small *very* fast.
>What are Z-shell and M-shell? How will they be superior to the HP? If
>you've played the arcade games (columns, tetris, etc) on the HP, and the
>equivalent on the TI-85, which was better (i.e. faster, smoother etc)?
>
The TI-85 has almost the same programming power as the TI-92, but one
day some people said, "why can't I hack into my calculator and run
programs *directly* off of the processor instead of through TI's basic
interpreted language?" Well, someone did it (using a clever trick),
and the shell ("Z" shell, since the processor inside is the Zilog Z80)
they created allows you to download raw machine code directly into
memory and run it, totally bypassing the calculator's rom. This has
allowed total programming freedom and some really good games have come
out. This, as you can imagine, revolutionized game programming on the
TI-85 and has literally taken over high schools across the country.
It's like having your own Gameboy in the middle of class! Go to the
Unofficial TI-Calculator Homepage and you will find about 30 games
made for this (I myself have programmed a relatively good version of
Columns). Unfortunately, there are really no "educational"
applications out, since games are mostly in demand. Since the HP
already has provisions for assembly language programming, the majority
of games are roughly the same size and quality as the ones I've seen
on the HP (I never did build the link for the HP, so I haven't seen
all the games). However, until more programmers learn to program in
Z80 assembly language, the HP games are bigger and better.
"MShell" is the theoretically equivalent version of Zshell, but for
the TI-92. When completed, games will be run on raw machine code and
will truly be awesome, utilizing the large screen and fast processor,
and hopefully allowing memory expansions and other new things.
>Finally, you mentioned 98% of the integreals you tried on the 92 were
>unitegratable? Is it that bad? Please clarify.
>
Sorry, that was a typo. I meant that 98% ARE integrable, and the
other's will just simply return the original typed-in function as the
answer.
G. Allen Morris III
| On Fri, 08 Mar 1996 15:53:32 -0800, Erek Hutto <shu...@vonl.com>
| wrote:
| >The deal is is that people
| >got hooked on algebraic, because it was able to win the market share.
| I always thought it was because algebraic has been the standard for
| thousands of years... Silly me!
Mohammed ibn Musa al-Khowarizmi (ca 825A.D.) is Mr. Algebra, so it
seems that algebraic notation must be only a little bit over a
thousand years old. I am sure that people said to his Al use
geometry it is thousands of years old and the Algebra is brand new.
| >Why? Because one company survived better than another company.
| There were mathematicians unsing algebraic notation long before there
| were calculator companies.
And there where people making and selling abacuses thousands of year
before algebra.
[snip]
justin dossey
> can find something similar on the net: it is mupad. I run it under
> Linux, and if I am not wrong the next version should
I dl'ed, compiled, and ran MuPAD... I wasn't too
impressed. It couldn't handle some of the integrals I
needed it to, and I'm spoiled to Derive. I'm in the
midst of trying out the other math programs for Linux...
I'll keep the newsgroup updated. I wish someone else
had done this... it's a pain.
G. Allen Morris III
| Morris III) wrote:
| >The point I'm tring to make is that TI-92 is not really a calculator.
| >The HP-48 in my opinion is right on the edge of being a calculator,
| >I would like it to be smaller,
| So you differentiate "calculator" from "computer" by size alone. They
| HAVE to be teaching you more than that at Berkeley...
And what would you propose that the difference is? Width of
the data bus? Speed of the processor? Number of keys on the
keyboard? Programmability? Weather or not it has Windows?
The only difference I see between a calculator and a computer today is
that the calculator is more portable and the keyboard is optimized for
numeric input rather than alphabetic input.
Paul R. Hart
> | There were mathematicians unsing algebraic notation long before there
> | were calculator companies.
>
> And there where people making and selling abacuses thousands of year
> before algebra.
I'll go out on a limb here, and state my opinion. I much prefer calculating in
RPN over anything else I've ever tried. With an RPN calculator you can breeze
through the most complicated mathematical expressions with ease, once you
understand the nature of RPN. Just one thing to remember: operators take their arguments
from the stack and place their results on the stack. Not too hard to use, with
a little practice and rethinking.
The edge that I see with RPN on the 48, when it comes down to it, is that you have an
arbitrary number of "memories" to store intermediate calculations. Each level
of the stack can be used to store an intermediate calculation, greatly
simplifying long and tedious calculations. Using one of the $20 scientific
calculators you might have three "memories", two stored memories and the
number in the display. With such a calculator you'll certainly find yourself
writing intermediate results on paper if you're called on to do any significant
calculations. That's where the real edge lies with the HP48. If you limit
yourself to the trivial "algebraic vs. RPN" calculations that have been proposed
here in comp.sys.hp48 lately you might not notice a difference, but the difference
becomes painfully obvious when you do some "industrial-strength" calculating.
-- Paul
--------------------------------------------------------------------------
Paul R. Hart Computer Systems Administrator
pa...@ee.byu.edu Electrical and Computer Engineering
(801) 378-5728 Brigham Young University, Provo, Utah
Bobby L. Childers II
>
> 5 sp 7 / 6 4 * +
>
> It is still 8 places and gives you the correct answer.
> Besidees,
>
>
> Is only 8 key strokes!
He must consider hitting Enter an extra keystroke. Poor, poor baby. Heh
heh....
Derrik Pates
dpa...@cavern.nmsu.edu
Gabriel B. Gonzalez
Morris III) wrote:
>Mohammed ibn Musa al-Khowarizmi (ca 825A.D.) is Mr. Algebra, so it
>seems that algebraic notation must be only a little bit over a
>thousand years old.
Thanks, I'm too lazy to look for the exact name and time in my math
history book!
I was merely trying to make the point that calculators and calculator
companies have had little effect on algebraic notation. It's not like
someone has a copyright on either postfix notation (RPN) or infix
("algebraic") for usage in calculators.
>I am sure that people said to his Al use
>geometry it is thousands of years old and the Algebra is brand new.
Unless I'm terribly wrong, geometry is a branch of mathematics and
algebra (in the context of this thread) is a kind of mathematical
notation. Completely different things.
justin dossey
> "Calculators with typewriter-style (QWERTY) keyboards will not be
> permitted, and calculators with infrared capabilities will be permitted
> only if the serial port is covered with electrical tape."
>
> Huh? The _SERIAL PORT_?!!! Boy, am I gonna have fun walking in to the
> test with my GX's serial port covered. hehe
I hate to break it to you, but the IR port is a type of serial port, if you get
technical...
justin dossey
>
> 4 * 6 + ( 5 / 7 ) =/enter 9 keystrokes
> 4 sp 6 * 5 sp 7 / + 9 keystrokes
>
> RPN has more potential speed than algebraic.
>
Get a little more complex, as suggested in "Hewlett-Packarard Calculator
Superiority" by (I think) Andrew Ho:
While smaller expressions utilize the same number of keystrokes whether
entered with RPN or algebraic entry modes, RPN saves an average of 5-10%
on keystroke usage on large expressions. Consider the example of the
larger real root returned by the quadratic equation. With algebraic
entry assuming the existence of a negation (Neg), square (Sq), and
square root (Sqrt) command as well as three numbers a, b, and c, we
would type the first line below; and with RPN, the second:
( Neg b + Sqrt ( b Sq - 4 * a * c ) ) / 2 * a = (21 keystrokes)
b Neg b Sq 4 a c * * - Sqrt + 2 a * / (16 keystrokes)
Already, a savings of five keystrokes in a relatively simple operation.
With larger expressions with multiple nested parenthesis, the keystroke
savings simply multiply.
I agree with this. Test it out.
Alfred A. Arduengo
>
> While smaller expressions utilize the same number of keystrokes whether
> entered with RPN or algebraic entry modes, RPN saves an average of 5-10%
> on keystroke usage on large expressions. Consider the example of the
> larger real root returned by the quadratic equation. With algebraic
> entry assuming the existence of a negation (Neg), square (Sq), and
> square root (Sqrt) command as well as three numbers a, b, and c, we
> would type the first line below; and with RPN, the second:
>
> ( Neg b + Sqrt ( b Sq - 4 * a * c ) ) / 2 * a = (21 keystrokes)
>
> b Neg b Sq 4 a c * * - Sqrt + 2 a * / (16 keystrokes)
> Already, a savings of five keystrokes in a relatively simple operation.
> With larger expressions with multiple nested parenthesis, the keystroke
> savings simply multiply.
>
> I agree with this. Test it out.
>
> --
>
I hate to be a smartass, but you should look again. Remember, SPC is
considered a keystroke as well as ENTER. 4 a c takes 5 strokes itself. In
addition, b NEG b should take 5 strokes otherwise you would get -bb.
G. Allen Morris III
[snip]
| I hate to be a smartass, but you should look again. Remember, SPC is
| considered a keystroke as well as ENTER. 4 a c takes 5 strokes itself. In
| addition, b NEG b should take 5 strokes otherwise you would get -bb.
The place that rpn saves you keystrokes is in terms that are in parnthses.
(1+2)*(3+4)*(5+6)= (18 strokes)
1 sp 2 + 3 sp 4 + 5 sp 6 + * * (14 strokes)
Justin Dossey
> =pentium; what we have between HP and TI are apples and oranges.
>
> You mean apples and PCs rite?
>
I would never consider the HP, with its programmability and all, anything
like an Apple computer.
Peter Woolliams
for a faster, more functions, more programmability calc. I want to programme
astronomical algorithms into it, this requires quite a bit of memory and
speed! How would you input something like:
z*COS(23*ABC-DC*DER)-..... (letters can refer to variable names)
in RPN, ie how do you put a large complex argument into a trig function.
My TI 85 had enough space for quite a good programme, but it has little
'fits' and crashes, so I have finally got cheesed off and want a new machine
that will not do this! As well as giving me more speed (vital) and memory
(vital), but at the same time not burning a cosmic sized hole into my stedent
sized wallet. The TI 92 looks good though lacks space and expandability , the
HP looks good though may be require 'getting used to' and also doesn't offer
the same level of symbolic maths and seems to have a wierd programming
language.
I am stuck as to what to do, I am currently waiting for TI and HP to release
the next 'Ultimate Calc' (until the next one comes along).
Peter Woolliams 2nd Year Undergraduate Physics Student Imperial College
"We came, we saw, we went away again" SICK
Seth Arnold
> I am stuck as to what to do, I am currently waiting for TI and HP to release
> the next 'Ultimate Calc' (until the next one comes along).
What requirements are you looking for? The ti-92 will differentiate and integrate pretty
damn quick, but there are things the HP does better.... eg, I want a simultanious equation
solver, (like SIMN) and a Multiple Equation Solver (like the built-in one. ) Also, aa units
conversion utility is sorely needed on the 92.
--
Seth Arnold | http://www.rdrop.com/~hideki | PGP available. Ask me.
Gabriel B. Gonzalez
Morris III) wrote:
>| So you differentiate "calculator" from "computer" by size alone. They
>| HAVE to be teaching you more than that at Berkeley...
>And what would you propose that the difference is? Width of
>the data bus? Speed of the processor? Number of keys on the
>keyboard? Programmability? Weather or not it has Windows?
There is no difference (you should know this), a calculator is a
specialized computer, running only one app all the time. IOW, the
tool is geared towards one purpose alone.
When one says "computer," one does not mean a CPU, since you'd have a
hard time time finding an appliance that does not have one at this
time. One means a "general purpose computer."
Within this context, the 48 is much more of a computer than any TI.
G. Allen Morris III
| How would you input something like:
| z*COS(23*ABC-DC*DER)-..... (letters can refer to variable names)
'z*COS(23*ABC-DC*DER)-.....'
| in RPN, ie how do you put a large complex argument into a trig function.
z 23 sp ABC * DC sp DER * - COS * ..... -
eric larson
>> =pentium; what we have between HP and TI are apples and oranges.
>>
>> You mean apples and PCs rite?
>>
> I would never consider the HP, with its programmability and all,
> anything like an Apple computer.
Right. The HP has a 4 bit CPU clocked at 2 MHz. My Apple has a 32 bit 604 RISC
CPU clocked at 150 MHz. My HP has 512K of RAM, expandable to 4 MB. My Apple has
64 MB of RAM, expandable to 1.5 gigabytes. My HP has a hard to learn text mode
operating system. My Mac has an operating system that everybody has been trying
to match for the past twelve years, without success. My HP has user RPL, Sys
RPL and machine code programming. My Mac has FORTRAN, Yerk, C, C++, machine
code, Java, Pascal, Object Pascal, Hypertalk, AppleScript, Frontier, Mogul,
Oberon, Modula, PERL, AWK, Tickle, Scheme, Lisp, Forth, Mops, and several other
languages. My HP has a serial port and an IR port. My Mac has full video and
audio I/O including PAL, SECAM and NTSC, dual SCSI busses, 10BaseT and AAUI,
dual RS422s, ADB, and two 17" monitors.
They aren't ANYTHING alike. The only problem is that I can't get my Apple into
my pocket, so I own one of each.
--
|Fidonet: eric larson 1:107/407
|Internet: eric....@shockwave.metronj.org
| Standard disclaimer: The views of this user are strictly his/her own.
raph
> I am stuck as to what to do, I am currently waiting for TI and HP to release
> the next 'Ultimate Calc' (until the next one comes along).
checked out the TI-92?
raph.
Charles Watts
larson) wrote:
>Right. The HP has a 4 bit CPU clocked at 2 MHz. My Apple has a 32 bit 604 RISC
>CPU clocked at 150 MHz. My HP has 512K of RAM, expandable to 4 MB. My Apple has
>64 MB of RAM, expandable to 1.5 gigabytes. My HP has a hard to learn text mode
>operating system. My Mac has an operating system that everybody has been trying
>to match for the past twelve years, without success. My HP has user RPL, Sys
>RPL and machine code programming. My Mac has FORTRAN, Yerk, C, C++, machine
>code, Java, Pascal, Object Pascal, Hypertalk, AppleScript, Frontier, Mogul,
>Oberon, Modula, PERL, AWK, Tickle, Scheme, Lisp, Forth, Mops, and several other
>languages. My HP has a serial port and an IR port. My Mac has full video and
>audio I/O including PAL, SECAM and NTSC, dual SCSI busses, 10BaseT and AAUI,
>dual RS422s, ADB, and two 17" monitors.
ObNitPick: your apple (powercomputing clone?) has an (at least) 64 bit
604 PPC chip.
ObComment: don't let the general public know where you live. that
thing is awfully nice, and we'd all come over to use/
borrow it. <grin>
charlie
Red October
>charlie
Your HP fits in your pocket, but your Apple sits on a desktop. Your
HP runs on 3 AAA batteris, your Apple plugs into the wall. Your HP
cost you about $200, your Apple, perhaps $5,000-7,000. I don't think
it's exactly fair to compare the two, do you? The HP was never
designed to be a personal computer. The HP is a calculator, and a damn
fine one at that. However, it is not a multi-thousand dollar
full-fledged personal computer, and please don't compare it to one.
That's not it's job.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Red October
chur...@cybercom.net
"Scientists tell us that the fastest animal on
earth, with a top speed of 120 feet per second
is a cow that has been dropped out of a
helicopter. How long, traveling at top speed,
will it take the cow to travel 360 feet?"
-Dave Barry Slept Here
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Red October
On Sat, 23 Mar 1996, Scott Marlowe wrote:
> Red October (chur...@cybercom.net) wrote to comp.sys.hp48 message <4ivqbb$4...@orion.cybercom.net>:
>
> : Your HP fits in your pocket, but your Apple sits on a desktop. Your
> : HP runs on 3 AAA batteris, your Apple plugs into the wall. Your HP
> : cost you about $200, your Apple, perhaps $5,000-7,000. I don't think
> : it's exactly fair to compare the two, do you? The HP was never
> : designed to be a personal computer. The HP is a calculator, and a damn
> : fine one at that. However, it is not a multi-thousand dollar
> : full-fledged personal computer, and please don't compare it to one.
> : That's not it's job.
>
> If you took the time to read what he was writing, you'd have seen that he
> was showing that it is NOT FAIR to compare the HP48 to a power mac.
>
> I.e. you agree with him.
>
>
Oops, my bad, sorry :)
-Red October
Brian Downing
>Brian Downing wrote:
>> "Calculators with typewriter-style (QWERTY) keyboards will not be
>> permitted, and calculators with infrared capabilities will be permitted
>> only if the serial port is covered with electrical tape."
>> Huh? The _SERIAL PORT_?!!! Boy, am I gonna have fun walking in to the
>> test with my GX's serial port covered. hehe
>I hate to break it to you, but the IR port is a type of serial port, if you get
>technical...
Yes, yes, yes. I know that the IR port transmits 1 bit of data at a
time. Well, they don't state _which_ serial port.
Still funny.
-bcd
--
*** Brian Downing / bdow...@somat.com | finger for
*** http://www.somat.com/~bdowning/ | PGP key
*** "It does not do to leave a live dragon out of your
*** calculations, if you live near him." --- J.R.R. Tolkien
Tom Cramer
>Right. The HP has a 4 bit CPU clocked at 2 MHz. My
>Apple has a 32 bit 604 RISC
>operating system. My Mac has an operating system that
>everybody has been trying
>to match for the past twelve years, without success. My
>HP has user RPL, Sys
There is other programming languages for the HP. Oh, by the way,
ever herd of an AMIGA? Probably not, since looking at the posting,
you don't look very hard. The Amiga had 32 bit, premtive multitasking,
and 4096 colors when it first came out. the OS was Icon, and text,
and is your choice what to use. And guess when the Amiga came out.
1995?
No before that...
1993, nope,
try 1985. Lets see, Apple 1996=Amiga 1985? Looks like someone beat it
before Mac was born.
(sorry to go off a bit, but just trying to get this poor person to
relise you can't compare apples to oranges (calculators to computers))
Tom
--
Fireball
> >> 4 sp 6 * 5 sp 7 / + 9 keystrokes
>
> even with
> algeabraic 's
>
> these things DO use proper precedence, after all
>
> >5 sp 7 / 6 + 4 *
> this is wrong... try 5 sp 7 / 4 sp 6 * + still 9
I don't know what you did but, I enter both of the above string of
commands and BOTH work. Different answers but BOTH work!
Fireball
I don't see much in the way of statistics. Specially, I would like to
know if one could do regressions on the hp48g? I have come across it
yet, and was wondering how to do it, what some of the capliblities of the
hp in this pictular function, how or in what form is it displayed? Any
light on this subject would be helpful :-]
----chris-----
cl...@axe.humboldt.edu