A369

I'm looking for HP42S calculator in Italy and possibly in Puglia
--
COMO PODRIA HACER LO SIGUIENTE EN ML ??
 1 2 3 4 -> 4 2 3 1
> COMO PODRIA HACER LO SIGUIENTE EN ML ??
>  1 2 3 4 -> 4 2 3 1
Sorry, I do not speak Spanish but this will do what you want.
It is quit a good example for ML, I think.
I reverses the stack order of any number of stack elements
( -> )
( ob1 ob2  ...   obn -> obn ... ob2 ob1)
CODE
SAVE
CD1EX
D1=C            % point D1 and D0 to the first stack level
D0=C
A=DAT0 A
?A=0 A -> EXIT  % if 0 then no objects on stack
*TOP
D1+5            % point to next stack level
A=DAT1 A
?A#0 A -> TOP   % do until top of stack
D1-5            % point to top stack level
*MAIN
A=DAT1 A        % read highest object (level n)
C=DAT0 A        % read lowest object (level 1)
DAT1=C A        % swich them
DAT0=A A
D1-5            % point to level (n-1)
D0+5            % point to level 2
AD0EX
D0=A
CD1EX
D1=C
?A<C A -> MAIN  % check if middle of stack is reached
*EXIT
LOADRPL         % return to rpl
ENDCODE
@
--
This message was written with 100% recycled electrons
Pivo
> I reverses the stack order of any number of stack elements
Wouldn't it be simpler to do:
<< ->LIST REVLIST OBJ-> DROP >>
-- 
Ralf Kleineisel  -  Regionales Rechenzentrum Erlangen
                     Kommunikationssysteme
> Wouldn't it be simpler to do:
> << ->LIST REVLIST OBJ-> DROP >>
This is very slow because all three involved
UsrRPL commands are particularly slow if the
size of stack reversion is high. It nearly never
pays to program this in ML since the following
SysRPL program with 3 commands is much shorter,
really fast and error protected, in addition.
::
 CKN  (check for sufficiently many stack objects)
 reversym (reverse stack order of meta-object)
 UNCOERCE (regain the real or zint in Level 1)   
;
The stack diagram of this program is
      ob1 ... obn %n  ->  obn  ... ob1 %n     
Briefly speaking, it reverses a user-meta-object.
- Wolfgang
PS. reversym is itself a SysRPL-program and uses
the meta-object commands pshzer and psh1&rev.
> I reverses the stack order of any number of stack elements
> Wouldn't it be simpler to do:
> << ->LIST REVLIST OBJ-> DROP >>
Yep but that wouldn't be in ASM now, would it?
--
This message was written with 100% recycled electrons
Pivo
>>Unfortunately I need the cumulative subtotals of the list, not only the
>>total.
> Wou won't get rid of me that easily
> <<  DUP SIZE { } SWAP
> 1 SWAP
> FOR I
> OVER 1
> I SUB
> 0 +
> GSLIST +
> NEXT
> SWAP GSLIST
> {1 2 3 4} -> {.1 .3 .6 1.}
> is that more like it?
> --
> Pivo
it's me again. It works great. Any idea how to output the subtotals ({1 
2 3 4} -> {1 3 6 10} -> {.1 .3 .6 1.}) at the same time?
I tried several things but didn't find a good solution.
Georg.
> it's me again. It works great. Any idea how to output the subtotals ({1
> 2 3 4} -> {1 3 6 10} -> {.1 .3 .6 1.}) at the same time?
> I tried several things but didn't find a good solution.
 of course: change the last SWAP to DUP ROT
<<  DUP SIZE { } SWAP
 1 SWAP
 FOR I
 OVER 1
 I SUB
 0 +
 GSLIST +
 NEXT
 DUP ROT GSLIST
 / >>
--
This message was written with 100% recycled electrons
Pivo
That's what I've been looking for.
Georg.
>>Unfortunately I need the cumulative subtotals of the list, not only the
>>total.
> Wou won't get rid of me that easily
> <<  DUP SIZE { } SWAP
> 1 SWAP
> FOR I
> OVER 1
> I SUB
> 0 +
> GSLIST +
> NEXT
> SWAP GSLIST
> {1 2 3 4} -> {.1 .3 .6 1.}
> is that more like it?
 
> Unfortunately I need the cumulative subtotals of the list, not only the 
> total.
Cumulative subtotals? Example please?
 
please look there:  http://www.internationalcalculator.com/
 
 
 
be sure to understand http://www.anti-matrix.net
 
> that true ? I'm eager for using the 1.19-7
 
 http://www.epita.fr:8000/~avenar_j/hp/48.html     lots of HP programs
 
This is a nearly philosophical question, yet to me there seems to 
be an obvious answer.
You are talking about _local_ variables.  They are local in the
sense that they are defined and usable in one part of the
program, and not in others.  So, in being local to one part of
the program, how does one know which part?  Well, it's the part
one know when that locality begins or ends?  What would a local
variable be without its locality?  (Global, I guess.)
--  
when
b ).
arquitecture, but
that
That's right, but there's another reason: You are allowed to use almost any
names for your local variables, unlike globals. You need a mark (<<) to
of locals you wish to create), and when the program starts. Look at the
following (easy) example, and predict what is on the stack after it has
been run (I know that you know this, Steve):
 as to 'why do we always have to start a secondary, or thirdary, or
fourthary...program delimiter set every time we declare new local 
variables...'
is....
not entirely correct.
you can easily get around this, and treat all your local variables as 
global
variables, inside one big program block, by simply....!!!...defining all 
your
variables, of the local variety at the beginning of the main program...!
you can additionally use the <- ( left pointing arrow ) as a prefix to 
create
traveling local variables...!!!...that will jump from program to program 
while
the ( directory/program group ) is running...!!!
Please describe these functions. You might get an antiderivative that
does not look like what you get with your TI89 but that does not
mean that Erable fails. And I know many functions that Erable can 
integrate but not the TI89: for example all the functions that can
only be integrated using the Risch algorithm like
(1+2x^2)*exp(x^2)
 Okay, this is exactly the kind of exchange I was hoping for, constructive,
with good ideas.  I just thought I'd mention that the day after I posted 
the
original, I got a reply from someone (who shall remain nameless) asking
where to find information on the new HP58...  (I thought mythical and
wish list would make it abundantly clear . . .)
As for the QWERTY question?  I still find that the TI-92 is entirely to
unweildly to use effectively.  I would accept the clamshell design of the
28, maybe even with a sideways QWERTY, but that is most defenitly low on my
list, as I can type 30 wpm on my 48...  (Lots of practice)
And, I just saw a Gameboy color, and I have to say, if HP pulled color off
as well as on that machine, and kept long battery life (at least a month of
heavy use), I would accept color.  But again, that's low on my list.
Support, needed, badly.  HP used to (even unofficially) have great support
(when the calc division was still just down the road from me in Corvallis,
OR), and strong (again, even unofficial) support would be great.
Finally, the feel.  I agree, the TI's feel like toys.  I remember way 

back
in high school, doing a physics experiment, and I was (of course) the only
one with an HP, everyone else had the school-provided TI-81s, and one 
person
dropped a 500g weight on his, and it snapped the case like a small twig,
destroying his calc.  Just for fun, I started dropping the same mass weight
slightly
bend the faceplace where the 'hp' logo was (upper left of the SX)  So, I
have to say that the feel should not be changed for ANYTHING!  I may want 

my
watch titanium so it's light, but I want my calc to feel solid!
-- 
programming -
learn
Just a little nit-picking of terminology: The kinds of programing
languages you aree talking about are actually called _low_ level
languages, because they are closer (lower) to the hardware.
As for what the next HP should have....I agree so far with what has been
said. Keep it about the same size, although a larger screen would be
nice, don't _make_ us use a QWERTY keyboard, etc. My personal opinion is
that if you want a QWERTY keyboard, buy a WinCE device and get the HP
emulator for it. I realize that this means that you still need to buy
the calc, but a lot of us would rather use a computer for the kind of
things that a QWRERTY keyboard would be used for.
The most important thing for me would be that they don't change the
_feel_ of the calc. I've owned several Casios and used TIs in the past,
and they all felt like toys. My 48GX, on the other hand, feels like a
tool. If they change everything else, heaven forbid, they should keep
this quality.
 
   
Please, not BASIC or Pascal.  Both languages have been deprecated strongly
by their authors, respectively Dartmouth College and Nicholas Wirth, for
encouraging poor programming styles.  Originally, BASIC and PASCAL could 
not
initialize variables in their declarations, BASIC has goto statements, both
are missing reasonable data types, orthogonal statement types and 
operators,
etc.  Yes, I am aware that _some_ later variants of these did address some
of these problems.
Nicholas Wirth deliberately had omitted data types, statement types,
available functions, variable initializations so that students would not
have more than one choice for any particular need to reduce student
confusion by having multiple choices and to ease grading student programs.
Unfortunately, most people are not willing to learn more than one language
and apply the ``if the only tool I have is a hammer, then every thing is
a nail'' technique to all their programming.
I would suggest subsetted and interpreted Java, C, or C++ as appropriate.
I could accept Perl, Tcl or Python as other reasonable models (with some
provision for initialized variables).
C?  World's highest-level assembler?  The one that
insists on passing all structs as values and all
arrays as variables?  The one last language that
still has a global goto, except for C++?  The one
where int could be 16, 32 or 64 bits long depending
on what compiler you are using and what flags are set?
That C?
Java I can relate to, though I think syntactically
it's as big a mess as C.  foo==bar is not a very
obvious way to tell if both foo and bar refer to
the same object when foo.equals(bar) is required
to tell if the values to which foo and bar refer
to are the same.  Those are implementation details,
they shouldn't be shining through.
Tcl vars _are_ always initialized.  They are initialized
to  unless you set them to something else.  And Tcl
has no provision for defining new data types, nor does
Perl.  Python does, but Python also uses indenting to
define program blocks - not a fun feature on a tiny
display.
 Sort of impossible, no ?  I thought an RPL program is just
a list of addresses of ROM-pointers which point to objects,
(which point to objects, iterate if needed) which are machine
language. Let me know if I'm wrong...
 
I would like that the new HP58 have a plug to plug it to the electric 
socket
of the house (my english is baaad, so I hope you understand me). I use my
calculator always in my house, I could plug it and don't use the bateries,
if I have to go to the university I use the bateries.
I don't know if this can be done, but I think yes.
etc,
 
If ever a new model  should be available, it have to have a ergonomic 
format
like the 28 was built !
With actual technologies it is possible to shrink 48+ facts into a smaller,
more attractive outfit.
 b) smaller, bigger screen, fewer keys, more menus, easier to access menus, 

user
defined/orderable menus...( allow the user to redefine the menu access
trees...so instead of simply creating more menus, the user could customize 

the
user interface...( this however would make it very frustrating when you lend 

it
to someone...??? ))
c) colour with 300 pixels per inch resolution / with alcohol 
batteries...!!!
d) built in spectrum analyizer.
e) tiny camera that tracks your eye movement to act as 'mouse' controller.
Of course.. I changed it as an afterthought for 1 GET in a (futile) attempt
to outwit you.
The
attempt
from vectors (use with care on 2 and 3d vectors, flag-dependable, that's
why it's slow.)
Here's a little mc: Write the best program (lowest b*t*speed) to extract
each of the following:
[ -1.2 ]
[[ -1.2]]
(2 programs, if you like)
 
If necessary, they should add a new catenate operator.
gether.  This would also allow us to define a rigorous
mathematically correct semantic for it that would apply
to numbers as well as strings and string/number combina-
building up certain types of expressions.
 
this sounds like a good idea, but if they're going to add (Eeeeeeeek!) a 
new
catenate operator/function then perhaps they should use the ones already
available to Set Theory, and along that include some new characters for
Union...and uhhh...what ever else is part of Set Theory...
along with some characters/operators/functions for Symbolic Logic, allowing 

for
fractional logical values ( Fuzzy logic ) and the Set Theory operators that 

go
along with that...
as i suggested/mentioned several months ago during the Object Oriented
Fiasco...
whenever we say add or subtract or whatever, we are very often refering to 

many
operations with a single term...maybe HP should extend the Operator base to
include all the sub-operators as User level functions...or SubUser Level
functions...???...that is, 'rare' functions...???
they could be hidden away in a buried directory somewhere, and have limited
error checking, so that if they fail, they wouldn't fail catestrophically, 

but
only crash the individual program--while allowing the user to write, 
somewhat
less friendly, but much faster code...???
hmmmmmmmm...???
 To Recap:
Is there any way to merge 2 arrays into one?
    Eg: [ [ 1 2 3]  [ 4 5 6]  [ 7 8 9] ]
merge with   [ [.2 .4 .5] ]
------------------:: o
this program is about as lagubrious (?) as yours...
 
If you have a G series:
 
Gahm's Remote 3.4 - using a setup that has worked previously. 
When I use the self-test (On-D), and press H, I see U_LB 20000. Is
that the IR loopback test? The batteries in the calc are a little
elderly - could it be that they're running out, meaning that there's not
enough power to operate the IR?
 I'm going to ask around a little on Compuserve but at this point I think I
will probably buy a Cynox card.
I'm interested in who the other source is so I can look into them as well.
I have no idea what FUD tactics are, but this is a newsgroup who's purpose
is to exchange information.  If we're afraid to do that freely, what's the
point of the newsgroup?
Barry
 
in the
----------------------------------------------------------------------
 
Hi Barry,
have you already looked at the hpcalc.org? There's a buying guide
which lists *all* addresses.
 
I just purchased a 128K card from them.  They answered all my emails
promptly and the card took 8 days to arrive from the day it was
ordered.  I got a covered card and it looks pretty good (not
homemade).  It is working perfectly so far.  As for the shrouds I
don't know.  Their literature says to ground  ones self before
touching the card, which seems to be a standard thing when touching
any kind of internal electronic components.   
   
 
[Answer in Bad English]
Very simple. When i was using it in practical classes ( Travaux pratiques
 in French, dunno the exact word in english) of chemistry, i had gloves and
 using a point-mouse device would have not helped me.
Moreover, my hp have a kind of protection case that is in place each time
i use it not to do maths (electronics, chemistry, mechanics) and a 
point-mouse
device is not really useable through transparent plastic.
A waterproof version of the new hp would be valuable.
 There are times when a GUI is better than the menu system.  Of course, I 

can't say
that is true for the 48's built-in GUI...  But for flags and such (if you 

are
unfamiliar with flag numbers) then it's semi-useful.
-- 
Only if the menus are labled with the cryptic flag numbers instead of
lables that make sense.
 
----------------------------------------------------------------------
 
Any kind of over-clocking is potentially dangerous to the hardware.
Running something at a frequency that is was not designed to be ran at
is putting your hardware at risk.
--
Aaron.
----------------------------------------------------------------------
 
No way!
RPN is great.
I get so slow when I try to use an algebraic calc now.
My first scientific calc was a TI58 but I soon learned to love the
HP's. I use my HP32 at work all the time and hate having to think
down to use algebraic calcs now.
Yuck!
 
------------------------------------------------------------
Mark Richardson
m.richardson@utas.edu.au
____________________________________________________________
 
it
almost
properly
learn
is on
hit me
Aak!  No!  Keep RPN, no matter what!  RPN is so much more efficient than
algebraic, I would not buy the new calculator if it didn't use RPN.  Now,
having algebraic entry as an option (as it is now on the HP48s) poses no
problem, and lets you work your way, and me mine.  True, their algebraic
entry could use some work, and hopefully that gets improved in the next
version, because some things just work better in algebraic mode, however 
for
95% of what I use the calculator for, RPN is far superior.
 
Eric is charging $20 for CDROM of the archive?  Hmm..  $20 is a bit much for 

media
costs.  I can purchase Mitsui Gold media for $1.50 a peice (including jewel 

case).
I think $20 covers media costs fairly well..  I would guess that there is 

something
more which he is looking to be reimbursed for.  Managing a site like that 

takes
time (I would imagine).  I don't think $20 is unfair, but it is more than 

what CD-R
media costs.  Anyway, just because Eric is selling his archive on CD doesn't 

mean
that I wouldn't send him my programs.
--
Aaron.
 
Eric is charging $15 for the archive.
 
----------------------------------------------------------------------
 
 
It is a little more complicated, but yes.
Store your equation in 'EQ' then do 30.01 MENU.
(there are lots of other ways as well.)
------- 
 
will
still
trig
that
equation
they
with
key.
the
can
----------------------------------------------------------------------
 
To recap:
HP keyboards have, in my opinion, the best *feel* of any calculator
keyboard.  The solid feel is unparalleled by anything from TI or Casio.
However, their quality has progressively gotten worse.
---------------------------:: o
as i occasionally mention to anyone that will listen...( the spiders in my 

room
are VERY attentive when i'm ranting )...is that while i own a Mac, and 
several
HP Calc's...i'm not an especially fanatical owner &/or user of either...
when a better product comes out, i won't hesitate to switch brand 
allegance...!
then also: i'm VERY disappointed and annoyed that the key labels are now 
being
printed on...
it defies the essential 'Zowieness' of a super kewl product....
the HP Calc's are becoming just another lump of plastic, ceramic & metalic
bits...
hardly different from a General Electric Toaster Oven.
 
----------------------------------------------------------------------
 
UPDATE:
It now draws ~7mA idle ~13mA editline and ~27mA when a key is pressed
 
----------------------------------------------------------------------
 
I have been amused that few people will print their battery drain
currents for OFF (Deep Sleep) Mode.  To get this value, I usually
had to turn the calc off and let it sit for - oh - 15 or 20 minutes.  
That's how long I recall it taking for the main supply battery to drain
to the level of the battery voltage, and start drawing current through
the series diodes.
Now, if I had my old lab notebook at hand, I'd tell what the readings 
were.  I seem to recall something in the 200 uA range, but that's a
vague recollection.  I have argued that for many people it is the OFF
current that most affects their battery life.  Even full-time college 
students have it in RUN mode only a small fraction of the average day.
They had to measure COMA current on the production line for each unit, 
so I had them put certain resistors across the main caps for a second 
to pull down the cap charge levels, remove the resistors, wait another 
second for stability, and then read the current drain. I don't even
pretend to recall the actual values, only that it used to lowest digit 
on my benchtop multimeter.  So quoting Coma currents is nearly useless.
Good day!
Dave.
 
----------------------------------------------------------------------
 
Do any of you guys know of why my 48sx, w/128k ram card, might be using
up the card batt in only 5 or 6 days?
 
 
----------------------------------------------------------------------
 
I'm starting to work with sysRPL and I'm making a small program. Imagine 
that
I'm already with troubles with it.
I want the program to recall the flags that control the angle mode, store 

it,
set to grad mode, make some calculations and finally restore the angle mode 

to
the original. The interest is obvious: make the program use a specific 
angle
mode in its calculations without interfering with the user settings.
Can anyone help me making such a program. My attempts have completely 
failed.
 
----------------------------------------------------------------------
 
The following will work on individual system flags:
If you have several flags you would like to modify it may be easier to
recall the flags at the beginning of the program and then restore them at
the end with the following:
 
that
it,
mode to
angle
failed.
 
----------------------------------------------------------------------
 
 
The usual way is to do RCLF (recall flags) keeping the result on the stack
or storing it in a local variable, then at the end of the program put it
last on the stack and do STOF (store flags). This restores all flags,
system and user, to their status quo ante.
This allows you to change any user or system flags during the program and
restore them at the end to what they were before entering the program.
----------------------------------------------------------------------
 
the
drinks
in
including
As a matter of fact, Melbourne is not such a good place for surf or
windsurf. We are in a kind of large bay and neither the waves nor the wind
are strong enought. North of australia is much better for this.
--
A+ Cyrille de Brebisson
Le Meilleur moment pour planter un arbre etait il y a 20 ans. Le Deuxiemme
meilleur moment est maintenant
The Best Time to plant a tree was 20 years ago. The second best moment is
now.
http://www.capway.com/brebisso
----------------------------------------------------------------------
 
  A better predictor would be the development times,
  once the project was underway.  We do know that the project
  is in fact underway, and when it began, so just fill in the
  actual project start and completion dates for all of the above,
  and then do another fit :)
  
Never forgetting, however, that the first 90% of a project takes 90%
of the time and that the last 10% of a project takes the other 90% of
the time...
 
 
 
----------------------------------------------------------------------
 
You will see that most of the people working for ACO are drinker expert...
I personally got 195 (215 in fact, but it's not politically correct) to the
 
to
in
 
----------------------------------------------------------------------
 
Well, in this case, there is nothing else left, but THE ONLY SOLUTION !
We have to contact former CIA or KGB guy in order to buy special
modified ethanol that open human memory :-) This way we will
utilize Cold War effort in order to discover HP plans :-)
Jack
----------------------------------------------------------------------
  
High tolerance to alcohol is part of security, right?
Looks like your group is celebrating. Hmm.
for
bars
 
----------------------------------------------------------------------
 
I want to put some libraries of mine in the Net. I have already a list of 

some
of the existing ID's. Is there some kind of authority to whom I can send 

the
ID's of my own libraries.
 
----------------------------------------------------------------------
 
Well I want to thank everyone who responded.  The breakin is complete.  
Anyone
who choose to do this must me aware of some things.  I talked to the hp rep 

in
Califorinia and all repairs const the same regardless of the calculators
condition.  When press he said that you are essentially paying ninety some 

on
dollars to have the calculator replaced with a new one.  Kinda lets you 
know
what that $350.00 calculator is really worth.  So if there is any serious
damage done to the calculator it is worth just sending it in.  The only 
thing
that needed to be done to mine was simple cleaning to remove the suryup(sp)
from the keyboard.  This wasn't worth the cost.
   * Remove the batteries and any exterior covers.
   * With a small screw driver slide it under the aluminium keyboard cover.
     This will deform the cover so be as careful and reduce the amount of
     damage to a minimum.
   * Once the keyboard cover is off there will be a number of recessed pins
     that snap the back cover to the front.  I drilled these out using a 
drill
     the save size as the exterior hole as a finger drill.  I DID NOT USE A
     DRILL.  The plastic is soft enough to get through.
   * Once these catches have been removed then the covers catches can be 
poped
     fairly easily.  Still take care care.  You will need these catches in 

good
     shape to reassemble the calculator.
 
 
----------------------------------------------------------------------
  
Figuring out such patterns was the basis of a game called Eleusis,
so just grab the game and play a while, and it will come to you :)
In STD display mode, so long as the HP48 can display all twelve
significant mantissa digits without reverting to scientific notation,
it does so.
For (absolute) values less than 1, this means that if the value does
not require more than twelve digits following the decimal point
(counting any necessary leading zeros, but not final trailing zeros),
then it does not need to be displayed in scientific notation.
For example, 1E-12 is displayed as .000000000001
but 1.1E-12 is displayed as 1.1E-12
So, change to scientific notation, even temporarily;
for example, the following program will temporarily
re-display the stack in scientific notation; it will
revert back to the previous display mode at the next keypress:
Be sure that the hex address is exactly as shown,
or else something untoward might happen.
 ---------------------------------------------
 
i think that it has to do with precision...
if the 48 thinks that the number can be accurately displayed with zero's 
then
it will do so in STD mode, but if the 'significant' digits leak out past the 

12
digit limit, then it switches to SCI mode...for that number...???
this might help a little...
( assign this to a user key )
i have a more complicated program which displays the 'object' on 1: anyway 

you
want on 1: for a few moments, and then reverts back to 'normal' display
parameters...
Here are 3 display routines that take an object on 1: and modifies it 
somehow
and then pastes it back on the display in a new format in such a way that 

it's
slightly illegal...!!! 
For viewing purposes only...!
Takes an Object on 1: and converts it to SCI notation while leaving 
everything
else on the stack in a pre-SCI format...!
To escape, hit any key, and that will return the object on 1: to it's 
former
state...!
<< RCLF OVER 4 SCI
1:                     @ 20 Spaces
1 22 4 PICK SIZE -
LCD215 { # 0d # 46d }
WAIT DROP STOF
-------------------------------------------: <<
fraction as an integer fraction.
Before:
3:               123
2:         Bite me
1:              2.25
After
3:               123
2:         Bite me
1:             2 1/4
<<
  <<  STD DUP IP  
- SUB +
-------------------------------------------: <<
The general solution.
Takes 2 objects;
The object to be acted on,
And the modifing function or program.
2:            object
<< RCLF 3 PICK ROT
1:                     @ 20 Spaces
1 22 4 PICK SIZE -
  IFERR 0 WAIT
  THEN
  END DROP STOF
-------------------------------------------: <<
 
I want to transfer my files between my PC and my HP 48GX.
I did it before when working under DOS, but I've changed for Linux OS.
Does anybody know existing apps for doing that purpose?
 
----------------------------------------------------------------------
 
There are several ways:
You can use seyon or minicom or simply rx/sx from the lrzsz package, if you
want to use xmodem, or get kermit (more comfortable, but much slower). The
Kermit home is at  http://www.columbia.edu/kermit/
 
----------------------------------------------------------------------
  
My first HP Calc was a 28C.  I seem to recall that there was a 
business version of that machine that actually came out a few
months earlier.  Did it have a 4-level stack or an indefinite-depth
stack?  I'll admit that I never bothered to look at it very hard,
having been a somewhat myopic Engineering student.
Dave.
----------------------------------------------------------------------
 
  Newbie!
  If you mean the HP18C (the first calc in that clamshell case),
  it has no actual stack, being exclusively algebraic,
  but it does save up to four items for display, which can be
  rotated via the up/down arrow keys.  What you might call
  level 2 can be inserted into a new expression via a LAST function,
  which is the limit of the re-usability of previous values.
  What we might now call a tagged value appeared as
  name=value in the HP18C display (as I might guess
  it still may do on the HP19B[II])
  Since the more recent HP19BII does offer stack-based calculation,
  the question might still remain, how is it implemented on that model?
  I bet two bits on the same old 4-level system.
 
----------------------------------------------------------------------
 
The 28C and later the 28S had indefinite stacks and reverse Polish.
----------------------------------------------------------------------
 
User flag 56 controls this behaviour - however it isn't in the STK screen
as we didn't feel it warranted going to multiple pages or implementing
scrolling.  So, to disable commas you can just do: 56 CF
--  
----------------------------------------------------------------------
 
well i saw the key word USER...so for anyone else who missed it, it is the 

USER
FLAG and NOT the normal system flags that control the calculator that needs 

to be
 
----------------------------------------------------------------------
 
  It's not OK if you look at your computer text file (e.g. with Notepad)
  and already see XLIB ... ... in the file, which means that you sent
  the file as text (i.e. using Kermit in ascii mode) when the necessary
  library was not attached; this causes the names of the library commands
  to be lost, and only the the library and command numbers to remain.
  True; when you read ascii text files back using Kermit, they must
  be re-compiled into a program as if they had just been typed in,
  but the internal compiler does not recognize XLIB mmm nnn
  Only if it was installed *both* when first SENDing the program
  (in ascii), and then once again when RECVing it.
  While SENDing (in ascii mode), library calls (which exist internally
  only as numeric references to library and command numbers)
  are looked up by scanning the library for the command name;
  if a name can not be found, then XLIB mmm nnn is substituted.
  While RECVing (in ascii mode), command names are looked up, and
  if not found in any library, then they are compiled as variable
  names, rather than as library function references.
  Thus, to be successful in transmitting a user program in both
  directions in *ascii* mode, all libraries which it references
  must be attached during both sending and receiving; OTOH,
  there is no problem when using *binary* transfer (except that
  you can't then read or edit the program on your computer).
  If after binary transfer you see XLIB ... ... displayed in the
  program, just attach the library(ies) you need and then proceed;
  the program will work once the library(ies) are attached.
 
----------------------------------------------------------------------
  
        I have this problem too, but if you changes your modem configuration 

it
will not work well... When I want transfer some files I reboot my
computer in MS-DOS safe mode, disconnect my mouse and run Kermit using
COM1... I know that my COM4 (in Win95b) works because I can make direct
cable connection using it...
Sorry. My english is the worst in world.
Leonardo Vasconcelos Alves.
Computer Science Student.
http://www.dcc.ufmg.br
 
----------------------------------------------------------------------
 
I just want to upgrade my 48G and I've heard about Cynox ram modules. I've 

a
question:
        - What I need to install a ram module?
ram only if you want 128k and the ram and other stuff if you want more 
(like
256 or 512k). I don't understand if the Cynox chip are enough for an
upgrade.
 Does anybody knows something?
 
----------------------------------------------------------------------
 
my homepage moved from http://www.cyberline.de/kkgbr to
http://www.uuhome.de/oklotz
 
----------------------------------------------------------------------
 
I am trying to use the HP48 Program bkup to backup my entire calculator
to my computer.  I am using Columbia's Kermit software under linux.  I
want to setup kermit so that it auto-matically recieves files from the
recieve a new file.
Anyone who has any experience with this sort of thing, please let me
know.
 
----------------------------------------------------------------------
 
I am using PBAK version 1.4 by Jim Daniels.  It runs entirely from the HP, 

and
will back up all I mean all, ports, memory and all flag status. I remember I 

had
a bit of trouble setting kermit up
try this mskermit.ini file, which kermit looks at when starting up.  I run 

rom a
floppy in a: and this sets up the server, waiting for commands issued from 

the
HP.  Version 3.14 May 1995 is backup up my HP as I type this.  If you type 

Ctrl
C to exit, just type server to return to server mode, once Kermit is running 

in
a DOS window (assuming Win95
set speed 9600
set file type binary
set parity none
set term bytesize 8
set port com2
set speed 9600
cd a:
server
 
----------------------------------------------------------------------
  
Well if you already can receive files from your HP, you can just start the
server under kermit and then send all the files you want from the HP,
without having to bother anymore about the kermit.
Try it up, and tell me if it doesn't work or doesn't work the way you want.
----------------------------------------------------------------------
 
Maybe you could tell me how to setup colubia kermit under linux, cos I 
can't
get a connection to my hp48 (So many options!).
 
----------------------------------------------------------------------
set file type bin
set file name lit
set modem hayes
set terminal apc off
set server display on
getting files from the HP:
kermit -C set line /dev/ttyS2, set speed 9600, receive, exit
sending files to the HP:
kermit -C set line /dev/ttyS2, set speed 9600, send $1, exit
kermit server mode:
kermit -C set line /dev/ttyS2, set speed 9600, server, exit
Of course, you have to change /dev/ttyS2 to match the right com-port.
The correct IOPAR on the HP is 9600,0,1,1,3,3
This should be enough to get things up and running.
 
----------------------------------------------------------------------
 
Try running the kermit software in server mode and send the files from
the HP rather than get them from the HP. This is how I used to do it. 

You
can even write a user rpl program to send the HP directories that you
want to backup with the push of one button.
 
 
 
 
----------------------------------------------------------------------
 
You may come at my web page and download my french and stupid assembly doc:
http://www.chez.com/sunhp
Julien Meyer
cbez@unisul.rct-sc.br a 216crit dans le message
----------------------------------------------------------------------
 
My guess is that the HP uses its root solver to find eigen values. If I
remember my diffeq correctly, then eigen values are the roots of the
charateristic equation to differential equation. For systems of diffeqs, I
would say that is uses its matrix inversion routine to solve for the eigen
vectors. I am not sure which numerical method its uses, but my best guess
would be Newton-Raphson or some variant.
 
----------------------------------------------------------------------
 
I send the latest version of MeMan library.
Last time I received just a few mail to give me impressions on this
prog, so I renew my request to mail or reply to this message, also to
make criticisms or to tell me the program is useless: I need to know if
what I do is useful for someone.
This version is still a beta, because some functions have been not
completed yet, but I implemented most important ones, and I think the
program is stable and quite bug free.
I also included a small documentation.
 
----------------------------------------------------------------------
 
  What it do?
  
----------------------------------------------------------------------
 
I'm making another circuit which needs to talk to the HP48, but
this time I'd like to make it through an IR link. Has anyone already
tried this? I'd like to know if it's too hard, and if you have any
circuit suggestions... Please tell me about it!
 
 
----------------------------------------------------------------------
  
This is probably a stupid question, but does anyone know where I can 
download
a copy of the users manual for the HP 20S. I checked the HP website and
couldn't find anything.
I download a copy of the HP 48G advanced users manual.
 
----------------------------------------------------------------------
 
Just recently I competed in a county wide math compition in Norther
California. After taking the test me and my friend went over to the
assembly hall where the awards would be awarded. As we were waiting we
joined up with another group of High School students from another school
after we noticed their HP's. We got to talking and discussing the
awesome features of the HP. As more people started pouring in a small
group of HP users congregated and at the peak we estimated we were
surrounded by about 20 HP48G/GX users!!! Soon we dispursed and took our
seats. As awards were handed out it seemed like a good portion of the
winners (including me 2nd Place) had HP's. Guess this just goes to show
what type of people use HP's. HP48G/GX is awesome, I sincerly hope the
next HP is better. A compiled ROM would be nice!!!
 
----------------------------------------------------------------------
 
On the other hand, when I took Stat in grad school, the prof said that we
could roll in any text, textbook or IBM 370 if we wanted, and it, by 
itself,
of
the writer.)
It's like prepping for the SAT - by learning to make full utilization of 
the
HP48 in high school you are opening yourself up to all the wonderful
intricacies of math and logic (is that philosophy or math?).  In fact, you
will find yourself better prepped to do intuitive excercises because of the
library you've built in your mind.
 
----------------------------------------------------------------------
 
I'm a little confused. Are you saying it is good to get an early start with 

a
calculator that will last you well through college and beyond or are you 
saying
that by using a high end calculator like the HP you are doing a disservice 

for
yourself? I guess I would have to take a neutral position on that sort of
argument since the HP has gotten me out of tight situations in math more 
than
once.
 
----------------------------------------------------------------------
 
I'm going to upgrade my poor HP48g, and I don't understand why I need this
two chips:
 - A 74HC00
 - A 74HC174
 
----------------------------------------------------------------------
 
The 74HC00 is a quad NAND gate IC.  The 74HC174 is a hex D Flip-flop
chip.  
chips are required if you simulate a Port 2 configuration.
The NAND chip is also required if you simulate Port 1 with a device
that requires the NOE (Output Enable, Inverse Logic) to be driven
at the right times.  
chip design permits this simplification.  I looked at the chip data 
sheet before making this design decision.  Not all memory devices 
have this option, so proceed with caution.
I usually advise against fiddling with the HP48 hardware, especially 
when the person doing the 'upgrade' doesn't understand electronic
assembly methods, or doesn't understand the circuit they are creating.
My reasoning is that mistakes happen, and if you don't know how to 
find and correct mistakes you run a greater risk of losing use of 
the machine permanently.
Best of Success!
Dave.
-----
I don't speak for HP when I post here.
Those who modify their own HP48s do so at their own risk.  Use this 
information with caution.
----------------------------------------------------------------------
 
The first one is a set on nand used as inverters, they are used to generate
the chip select signal on the port2
the second one is a set of 6 d-latches used to store the 5 bit extended
address for the port2. It's the banks switcher.
--
A+ Cyrille de Brebisson
Le Meilleur moment pour planter un arbre etait il y a 20 ans. Le Deuxiemme
meilleur moment est maintenant
The Best Time to plant a tree was 20 years ago. The second best moment is
now.
 
----------------------------------------------------------------------
 
The list price for a new calculator is $267.00, each of the 3 card packages
are list at $125.00, the links48 software and serial cable listed for 
$75.00
new.  Total list is $717.00.  Please confirm if you'd like.  I ask $300.00
for the whole package.  ($325.00, I forgot I have to ship it to you.
----------------------------------------------------------------------
 
I have a problem on: 
    Numeric diff solving with matrix.
The thing is that on the hp48 there are a numeric diff solver,
that solve numerically diff equation of the form of:
    y'= f(y,t)
And it works on matrix too, look on section 19.
The declaration in this is only f(y,t) 
Higher order of diff equations must be represented 
as matrix form to one order diff.
My problem is now:
  I want to solve a diff matrix problem of the form of:
  
  = _    _      = _
  C*y' = W(t) - G*y
            
            = =
  Were the  C G  is sq-matrix, and the rest is vectors.
! I cant jest inverse C and multiply it with the right size, cause
  the C-matrix are under defined, and by definition does not have any  
  inverse or determinant.
The equation is a equation from the circuit simulation program 'Csim'.
Were C-matrix is the representation of component that depends of j*w,
like condensators inductans and mutual things. And the G-matrix is 
the representation of resistive components, like R and S.
And the W-vektor is voltage.
A solution of a 'AC' problem looks like this:
  _   _  =     =
  y = W/(G+j*w*C)
( my quest is not about this, it's about the 
   'transient'-analyse that need diff solving )
Some examples of matrix is here:
W=[[ 0 ]
   [ 0 ]
   [ 0 ]
   [ 0 ]
   [ 0 ]
   [ -1 ]]
  
C=[[ 0 0 0 0 0 0 ]
   [ 0 0 0 0 0 0 ]
   [ 0 0 0 0 0 0 ]
   [ 0 0 0 .0000001 -.0000001 0 ]
   [ 0 0 0 -.0000001 .0000001 0 ]
   [ 0 0 0 0 0 0 ]]
  
G=[[ .1 0 0 0 0 -1 ]
   [ 0 .01 0 -.01 0 1 ]
   [ 0 0 .002 0 0 0 ]
   [ 0 -.01 0 .01 0 0 ]
   [ 0 0 5.E499 -5.E499 0 0 ]
   [ 1 -1 0 0 0 0 ]]
I have made a program were i can choose between different 
numeric diff solve algorithms. I choose between Eulers, trapezoide,
Runge-kutta and the two different methods the hp48 calculator have
inbuilt ( what ever they is based on ).
I will attach my library to this mail, And it needs Fin's HACK library,
and its a small bug in the attachment of it, type 1040 ATTACH, for
properly use.
I wonder how the built in numeric solver takes something called 
 Absolute error tolerance how in the heck can it figure out the error
of the answer?
All of them support solutions with matrix, so i thought that i have made
a good play-ground to se what algorithm is best later on. BUT it didn't
work with the Csim program.
the author of Csim claims in his documentation, the he implemented the 
transient analyze on only 15 minutes!!
So i thought i could make a better job. But i have definitely tried to
over come this problem more than 15 minutes!
And i can't do it...
The Csim program solves it's diff with the following algorithm:
 _        =       =         = _           _
 y.n+1 = (C+Delta*G)^-1 * ( C*y.n + Delta*W.n+1) 
        { with euler algorithm }
And
 _        =           =         =           =  _    
 y.n+1 = (C+(Delta/2)*G)^-1 * ((C+(Delta/2)*G)*y.n +
                   _     _
        (Delta/2)*(W.n+1+W.n)) 
                                { with trapezoideal algorithm }
Yea well i can't chow all the steps to prove it here now,
but it is in the author of Csim Per Stenbius documentation.
you find it here:
http://www.aplac.hut.fi/publications/ct-11/main.html
The question was agen:
   How can i solve this differential equation numerically,
     = _    _      = _
     C*y' = W(t) - G*y
   on the hp48, and the C matrix is under defined.
   And how can the hp48 check Absolute error tolerans that i need to
   specify in the diff solver? ( way not only the step ?) 
Any help or comments are wery welcome!!
 
----------------------------------------------------------------------
 
They cost $90 at the GA Tech Bookstore, brand new.
 
 
----------------------------------------------------------------------
 
without deviating too very much from the topicalness of the HP48...i would 

like
to propose that people start posting more 'problems' that they're having 
that
they hope to solve with the HP48.
so that the new set of problems will be for the '48, 
and not so much about the '48...???
-------------- <
First two problems:
1) 
i was asking awhile ago about a program or technique, or process or approach 

to
solving a spiral...
i got several 'starts' to this, but i never got around to finishing it...
what i would like is a program in which you could input
2: The angle of the spiral 
   0 = circle or singularity at 0 radius
1: Desired length from center
   center would always be (0,0)
So that the program would then return the (x,y) of where the spiral would 

be
after extending outward along the spiral path for, say; 22 units...???
???
2)
the other thing that i got hung up on awhile ago was a tube maker for my
Perspectivator Directory...
this always seemed like a very simple problem, and yet...
the input would be a list of [ x y z ] coordinates that would define a path 

or
string in 3D space...
and the number of sides to the final rope...???
2:  { [ x y z ] [ x y z ] [ x y z ] ... }
1:  sides
the sides would be 4 if you wanted the final rope to look like a stretched 

out
box...???...or 12, if you wanted it look more roundish...???
the output would then be;
2: All the [ x y z ] coordinates for all the polyhedra, 
   which would probably be rectangles.
1: All the rectangles, defined so that they refer to the 
   coordinates in list 2:
using this approach...
a cube would be:
2: { [ -1 1 1 ] [ 1 1 1 ] [ 1 1 -1 ] [ -1 1 -1 ]  
     [ -1 -1 1 ] [ 1 -1 1 ] [ 1 -1 -1 ] [ -1 -1 -1 ] }
1:  { { 1 2 3 4 } { 1 5 6 2 } { 2 6 7 3 } 
      { 3 7 8 4 } { 4 8 5 1 } { 5 8 7 6 } }
all the rectangle ( squares ) should also be defined so that they 
occur/appear
counter-clockwise from the center so that the hidden line removal technigue
&/or play-shadows may be applied to the final object ( rope in this case 
)...!
thanx...
what i would like to see are problems having to do with number theory, 
number
series, perspective, geometry...
or something completely different...!!!
----------- :: o
.---..-..-..-..-..-..-.  .  .-. .-.  .-..-..-..   .-.  
 `-' `-'`-'`-' `-'    `. ^ .' `--^--'`-'`-'`-'' `--^--'
   Thinking Is The Enemy; Dada Your Way To Happiness!
----------------------------------------------------------------------
 
yes, I agree that there should be an increased focus on specfic problem 
solving
for HP calculators, however, I'm an old HP-28S user recently becoming a
power-user with no intention of upgrading to a 48 series until my calc 
breaks. 
The HP-28S is a mighty powerful younger brother to the 48 series. The 
Hp-28S
was discontinued in 1993 with the introduction of the HP48G.
Loyal HP-28S power users still out there are especially welcome.
 
whether you're using a 28 or a 48 is 'almost' irrelevant' as they both ( 
naked
calcs ) speak essentially the same language.
so that problems that are intended for one, would be easily converted to 
the
other... (?) ( mostly )
and then: actually--the 28S was replaced by the 48SX, the G came much 
later,
after the GX.
And finally...have you even tried fooling around on a 48...???
i loved my 28 dearly, and could kick myself for giving away my 28C...but 
after
i got my 48, i can't stand even touching the old 28...it is just so 
amazingly
clumbersome to use...!!!
there are a couple of things that i liked about the 28, like the menu
arrangements...i still have a very hard time finding things on the 48, if i
don't already know where they are...i often resort to looking up the 
spelling
of various functions in the pocket guide after searching for them in the 
menu
trees...!!!
but all in all...the 48 has a much easier to use interface which is
considerably more valuable than the bigger screen or larger memory...!
try one...you'll ( almost certianly ) like it...!!!
 
I have a problem on: 
    Numeric diff solving with matrix.
The thing is that on the hp48 there are a numeric diff solver,
that solve numerically diff equation of the form of:
    y'= f(y,t)
And it works on matrix too, look on section 19.
The declaration in this is only f(y,t) 
Higher order of diff equations must be represented 
as matrix form to one order diff.
My problem is now:
  I want to solve a diff matrix problem of the form of:
  
  = _    _      = _
  C*y' = W(t) - G*y
            
            = =
  Were the  C G  is sq-matrix, and the rest is vectors.
! I cant jest inverse C and multiply it with the right size, cause
  the C-matrix are under defined, and by definition does not have any  
  inverse or determinant.
The equation is a equation from the circuit simulation program 'Csim'.
Were C-matrix is the representation of component that depends of j*w,
like condensators inductans and mutual things. And the G-matrix is 
the representation of resistive components, like R and S.
And the W-vektor is voltage.
A solution of a 'AC' problem looks like this:
  _   _  =     =
  y = W/(G+j*w*C)
( my quest is not about this, it's about the 
   'transient'-analyse that need diff solving )
Some examples of matrix is here:
W=[[ 0 ]
   [ 0 ]
   [ 0 ]
   [ 0 ]
   [ 0 ]
   [ -1 ]]
  
C=[[ 0 0 0 0 0 0 ]
   [ 0 0 0 0 0 0 ]
   [ 0 0 0 0 0 0 ]
   [ 0 0 0 .0000001 -.0000001 0 ]
   [ 0 0 0 -.0000001 .0000001 0 ]
   [ 0 0 0 0 0 0 ]]
  
G=[[ .1 0 0 0 0 -1 ]
   [ 0 .01 0 -.01 0 1 ]
   [ 0 0 .002 0 0 0 ]
   [ 0 -.01 0 .01 0 0 ]
   [ 0 0 5.E499 -5.E499 0 0 ]
   [ 1 -1 0 0 0 0 ]]
I have made a program were i can choose between different 
numeric diff solve algorithms. I choose between Eulers, trapezoide,
Runge-kutta and the two different methods the hp48 calculator have
inbuilt ( what ever they is based on ).
I will attach my library to this mail, And it needs Fin's HACK library,
and its a small bug in the attachment of it, type 1040 ATTACH, for
properly use.
I wonder how the built in numeric solver takes something called 
 Absolute error tolerance how in the heck can it figure out the error
of the answer?
All of them support solutions with matrix, so i thought that i have made
a good play-ground to se what algorithm is best later on. BUT it didn't
work with the Csim program.
the author of Csim claims in his documentation, the he implemented the 
transient analyze on only 15 minutes!!
So i thought i could make a better job. But i have definitely tried to
over come this problem more than 15 minutes!
And i can't do it...
The Csim program solves it's diff with the following algorithm:
 _        =       =         = _           _
 y.n+1 = (C+Delta*G)^-1 * ( C*y.n + Delta*W.n+1) 
        { with euler algorithm }
And
 _        =           =         =           =  _    
 y.n+1 = (C+(Delta/2)*G)^-1 * ((C+(Delta/2)*G)*y.n +
                   _     _
        (Delta/2)*(W.n+1+W.n)) 
                                { with trapezoideal algorithm }
Yea well i can't chow all the steps to prove it here now,
but it is in the author of Csim Per Stenbius documentation.
you find it here:
http://www.aplac.hut.fi/publications/ct-11/main.html
The question was agen:
   How can i solve this differential equation numerically,
     = _    _      = _
     C*y' = W(t) - G*y
   on the hp48, and the C matrix is under defined.
   And how can the hp48 check Absolute error tolerans that i need to
   specify in the diff solver? ( way not only the step ?) 
 
----------------------------------------------------------------------
 
http://www.dejanews.com/getdoc.xp?AN=440247078
But non of this hotshots on that news group have replied anything.
So now im posting an extra orientation about the problem, 
and hope somebody have ANY idea! 
 I have a problem on:
     Numeric diff solving with matrix.
  Look on the following thing:
a simple circuit like:
             ____ 100ohm 
   n1 *-----|____|----------* n2
      |                     |
      |                    ___  
 2V  ___            ___  100nf
      -          |
      |          |
      * gnd               *  gnd
    
Can be represented in the equation:
[[ 1/r  -1/r        1 ]    [[ v1 ]    [[  0  ] 
 [ -1/r  1/r+i*w*C  0 ]  *  [ v2 ]  =  [  0  ]
 [ -1    0          0 ]]    [ v3 ]]    [ -2A ]]
Were the extra node here is a extra stamps for the 'voltage'
 source( it's only v1 and v2, who are interesting in the solution). 
The system is G*v=J!
This system can be pointed up into two sq-matrix and one vector:
i put
G= [[  0.01 -0.001  1 ]
    [ -0.01  0.001  0 ]
    [ -1     0      0 ]
And put a new matrix for the frequency depending things:
C= [[ 0 0         0 ]
    [ 0 0.0000001 0 ] 
    [ 0 0         0 ]]
And the current vector:
J= [[  0 ]
    [  0 ]
    [ -2 ]] 
A simple solution of a AC problem is solve like this:
V = J /( G + jw*C )
BUT the problem is now how do i solve for a transient analyze problem?
The transient needs to be soled with a numeric diff solution!!
I have lurn some methods from school two solve a diff numerically. 
And the form the equation has to be in is: y'=f(x,y)  
So in my case the diff equation is:
  v*G + v'*C=J    (were v' is the dv/dt thing)
But i can not solve out this v' , the closest i can get is this:
 v'*C = J - v*G 
C does not have any inverse matrix because it's underdefinend!
What can i do???? How do i solve
this??????????????????????????????????????????????????????????????????????
 
----------------------------------------------------------------------
 
I expect no one understands the question. You have the symbol for a 2V
battery driving a passive circuit made up of a resistor and a capacitor. 
V = I*R + Q/C or, putting I = dQ/dt (where Q is the charge on the
capacitor) R*C*dQ/dt = V*C-Q.
dQ/(V*C-Q) = 1/(R*C)*dt
-ln(V*C-Q) = t/(R*C) + K
if Q = 0 when t = 0 then
K = -ln(V*C)
ln(1-Q/VC) = -t/RC
Q = VC(1-exp(-t/RC))
I = dQ/dt = V/R*exp(-t/RC)
Alternatively, if you really mean an AC voltage source V*exp(i*w*t) then
the circuit impedance is Z = R - i/(wC)
I = V*exp(i*w*t)/(R-i/wC) = V/(R^2+1/(wC)^2)^0.5*(R+i/wC)*exp(i*w*t)
What is the problem? Why do you need matrices and differential equations
for this simple circuit?
 nson
----------------------------------------------------------------------
 
If you have the value v0:=v(t) and want v1:=v(t+delta_t), solve
v1*G+(v1-v0)/delta_t*C=J(t+delta_t)
for v1 and you should be done.
This is called 'implicit euler method', the simplest possible method
you can apply. If you happen to see a method called 'explicit euler'
(which would correspond to v0*G+(v1-v0)/delta_t*C=J(t) ) - forget it.
It's useless for circuit simulation for mathematical reasons, though it
would be much faster if it would worked, but it doesn't work for 
realistic examples.
If you have further questions, mail me.
Hope that helps a bit,
Martin.
 
----------------------------------------------------------------------
 
 
Ok.. i get it.
But I was looking for a form of typing the equation like:
      ___________
     | y'=f(x,y) |
     |___________|
when i think about Euler it's : Y.n+1= Y.n + delta * f(X.n,Y.n)
AND X.n+1= X.n+delta_t    .
Now when i followed you advise i get the equation:
  V1 = ( J(t+delta_t)+ (V0/delte_t)*C) * (G + C/delta_t)^-1
delta is found here on several places, and i can absolutely not
figure out were Y.n is. So i can solve out corresponding:
       f(x,y) 
I need  V'  !! How would I else easy switch between different
numeric algorithm that i want?? 
 
----------------------------------------------------------------------
 
This is not possible, since you don't have a differential equation
system but a differential algebraic one. (This means e.g. you cannot
provide arbitrary initial values for all voltages)
That's the one i told you to forget about!
You need 
X.n+1= X.n+delta_t    .
Y.n+1= Y.n + delta * f(X.n+1,Y.n+1)
(of course, you have to solve that for programming, first.)
what else do you need? v1 is v(t+delta_t), which corresponds
to Y_n+1 and X.n+1=t_n+delta_t
As i explained above, that's not possible in general.
(Otherwise C would be regular). You cannot apply methods
designed for explizitly stated differential equations (y'=f(x,y))
for implicit given differential algebraic equations out of the box.
What algorithms do you want?
e.g. choose trapezoidal rule:
then you would have to solve
(v1-v0)/delta_t*C=0.5*[(J(t)-v0*G)+(J(t+delta_t)-v1*G)]
for v1.
If you want other algorithms, you have to adept this.
(implicit Runge Kutta methods are a bit more complicate for this)
 
----------------------------------------------------------------------
 
Is there any way to defeat the following bug:
In Java37b4 bringing a line higher than level 1 into the standard HP
editor crashes the calculator and deletes the stack.
Bringing a matrix higher than level 1 into the HP matrix editor crashes
the calculator hard and deletes all the memory.
 
----------------------------------------------------------------------
 
What is ZeldaHP?
What is GPL?
 
----------------------------------------------------------------------
 
I will try to give an answer for each questions you asked:
- ZeldaHP is (will be) one of the best and cute game designed for the HP48
- GPL means General Public License (www.gpl.com)
- On mars ? Well I'm waiting the release of: Total Recall II to get an
answer :)
 
----------------------------------------------------------------------
Mac
http://studwww.eurecom.fr/~grundsch/ramcard.html
 
----------------------------------------------------------------------
 
when everything is set up for transfer, you just have to type 'server' in
the kermit.
You can put all youroptions in a '~/.kermrc' (~ is a shortcut for your HOME
directory). You can create the file with any text-editor (ASCII only).
Mine contains the following:
set line /dev/ttyp2
set file type binary
set speed 9600
and that's all.
The 'set line /dev/ttyp2' command usually doesn't work, because the line is
set by x48 when it starts, and it depends on the number of pseudo-terminals
(xterm or other) that you have on your screen. As you surely know, the line
that x48 uses is written as 'wire: /dev/ttyp*' just under the screen.
 
----------------------------------------------------------------------
Your calculator is probably set in 'a complex mode', the ti-89 has the same
behaviour.
JP BINI
 
----------------------------------------------------------------------
 
As I recall most calculators produce an error when you try to calculate a
fractional exponent of a negative number because it calculates the fraction
first, so what you are trying to do is calculate -8 to the .666666666667 
power.
Casio and HP have complex number functions that let them calculate the
primary answer for these problems, which is why you get the (actually 

one of
the) correct response from these calcs!
If you want the cube root of -8 squared, you should try (-8^2)^(1/3) or
something of that nature :)
 is
----------------------------------------------------------------------
 
function:
'FRACPWR(X,N,D)=XROOT(D,X^N)' DEF
So you can use FRACPWR(-8,2,3) in place of (-8)^(2/3).
function.  That way you can graph 'FRACPWR(X,2,3)', and then look at its
derivative with F'.
'derXROOT(r,x,dr,dx)=1/r*XROOT(r/(1-r),x)*dx' DEF
and store somewhere accessible, like HOME.
Deciphering is left as an exercise for the reader. :)  (hint: what is
1/((1/r)-1)?)
 
----------------------------------------------------------------------
 
here is a old story I was 17 when I bought my CPC, 1989. I bought my first
'48 hm.. three years later ? There was no serial connection with the CPC,
just a parallel port with missing 8th bit (THANX, Amstrad) There was a 
small
HOWTO_get_data_from_CPC_to_PC in germans most popular magazin c't
(http://www.heise.de/ct/). Remember: the CPC used 3 discs( THANKS again,
Amstrad to be three weeks faster than IBM in setting standards). A simple
serial interface by sending through data pin 1 and receiving with the busy
signal ;)
i tried to connect my HP that way. I remember a long night playing around 

with
the wait&send loop on the CPC and on the HP, until i got fitting values.
Ok, here is the source for the '48. Just wait, read the right pin, write 
into
a string ( DIR gave me the amount of blocks to send). And so on.
What speed did i reach ?
( Recv19k2 )
::
 CK1NOLASTWD
 ::
  CK&DISPATCH0
  # 1
  COERCE
  # B
 ;
 BINT80h
 OVER#<
 ?SKIP
 ::
  BINT80h
  #*
 ;
 DUP
 #+
 NULL$
 SWAP
 EXPAND
 CODE
        GOSBVL   =SAVPTR
        P=       0
        GOSBVL   =GetStrLenStk
        C=C-1    A
        D=C      A
        D0=(5)   =IOC
        LC(1)    8
        DAT0=C   P
LoopIt  LC(3)      #1FF
        OUT=C
        GOSBVL   =CINRTN
        ?CBIT=1  15
        GOYES    Exit
        D0=(5)   =RCS
        C=DAT0   1
        ?CBIT=0  0
        GOYES    LoopIt
        D0=(5)   =RBR
        C=DAT0   B
        DAT1=C   B
        D1=D1+   2
        D=D-1    A
        GONC     LoopIt
Exit    GOVLNG       =GETPTRLOOP
ENDCODE
;
 
----------------------------------------------------------------------
 
While working with exponents in my Algerbra 2 (Alg 3,4) I performed the
following operation
(-8)^(2/3) This problem yielded an imaginary number  (-2,3.46410161514)
when the answer should have been 4. (Did it the long way (-8)^2   =
64^(1/3)=4)
---------------------------------:: o
This sort of thing is outside of the way my brain normally works...
but...
i was fooling around this a little, and noticed that the Absolute ABS of 
( -2, 3.46410151514 ) is 4.
which means that it's 4 units from (0, 0) as the parakeet flys.
so that i would guess, that, the 48 provides you with the really real 
solution
in 2 dimensions, to account for the negativity factor, but 4 is the correct
answer for 1 dimension ( 1 axes ) (?). 
???
 ----------------------------------------------------
 
| While working with exponents in my Algebra 2 (Alg 3,4)
| I performed the following operation: (-8)^(2/3)
| This problem yielded an imaginary number  (-2,3.46410161514)
  You could just do ABS to get a real final result, *if*
  the result should be positive, but some real results
  should in fact be negative, e.g. -32 '3/5' ^ should be -8
  But as you point out, if you are more interested in the magnitude,
  and don't care much about the sign, you can simply do ^ ABS
  Hmm..  Is there any definite relationship between the ARG of the
  complex result and what the final sign of a real result should be?
  Y'all try this as a MC or something; my brain's too fried just now.
 
First thing you'll notice in a Malaysian airport are the
posters telling you
Don't take Dada ! The penalty for Dada is death
or some such (and they are not joking) ...
 
----------------------------------------------------------------------
 
  
Here is a simplistic program for obtaining a real answer
from the ^ function, when the power is a fraction which
is the quotient of small integers:
But wouldn't you know it, if the power (level 1 argument)
is negative, the above program fails, because, e.g.
If the power is a negative fraction, it's even worse:
  (in this case, we *still* have a fractional power!)
Oh, well; please update your program to fix the above:
    1 == -1 IFT ROT ROT DUP TYPE { XPOW } { ^ } IFTE
Please make sure that the program name within the program itself
agrees with the name under which this program is stored
(i.e. the program may recursively call itself).
Now it works better, e.g.
Well, it sure is a good thing that this didn't get into
a new ROM version before we fixed it!    :)
 
Why doesn't '(-8)^(2/3)' give me the right answer (4) ?
The reason is that the function ^ accepts and returns complex values
whenever appropriate, and also always returns the principal answer
in the complex plane, which may not be real if the argument is negative,
even if there exist other real-valued possible results.
TinyWanda observed that the combination ^ ABS always produces the
correct magnitude of any possible real result (give or take a
little potential round-off error), but in case we would also
like to get the correct sign attached, we would have to do
just a bit more processing in general.
Here are some slightly revised UserRPL programs which
handle this sort of general real power of real question,
even with fractional powers,
including the sticky issue of getting the right sign.
Note that the power argument (level 1) may be either
to convert it to a fraction, which may be necessary
to determine the correct sign of the answer.
Note that (-8) ^ (1/3) now gives the same result as XROOT.
The version above is generally satisfactory for fractional
powers which do not have a large numerator; otherwise an
intermediate result might overflow, giving an invalid result.
The next version is overflow-resistant, but there is
sometimes a slight roundoff error from using ^ ABS:
Neither of the above programs ensures that its arguments are
actually real; other arguments (e.g. complex) can sneak by,
possibly producing meaningless answers.  You can prevent
It would perhaps be nicer to have a SysRPL function
to handle real^fraction; the only problem here
is defining exactly what fraction means, but any SysRPL
mini-challenge fans are welcome to have a go at it.
Some other test cases:
Another frequently asked question is:
Why does '-8^(2/3)' produce -4 instead of 4?
The answer to this (slightly different) question is that
the ^ operator has higher precedence than -, which means
that this expression actually represents '-(8^(2/3))'
If you want '(-8)^(2/3)' then you had better surround
the intended argument -8 with parentheses before doing ^
(and similarly with any symbolic expressions).
 
----------------------------------------------------------------------
 
After problem 2 (25 jan), here's:
Problem 4:
                                                           3000
     To 13 significant digits, what is the coefficient of X     in the
     expansion of:
             2000  2     1000  4  3  2    500
        (x+1)    (x +x+1)    (x +x +x +x+1)
                                   500
This is equivalent to calculating P,
with P:
       4  2        2  4   3   2
(x + 1) (x + x + 1) (x + x + x + x + 1)
We'll need a program to multiply two polynomials, here it is:
Again, I will use the fact that:
. dot products are inherently more accurate than repeated addition,
  since a dot product is executed in long precision;
. matrix multiply (*) is much faster than DOT
@--------------------------------------------------
@ PMUL
@
@ multiply two polynomials
@
@ In : Pn Pm
@ Out: Pnm
@
@ A polynomial is given by a vector of its
@ coefficients, with the coefficient of the highest
@ order in first position
@
@ bytes: 410.5
@ check: #F8824h
@--------------------------------------------------
<<
  OVER SIZE EVAL         @ get sizes
  OVER SIZE EVAL
  IF DUP2 < THEN         @ order
    SWAP ROT 4 ROLLD
  END
  ROT ROT
  DUP2 + 1 -               @ get N+M-1
  OVER 1 -                     @ M-1
  <<
    IF M1 THEN
      1 M1
      FOR I
        PM M I - 1 + M SUB
        PN 1 I SUB * 1 GET
      NEXT
      1 NM M1 DUP + -
      FOR I
        PM
        PN I DUP M1 + SUB
        * 1 GET
      NEXT
      1 M1
      FOR I
        PM 1 M I - SUB
        PN N M1 - I + N SUB
        * 1 GET
      NEXT
    ELSE
      PN PM 1 GET *
    END
In our case, the polynomials are symmetric, and the program can
be simplified (and speeded up):
@--------------------------------------------------
@ SPML
@
@ Multiply symmetric polynomials
@ In : SPn SPm
@ Out: SPnm
@
@ bytes: 364.0
@ check: #7F37h
@--------------------------------------------------
<<
  OVER SIZE EVAL         @ get sizes
  OVER SIZE EVAL
  IF DUP2 < THEN         @ order
    SWAP ROT 4 ROLLD
  END
  DUP2 + 1 -               @ get N+M-1
  OVER 1 -                     @ M-1
  OVER 2 / CEIL      @ (N+M)/2
  <<
    IF M1 THEN
      1 M1
      FOR I
        PM M I - 1 + M SUB
        PN 1 I SUB * 1 GET
      NEXT
      1 NM2 M1 -
      FOR I
        PM
        PN I DUP M1 + SUB
        * 1 GET
      NEXT
      IF NM 2 MOD THEN TAIL END
    ELSE
      PN PM 1 GET *
    END
Using
three times with
P1 = [ 1  7  24  54  90  119  130  119  90  54  24  7  1 ]
(which is P, calculated with PSML)
we arrive at P125 in about a half hour.
To calculate P250, I freed up my memory and removed the 'make full P' part,
as it is not necessary.  I divided P125 by 10^250 to avoid overflow.  The
calculation took 62 minutes.
To get the coefficient of X3000, it is sufficient to take the DOT product
of the resulting 1501-element vector with itself, multiply by two
and subtracting X[1501]^2 - as can easily be seen if you multiply a 
symmetric
polynomial of degree 3, for instance, with itself and want to determine the
coefficient of the result's middle term.
The result (after multiplying by 1e1000 to account for the previous 
scaling):
    3.97394226543e1426
Obviously not correct to within 13 digits, but that would be kind of hard 

if
you only work with 12.
TI92 answer : 3.9739422655849e1426
PS. The PMUL routine as presented here, while written in plain RPL, will 
outperform any other similar routine I have been able to test (Eric your 
server is down!) for sufficiently large polynomials.  Of course, it accepts 

only arrays, no lists/symbolics..
    programs compared:
  PMUL by WH (RPL) - Werner Huysegoms  PMUL by MH (SYS) - Mika Heiskanen,
from arit.lib and polylib.lib ( with flag 5 clear)  PMUL by WHS (RPL) - 
Wayne
H. Scott, from poly.zip  PMUL by WH8 (SYS) - Wayne H. Scott, from poly8.lib 

PMULT by DNF (RPL) - Dang Ngoc Francois
    P1 = { 1 7 24 54 90 119 130 119 90 54 24 7 1 }
    P2 = P1*P1, degree 24
    P5 = P2*P2*P1, of degree 60
    timings in secs
                  P1        P2      P5
        WH        5.72    12.75       39.34
        MH        3.21    12.04       88.85
        WHS       8.09   31.82     200.10
        WH8       3.42   13.47     121.00
        DNF      21.80  105.00     853.00
PPS. You don't think I'm crazy enough to sit and wait x hours for some 
program
     to finish?  Emu48 is 10.4 times faster than the real thing on my PC, 

but
     since other stuff runs at the same time, the timings may be off by a 

few
     percent..
BTW I did problem 1 as well, and got cond-2 for a 256x256 Hilbert matrix
to 11 sig digits in about 1 (ONE) hour.  The result posted for the 512x512
matrix as posted on the TI site is wrong, it should be: 1.18114088216e781
obtained in a bit more than 4 hours.
 
----------------------------------------------------------------------
 
 
  Why doesn't anybody tell you , what processor really IS inside the
HP48G (and also, in fact, in the 48S(X) and the 28C/S and some older
models, too)?
It's the Saturn architecture from HP, first time used inside the HP-71.
----------------------------------------------------------------------
 
It's the TI-89 that uses the M68K.
 
----------------------------------------------------------------------
 
Nope.  The HP48G/GX uses a Saturn cpu that was designed by HP.
The TI-89 and the TI-92 do use the Motorola 68000 cpu.  That cpu is no
longer used in the Mac.  I think the last Macs with Motorola cpu's used the
68040 or maybe 68050.  But they stopped making Macs with Motorola cpu's a
few years ago, I think.
The very first Macs used the 68000.  That was around 1982 I think.  They
moved in a year or 2 to the faster cpu's.
However, the 68000 is a good cpu and gives a lot of power for a calculator.
That should make the TI's much faster than the HP, but since the 
programming
isn't as good, some features are faster and some are slower.  There's
general agreement on the HP48 newgroup, I think, that the HP48 is a little
faster for the things people do most.  There's also general agreement on 
the
TI newsgroup that the TI is faster than greased lightening and does
everything many times faster than the HP.  :)
That's the difference in asking the TI students, and asking the HP
programmers and engineers.  HP people are much more subtle in our
comparisons and inevitable conclusions.
 
 
----------------------------------------------------------------------
 
   ...There's also general agreement on the
I have both the TI-89 and the HP48GX. For certain symbolic computations, 
the
TI is superior. However, for almost all numeric computations and matrices,
the HP48, (despite being almost a decade older than the TI) is clearly
superior. I find myself in a strange never-never land of using a TI to help
me learn how to do certain symbolic computations by hand, yet for 
everything
else I use my HP48. Numeric RPN computations on the HP48 are just as fast 

as
on the TI-89. Hardly a day goes by that I don't deeply wish for the new HP 

to
come out, that will just blow all the TIs away... - J.C. Randerson
 
The PowerPC series are a Motorola CPU (including the fantastic G3).
Also, the first mac was released in 1984. Amiga and Atari released a
computer based on the 68000 cpu in 1986 and 1987.
Also, the Mac classic, released in the early 1990 used a 68000 CPU..
You also find a 68000 in every Playstation ! (to control the joystick).
It's an excellent CPU, fast, lot of registers (32 bits). It's a shame to 
see
how slow is the TI89/92 with such a good cpu... What a waste of @#$@#
Jean-Yves
programming
the
 
----------------------------------------------------------------------
 
No, an IBM CPU: IBM PowerPC
I have my first PowerMac home, and if you look at it, there's a blue
Want a digital pic of it? It's quite beautiful ;))
 
1290 A24B 46C4 DE4C 13EE  D881 9B24 3F6C AED7 F114
----------------------------------------------------------------------
 
You also find a 68000 in every Playstation !
(to control the joystick).
-------------------------------------:: o
OUCH!
 
----------------------------------------------------------------------
 
Hello Jean-Yves !
The worst thing about your messages is that they all can be
construed into hints about the next calc's design :-)
 
----------------------------------------------------------------------
them :)
 
----------------------------------------------------------------------
 
I have a Cassiopeia E-10 palm pc running Windows CE that I would like to
use to backup my HP.  I was wondering if anyone could give me
information on how to use the IR ports to communicate.  I have used the
HP IR port to connect to a desktop before but I can't seem to get this
to work.  Any help would be appreciated.
 
----------------------------------------------------------------------
 
Since a few days I'm using an HP48 GX. I'm heavily using the formula 
editor.
My formulas do have lower and upper case letters as variables (e.g.   J, 
j).
The problem that I'm now facing is that I can type lower case letters in 
the
editor. However, if I later switch to the solver the only letters that are
shown are just upper case letters for the variables (J,J), which might lead
to confusion or makes it impossible to identfy which variable is meant by
which letter.
Is there any advise that somebody can give me. Is this a known problem or 

is
it something that I need to do to change this?
 
----------------------------------------------------------------------
 
I'm using my husbands email account.
thx again
Renata
editor.
j).
the
is
 
----------------------------------------------------------------------
 
METHOD #1:
On the G/GX, press right-shift down-arrow.
On the S/SX, press left-shift down-arrow.
This displays the variables and their contents in menu order,
with upper and lower case clearly visible, and you can thus
discern which menu key stands for which variable.
METHOD #2:
Specify the desired order of your variables.  For example,
if you have 'J+j=5' in EQ (which would put J and j in the
menu in that order) but you prefer to have j before J in
the menu, then store { 'J+j=5' { j J } } into EQ.  The list
specifies the order.  You can customize the menu in other
nifty ways too; check your manual.
 
----------------------------------------------------------------------
Pat 
editor.
j).
the
is
HPSauce, a stack replacement program, is well suited for this.  It has four
different versions, according to what you need and how much space you have
available.  I know for sure the biggest version (9.7K) will change the menu
labels so that lower case letters are differentiated from upper case.  The
program documentation will tell you more.  More stack replacement programs
at http://www.hpcalc.org/utils/stack/default.html
 
----------------------------------------------------------------------
 
the menu labels can only display uppercase letters! So the only solution
is to only use uppercase letters or only to use lowercase letters which
have no uppercase counterparts.
BTW in programming the HP48 lowercase letters are by convention reserved
for local variables (formulas are sort of programming aren't they :-))
 
Unless you use a good stack-replacement :-)
There are several of these, you'd best check out the FAQ
 
----------------------------------------------------------------------
 
Don't see it in here but I ran across this in my wanderings.  Hope it
helps.
http://www.hp.com/calculators/service/y2k.html
Mac
----------------------------------------------------------------------
 
Remember, that usually, when you don't have a problem, you simply don't
post...
Therefore, the picture you can get here is quite innacurate on how exactly
is the HP48...
Who will post:
===
Subject: Help! my HP48 is working properly...
I've bought an HP48 9 years ago, and unfortunately nothing bad happened...
I've tried not to use it properly and everytime it has displayed the 
correct
error message...
Worst of all, last time I wanted to do a matrix operation and I got the
expected answer..
Please help me, is there anything I can do to change the behavior ?
Before I used to have a TI calculator and I've never been in such 
situation.
our
pain..
Jean-Yves
 
----------------------------------------------------------------------
 
That is true, however, the HP48 does have a bad screen.  I have seen
soooooooooo many reports of the blue blur, missing columns...... being 
reported
on the HP48.  I have yet to hear any about the TI's, Palm's or any minature
device that uses an LCD.
----------------------------------------------------------------------
===
Subject: Re: New HP should be built stronger
happened...
correct
situation.
our
----------------------------------------------------------------------
 
I would like to know if there is a really good replacement for the
Matrix writer in the HP48. I am looking for something that is similar to
the HP48 Matrix writer but can handle symbolics.
 
----------------------------------------------------------------------
 
Not really. The best I've seen is AI048 and I'd rather write matrices by 
hand.
 
----------------------------------------------------------------------
I like AOI48. It's available at the usual place.
Dave Zelkowski
 
----------------------------------------------------------------------
 
That was close... ;-)
Actually, it's AIO48
 
----------------------------------------------------------------------
 
Could somebody give my a little push in the right direction and help me
out with this. I have translated the ? from Swedish to English and used
Clinton and Lewinsky instead of the Greek letters.  If there is a good
harted Swede out there you can read the jpg.  This is standard
engineering book kosan---- ãThe cowä
P=(x, y) and Pâ=(xâ,yâ) have the same distance r from 
origo O in the
direction angels as ä Clintonä and ä 
Clinton + Lewinsky ä (m a p
positive
äLewinskyä around O.
Express x,y,xâ ond yâ in r, äLewinskyä 

and ä Clintonä and extrapolate
them step by step using there similarities in a coordinate system with
there development around origo.
  xâ=x cos äLewinskyä öy sin 
äLewinskyä
{
  yâ=x sin äLewinskyä +y cos 
äLewinsky
----
 
 
----------------------------------------------------------------------
 
I've never traded in my old HP-28S for the new 48's that came out in 1993. 

I'm
too comfortable with it to ever trade up until it breaks, however, HP has 

not
and will not conduct any tests on the 28S to determine if it's Y2K 
complient.
Also there's nothing on HP's web site covering the 28. All support 
terminates 5
years after a calc is discontinued. If anyone could help me out that would 

be
great!!!
I plan on programming a clock program for it to determine if it's capable 

of
rolling over past 2000.
----------------------------------------------------------------------
 
I think the 28S *should* be Y2K compliant, since I believe the real-time
clock is just an elapsed time counter, keeping track of the amount of time
since the batteries were put in or the last system halt. Any program which
uses the real-time clock to calculate the date and manipulates dates in 4
digit format should work fine. I haven't tested this, but in theory it
should work.
of
----------------------------------------------------------------------
 
The HP28S built-in clock-calendar has a 100-year range,
either from 1989-2088 or from 1991-2090, depending on
which ROM version you have (rev J and all G-series
had their 100-year window advanced by two years).
Any two-digit year displays or inputs automatically
correspond to whichever calendar year lies within
the given range; in fact, at least in the G-series,
with the input forms, even when you input a 4-digit year
(to set an alarm or a date) only the last two digits are used,
and those digits are then made to correspond with a year
within this range, causing confusion to people who thought
that they could set the calendar to their birth date, for example,
but then found that it seemed to come out to the wrong day of the week
(because the calc took the year to be in the next century instead).
As to date arithmetic (DATE+ and DDAYS), the range of
input/output dates remains the same: 1582/10/15 thru 9999/12/31
(so any HP48 buried in a time capsule may have a Y10K problem :)
I am a little shocked to see how many very late-model products
and systems do not seem to have thought of the world lasting
beyond 1999.  It's also amusing to read web sites of companies
having to declare their systems non-compliant for such trivial
matters as that they might print the wrong date on a printout
cover sheet, or in one case that the battery replacement date
for a UPS (uninterruptible power supply) might appear by mistake
in red on a screen instead of blue, so you'll need to download
a free software upgrade to fix this tragic error.  The whole affair
seems to be a very great boon to the legal industry, however, and to
many other folks who are making a fine career opportunity out of it.
Microsoft seems to be so paranoid that they have declared
MS Access 2.0 non-compliant just because if you enter
a year as only two digits, e.g. 1/1/00, then the input
parser will take it to be 1900 (you can always type 1/1/2000,
however), whereas in fact their date/time data type can actually
store any date from year 100 to year 9999; well, don't miss
a good sales opportunity to scare people into buying
all new software every two years, I guess.
 
----------------------------------------------------------------------
 
[possibly mistaken stuff about HP28S deleted]
Oops!  The fog must have rolled in, since I didn't notice
that it said HP28S, rather than HP48S.
I can't even cancel the answer (either Netscape or the
news server doesn't seem to believe that I posted it,
which may be a result of my different reply to address).
Anyway, it would be most surprising if the HP28S didn't
have a very similar functionality, except for the
exact range of dates, but as you said, you can find
out quite readily just by the experiment of trying to
set the current date; sooner or later, you will come
to a year which you can't set as the current date,
and that's what will be the end of the calendar span.
If it doesn't go well past 2000,
I'll stop buying HP products entirely!
 
----------------------------------------------------------------------
 
All HP enthusiasts are eagerly awaiting the arrival of the new HP machine.
There is no official information, only rumours. There was a rumour coming 

from
someone within HP that the next HP would have a 190 MHz processor, who 
thought
it would be a great joke.  As improbable as it sounded, I wanted to 
believe
it, and bit hard on that baited hook. - I learned my lesson on that one! I
believe nothing on a new HP machine it until I pay for it, put the 
batteries
in, and fire it up! - J.C. Randerson
   Search, Read, Discuss, or Start Your Own    
----------------------------------------------------------------------
 
yea right..TI surpassing the hp48..not going to happen...consider that
the hp48 is reinvented each time you download a new program for it or
operating environment like java and there are many very useful hp48 tools
for many real life uses in industry....when was the last time you saw a
serious engineering major 'go wait a minute let me push in a whole bunch
of numbers into my nonflexible TI '  while the undergrad just uses his
hp48 and then gets a coke and a candy bar for himself while the TI crowd
is still slaving away at the abacus...
the only upgrade i would like is increased resolution in the
display...i've tried to think how to do it better electronically but
there is nothing that i cant program what i don't like....
----------------------------------------------------------------------
 
While I agree the HP is the best thing out there, that doesn't mean it
couldn't stand some improvement.  A faster processor is easy enough to do
these days (cpu speeds go up exponentially, the hp's current one is several
orders of magnitude less than it could be today), and more memory could be
had pretty cheaply as well.  Built in rom support for erable level symbolic
calculus and alg48 level algebra, as well as a symbolic stack could do
wonders for those who don't know how to upload to the calculator, and would
raise the bar for addons, making new programs extend the operations even
farther.  A higher resolution screen would ease reading and useage in
general, and many of the inputs (esp matrix and equation) could be
streamlined and sped up significantly.  Right now when I'm working
feverishly on the calculator, I sometimes have to wait for the calculator.
The calculator should be waiting on me. (obviously there are
exceptions...mathematica even takes a while to integrate sometimes)
 
----------------------------------------------------------------------
 
While I agree the HP is the best thing out there, that doesn't mean it
couldn't stand some improvement.  A faster processor is easy enough to do
these days (cpu speeds go up exponentially, the hp's current one is several
orders of magnitude less than it could be today), and more memory could be
had pretty cheaply as well.  Built in rom support for erable level symbolic
calculus and alg48 level algebra, as well as a symbolic stack could do
wonders for those who don't know how to upload to the calculator, and would
raise the bar for addons, making new programs extend the operations even
farther.  A higher resolution screen would ease reading and useage in
general, and many of the inputs (esp matrix and equation) could be
streamlined and sped up significantly.  Right now when I'm working
feverishly on the calculator, I sometimes have to wait for the calculator.
The calculator should be waiting on me. (obviously there are
exceptions...mathematica even takes a while to integrate sometimes)
Ken
----------------------------------------------------------------------
 
While I agree the HP is the best thing out there, that doesn't mean it
couldn't stand some improvement.  A faster processor is easy enough to do
these days (cpu speeds go up exponentially, the hp's current one is several
orders of magnitude less than it could be today), and more memory could be
had pretty cheaply as well.  Built in rom support for erable level symbolic
calculus and alg48 level algebra, as well as a symbolic stack could do
wonders for those who don't know how to upload to the calculator, and would
raise the bar for addons, making new programs extend the operations even
farther.  A higher resolution screen would ease reading and useage in
general, and many of the inputs (esp matrix and equation) could be
streamlined and sped up significantly.  Right now when I'm working
feverishly on the calculator, I sometimes have to wait for the calculator.
The calculator should be waiting on me. (obviously there are
exceptions...mathematica even takes a while to integrate sometimes)
Ken
------------------------ 
It had been a long time since I did anything in plain vanilla
uRPL :-)
Someone recently asked for quaternion-stuff, here's a first shot...
You'll find a directory attached below, if it evolves into something
of more interest I might bother to translate it in sRPL and
LIBrarize it. But it's nothing much yet...
USE :
The quaternions are meant to be represented by lists of 4 elements,
as to allow for some symbolic manipulation in some cases.
Thus { a b c d } is simply 'a + bI + cJ + dK'.
The operators QUPLUS, QUMINUS, QUMULT, QUDIV  work on reals
and quaternions as defined above, further argument checks are
not made, but there is no risk involved if you feed the rubbish,
you'll just get more of it, this is uRPL (what ease and joy to
program :-) .
The DIR also includes
* QUINV  which as you may guess finds the inverse of a quaternion
* a couple of other simplistic things
* a bonus QUEXP which returns (not symbolically) e^{a b c d}
many improvements can be made, functions added, etc...
I'm afraid this is for the G-series, I forgot about the
S-series inability to do LIST-processing... I'll do better
upon request... Not all functions need modification...
MC : uRPL proggie to calculate LN(q) , COS(q) ... you name it :-)
 
----------------------------------------------------------------------
 
The other night, after dropping my HP while working on some homework, I
realized just how vulnerable the LCD of my HP48 was.  I took an old CD
Jewel Case, and cut out a small piece of plastic to cover the LCD, and
secured it with an elastic band.  The elastic looked kinda cheesy, but
at least my LCD was safe from scratches getting hit. 
I was just curious if any of you out there had your own ways of
protecting your LCD's, or maybe some suggestions for me on how to better
attach a piece of plastic to my HP to protect my LCD.  
 
----------------------------------------------------------------------
  
A couple years ago I cut the back of a jewel case from a CD to 3 inches by
4.5 inches with a hacksaw.  I taped the alpha key layout (including Greek
letters), photocopied from the HP48 Handbook, to the plastic and placed it
in the top of the pocket in the HP48 soft case.  I haven't had any dents in
the metal or broken keys since, and my screen has never broken, so it works
well.  Plus, I never have to look up an alpha key combination!
I think the ideal case for a calculator is a soft case (like the HP48's)
with hard reinforcements (like my plastic).  This way there are no 
scratches
(like a hard case gets) while still having protection for the keyboard and
the LCD.
Ideally I'd try to find a larger piece of plastic to protect the entire
front of the calculator, but this was all I had on hand at the time, and it
seems to work well enough.
 
----------------------------------------------------------------------
 
I've taken a piece of Masonite (the funny pressboard stuff that pegboard is
made of) and cut it to size and slipped it into the pocket inside the HP's
case. This protects not only the LCD but the keys as well. Stiff plastic
should work just as well.
Now I wonder if anyone has any ideas for cleaning a sticky key? I think I
got soda or something on it.
 
----------------------------------------------------------------------
 
The Grapevine (publ.) HP48G/GX Pocket guide fits inside the
standard original case, and affords some protection as well
(if placed in front of the screen side); its front cover
shows the complete alpha keyboard, with all alpha shift symbols.
Donnelly's HP48 Pocket Book, on the other hand, is wider than
the above, and would be even more protective -- if only it
fit inside the case, which it doesn't quite do (so I put
mine on the outside, held by unattractive rubber bands);
Donnelly's front cover is just a photo of the calc.
Since this question is about protecting the calc,
rather than the content of the books, here are just
the ISBN's of these fine companion publications:
0-931011-45-0  HP48G/GX Pocket Guide (Grapevine)
1-879828-05-7  HP48 Pocket Book (Donnelly, G-series)
Perhaps you can, after all, tell a book by its cover :)
-----------------------------------------------------------
 
That is a great idea.  I thought of that one too.  But the edges may look
like crap from cutting it.  Of course you could file those edges also to
make it look good.  What you could do is to make the plastic piece the 
exact
dimensions of the screen
cut-out.  Seems as though something belongs there anyway.  Another thing 
you
could think of is the way the plastic slip covers the names and numbers on 

a
cordless telephone.  You could just get a thin and clear piece of  hard
plastic ( but bend able ) to put between the gap of the screen and the 
cover
plate.  Like I said, seems as though something belongs there anyway.  Maybe
HP was in such a big hurry to get it out there that they didn't care about
that anyway; they left it to public engineers like ourselves.  Please let 

me
know if you use my idea so that I will know that it works.
My e-mail address is jlaugher@gulf.net
P.S. The good thing about the small slip of plastic is that you won't have
to tear up your CD cases to have a cover; you could just bend the slip of
plastic very easily and replace it fast. :-)  Have an HP48 of a day!!
 
 
 
----------------------------------------------------------------------
Thomas
----------------------------------------------------------------------
 
The KML file is wrong. There are small working frontends on Casey 
Pattersons
homepage at http://www.gulftel.com/~pattersc/win48/. I hope the URL is
correct, but there must be a link from Sebastien Carliers homepage too.
 
[...]
 
----------------------------------------------------------------------
 
 
I just gave it a try: very nice! However there is one
drawback: the mapping of the 0 (i.e. zero) does not
seem to work on my Toshiba Libretto keyboard. Has 
anyone a clue what scan code the 0 has on the Libretto
(it sure isn't 48 or 96 as used in the KML script)?
 
----------------------------------------------------------------------
 
----------------------------------------------------------------------
information on that.
Another thing: how can I get the Jim Donnelly book for sysRPL in Europe. If
possible near my country - Portugal. I imagine that ordering from the USA 

is
very timeconsuming.
 
----------------------------------------------------------------------
 
Download the new version 2.4:
http://over.to/hpcomm
I decided to release another 2.x version, before the 3... Too many people
had problems with the current version.
Good news for all S(X) users! Use the XModem-feature of hpcomm, too, with
the library xmodem31.zip found at www.hpcalc.org/utils/comms It works even
faster than the built in XModem of the G-serie (about 6x faster than
Kermit)!!!
The next version will have a built in viewer for all kind of GROBs
(including animations) and many more convenient improvements.
 
----------------------------------------------------------------------
 
If the first four levels of the stack look like this...
56
25
4
10
then what does a programme need to look like to 
change these arguments to look like this on level 
one of the stack?
[ 56 25 4 10]
 
----------------------------------------------------------------------
 
----------------------------------------------------------------------
 
Hi
 
 
----------------------------------------------------------------------
 
First put the number 4 on the stack.  That's how many elements you have
Dave Zelkowski
 
----------------------------------------------------------------------
 
Drop me a line if you can.
Mike
----------------------------------------------------------------------
 
Can anyone point me to web site (or other source) featuring a legible
image of an HP-48GX keyboard?  I use a commercial keyboard overlay
that's not much fun to remove and replace, and on those infrequent
occasions that I have to do something that requires access to the
system utilities, I have to pry off the overlay to figure out which
keys do what.  The 48SX manual had a nice, full shot of the native
keyboard on the cover of the owner's manual, but the 48GX has only a
cropped shot that doesn't expose all they keys.  And, unless I've
missed something, there isn't a keyboard diagram in the manual, which
makes accessing the native shifted keys difficult when using an
overlay.
 
----------------------------------------------------------------------
 
The thing to do is to cheat: Let me recommend using one of the BMP
files for Emu48 as your diagram - check out
http://www.gulftel.com/~pattersc/emu48/
The one for 1024x768 seemed pretty good.
 
----------------------------------------------------------------------
 
There are many excellent pictures of the HP48 here:
http://www.hpcalc.org/pc/pictures/
-Joe-
This | is not a pipe.  -- Rene Magritte
http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own    
----------------------------------------------------------------------
 
  
Well i saw some other strange mails on this matter, who recommended
working with list????
You i CAN use the Edat list, the same list that is used for the
static things in the calculator.
[ x1 y1 ]  E+          (* 'E' is sigma *)
And the point is in the vector of Edat!! NO LIST JOX NEEDS!
----------------------------------------------------------------------
 
best
I am in the process of creating land surveying data collection software for
the HP48.  
My original idea for storing coordinate data was to pair the point
identifier with the coordinate into a list as follows:
{$PointID1 {Crd1} $PointID2 {Crd2}... }
where {Crd} is equal to {%N %E $PointDescription} or {%N %E %EL
$PointDescription}
This allows the point identifier to be alphanumeric, and does not waste
precious memory by storing an elevation when it is not needed.  In SysRPL
you can retrieve the data with the use of NTHOF and NTHCOMPDROP.  For
updating a coordinate use PUTLIST and for adding a new coordinate use
This works well until the list becomes very large in which case execution
times start to become very slow.  Presently my software is using a data
format very similiar to the above except I have written assembly language
routines to convert the {Crd} portion into a string object.  I have also
written assembly language routines to extract, modify, delete and append to
the list which execute very fast even on very large coordinate files.
My software is presently in a testing phase.  I plan to release it as
shareware next month after I have been able to thoroughly test it and make
it compatible with a few more total stations.  If you want to save yourself
some headaches wait until my software is ready and give it a try!  I think
you'll like it!  I should have a web page ready very soon and when
everything is ready I'll post to this newsgroup and a few others.
Bill Webber, PLS
----------------------------------------------------------------------
 
Have you ever used the 'LSQ' function on the HP to do traverse or level
run adjustments?  It seems as though all the tools are there....just how
to implement them.
 
----------------------------------------------------------------------
 
I have only used it for a system of linear equations.
Where I have a vector of constants on level 2.
And a coefficient matrix on level 1.
??
Mike
 
----------------------------------------------------------------------
 
Could someone please inform me on the best
way to store the points in a data file?  How would the data file be 
constructed
and accessed?
---------------------:: o
i can't help but think that someone is going to insist that you're going to
have to load up some huge peripheral library to do this...
while the easiest solution is to simply use the Matrix writer...
open it up with rightshift ENTER
then enter the first 2 or 3 axes for one coordinate, 
ENTER those, and hit the down {Q key} key...
then thereafter, all the points will be arranged in sets of 2 or 3, 
depending
on your initial input set.
then when you're done, hit ENTER to move the Array to the stack...
[[1 2] [3 4] ... ]
if you want to change them to a list of (x,y) coordinates
hmmmmmm...how to do this...???
DUP TAIL 0 + SWAP
DUP SIZE 1 - 1 SWAP SUB
is that it...does that work...???
 
----------------------------------------------------------------------
 
I made a useful library that I would like some of you to try.
It's basically a program manager with a lot of functions for better 
handling
of files and data.
It allows you to have different working directories where a file is kept
with information on what programs were run in the directory, or what
equations were used, or the name of the used data files, etc.
The main routine prompts these names, and once we choose on it runs the
program or opens up the file in the correct environment. Some of these
environments are build-in environment, but some not.
For example, one of them that I find very useful in my labs is the
environment to work with data. It easily creates a file that has the 
entered
data in form of a matrix, and name for each column of the matrix. Once the
file is created, there are several functions to manipulate this data, like 

a
much useful data fitting subroutine that the built-in one, or a command 
that
creates a new column in the matrix with the new values as a function of any
of the other columns. All this information is stored in the data file, and
can be easily viewed.
One important thing: there is also a routine to install programs so that
they can be managed by the library (all the programs are stored in the
HOME directory) and an uninstall routine too.
It has many more useful functions so please if you are willing to try it,
send me an email to mnt2@lehigh.edu
It would be very usefeul for mr if i could have other hp user's opinion.
----------------------------------------------------------------------
 
I have one problem with the EMU48 software. First I had dificulties with
the correct romdump, then with the .bmp that wasn't in the right
directorie. Then I tought everything should be right. Now I get the
error: Under german Nt 4.0 : Dr.Watson f237r NT:
Ein Anwendungsfehler ist aufgetreten. 
Ausnahme: Zugiffsverletzung (0x0000005), Adresse: 0x004195ff.
Wich means something like: violation of memory-zone.
I get something similar under win 95. 
 
----------------------------------------------------------------------
 
SP3.
It works fine on some computers (486DX2/66, 486DX4/133, P/200, P2/233,
P2/333). All needed Emu48 modules (Emu48.exe, ROM.48G, *.KML, *.BMP,
(BEEP.EXT)) must be in the same directory. Check and/or Convert the
downloaded ROM with the CONVERT.EXE utility deliverd with the original 
Emu48
V1.0 version. The checksums of the ROM must be OK.
 
----------------------------------------------------------------------
 
cann someone say me, how to convert some lists like:
{1 2 3}
{2 3 4}
{3 4 5}
in a Matrix:
[[1 2 3]
 [2 3 4]
 [3 4 5]]
 
----------------------------------------------------------------------
 
If on the stack you have:
{{ 1 2 3 } { 4 5 6 } { 7 8 9 }}
AR~LST will give you :
[[1 2 3]
[4 5 6]
[7 8 9]]
If you have
[[1 2 3]
[4 5 6]
[7 8 9]]
AR~LST
You will get:
{{ 1 2 3 } { 4 5 6 } { 7 8 9 }}
This works on any kind of matrices: Symbolic, string, real, complex etc...
Jean-Yves
 
----------------------------------------------------------------------
 
If you have ERABLE installed, the command AXL will take vectors or
matrices in standard form and convert them to list form and will convert
vectors or matrices in list forms to standard forms if all entries are
numbers.
 
----------------------------------------------------------------------
 
When you did this to all of them you just enter the number of columns you
or (if you really need it) mail me and I'll try to work something out.
I already did it:
<<
  <<
    1 n FOR c
    NEXT
You just have to put the number of lists you want to convert on the stack
below the lists and EVAL the prog.
 
----------------------------------------------------------------------
 
in this time I tried on my way and got that:
<< {1 2} SWAP 2 SWAP PUT
if {a b}
   {c d}
is the stack
the result ist {{a b} {c d}}
And i found, that ALG48 sees that as a matrix too..
julien
----------------------------------------------------------------------
 
 
----------------------------------------------------------------------
 
It started with 'minus' key button on my 48GX that work intermittenly.
it works one day but not the next day.
Now,... all keys on the right row work the same way as the 'minus'
key.
Does anyone has the same problem ? Any solution ?
Could this be just a contact problem ?
Could this because of the wheather of oregon (rain a lot) ?
----------------------------------------------------------------------
 
Keys are easier to operate when using gloves instead of mittens.
-Joe- 
----------------------------------------------------------------------
 
I sometimes have the problem that all the keys work as the 8 key, including
the on key, so I can't turn the calculator off. :,(
Francisco Hidalgo Sol207
fhs@satlink.com
----------------------------------------------------------------------
 
--------------------:: o
someone might easily accuse me of being hysterical, but i think it's 
broken.
call HP* and maybe (?) they'll replace it for FREE...! otherwise it'll 
probably
cost around a C note.
...............
* 1 541-757-2000
  ??? 
it seems to me that whenever i call HP, they've changed their numbers...???
 
----------------------------------------------------------------------
 
 
(1) Will the development tools such as rplcomp, sasm, etc that reference
the SX (such as ROM addresses, etc.) also work with the GX?
(2) Since these tools were developed prior to the GX being released, are
there certain GX features / commands that these tools do not exploit?
Thanx...Roger
----------------------------------------------------------------------
===
Subject: HP program
Steve
----------------------------------------------------------------------
 
In addition to Dave's answer:
If you like to group your CST-menu (to have some sort of directories)
you can try this structure:
{Name2 << {...}
 and so on (inlcuding the neccessary closing brackets)
Name
 is the name you want to give to the *directory/folder*, which you are
adding to the CST-menu
{...}
stands for your list of CST-functions (like for the normal CST-menu)
you whish to have in that specific first *directory*
Name2
is the name for the second *directory*....
Store the above list of lists under the name CST (like usual)
I hope answer is clear to you.
Happy custom gouping :-)
 
----------------------------------------------------------------------
 
This is all good.  One more question, though. 
Is there a way to put things like IF, END, etc. in a custom menu?
I now know I can put things like UBASE, UFACT, and so on.
I guess this is one difference between commands and syntactic keywords 
(what's the HP term for these?).
-- 
Steve
----------------------------------------------------------------------
 
HP calls them conditional structure commands.
Several requests for clarification have been posted.  However, if what
you want to do is to use these keys as typing aids (i.e. you press the
appropriate key in your menu and IF is inserted into the command line;
would be useful for keying in a program), just insert the word as a
string.  Ex:
{ IF } 'CST' STO
It may be helpful to make it  IF , with spaces.
 
----------------------------------------------------------------------
 
Hi Steve,
I'm not really sure what you mean by adding IF, END... to  a custom
menu. Do you want to call those commands from a custom menu, or do you
want to create a custom menu, that acts depending on a specific
situation?
  
----------------------------------------------------------------------
 
Anything you can put in a list, you can put into either a temporary menu
or a custom menu. What you will get from pressing the appropriate softkey
will depend on what kind of an object it represents. Commands will act
like commands, etc.
----------------------------------------------------------------------
 
Just put them in a list as I've shown below and save it as CST.
{ '1_m' '1_s' '1_kg' '1_K' }
Then when you are in the directory where you have it stored press the
CST key on the keypad(not the CST softkey).
Dave Zelkowski
Am I becoming a geek? I'm starting to be able to answer questions on
this group.
 
----------------------------------------------------------------------
 
button, and allowing me to have a key for things like V/mol*cm.
 
----------------------------------------------------------------------
 
Anyone know a Really good program that can transfer PC to HP 48GX from
serial port
We do not receive wisdom, we must discover it for ourselves,
after a journey through the wilderness which no one else can
make for us, which no on else can spare us... - Marcel Proust
 
----------------------------------------------------------------------
 
I won't say it's very good, but try my HPCOMM:
http://over.to/hpcomm
It has a explorer-like graphical user interface and supports fast XModem
transfer (even for the S(X)).
Check it out and send you comments to me....
James
 
----------------------------------------------------------------------
 
Not sure what the best one is, but my favorite so far is Hp-Ymodem2.2 by
Laurent Bernard.  It's super fast and very reliable.  You have to run it
through Dos, though, and no graphical interface.  The link is below.  Give
it a shot and let me know if you need help with some of the syntax.
http://hpcalc.org/utils/comms/
 
----------------------------------------------------------------------
 
I use Hyperterminal with Win95.  However the Hyperterminal kermit is
defective.  If you go to www.hilgraeve.com you can download Hyperterminal
Private Edition as a free upgrade (and it adds a lot of features).  It has 

a
good kermit that works just fine with the 48gx.
 
----------------------------------------------------------------------
 
It depends on what you mean by really good. Do you mean guish? Or
 dependable? 
 
----------------------------------------------------------------------
 
Nate,
    I don't know the answer to your question, but when you find out let me
know.  I tried to connect via IR and have had no luck.  Cable, no prob.
Problems are just with ir.
Mike
 
----------------------------------------------------------------------
   
No problem: Extract the ROM of your HP48 with ROMDUMP, which should be 
provided by the Emu48 package.
 
----------------------------------------------------------------------
 
Just use the romdump wizard that comes with
EMU48 to download the ROM from your OWN HP48...
 
----------------------------------------------------------------------
 
I too downloaded EMU48.
But I am very leary about dumping my ROM when the install instructions
suggest 2 .dll files are missing. Also there are extra files that are
recommended for download that conflict with existing files.
I opened ROMDump and it just laid there.......
It seems that EMU48 is very popular, so I must be missing something.
HPShell is Great!
Mike
 
----------------------------------------------------------------------
 
To anyone who has experience in the Cris Ramos' c-compiler ( Th216 C-
compiler ). Does anyone know if it is possible to compile any C-program, 
I mean his version of the program which just says printf(hello 
world!/n) is a lot different. It includes a lot of code to do the trick. 
I have the code which came in the examples directory of the C-compiler of 
his site. I also investigated the entire newsgroup for the subject of 
hp48 c-compilers but I have found nothing interesting.
----------------------------------------------------------------------
 
About once a month someone asks a similar question.  The best answer I have
seen was posted recently.  The problem is that there is no stdio library 
for
the compiler.
----------------------------------------------------------------------
 
This has to do with the fact that printf is not really part of the C
language, per se, but is just a function in a library of functions (often
written in C) for doing terminal i/o.  The HP48 is not really set up to do
input output in the same way most desktop type computers are, so things like 

printf are not as simple as it might at first seem.  There is a lot going on 

-- under the hood -- when you run the C one-liner int 
main(){printf(hellon);}
Checking out the size of the executable it creates on a normal computer will 

give some idea of the hidden complexity.
Consider, also, that while the computer in your vcr or nintendo may have
been programmed in C, it is doubtful that printf would have much meaning in 

this context.
 
----------------------------------------------------------------------
 
A beta means a test version of software, that is not supported in any 

way.
It is a normal method of testing software, you make a pre-release copy of 

it
available before it is finished so people can find the bugs for you.
And, an alpha generally means a version that is just barely ready for 

use,
and has gone no testing whatsoever.
 
features.
are
up or
GX).
DOS, Win
 
----------------------------------------------------------------------
 
I read a while back about a bad batch of hp48's where the screen went bad 

in
the upper left hand corner. Mine had started already nback then but hasn't
really gotten much worse. Anyways I broke the serial port so now it has to
go in for service.
So the question is Does anyone know if this was true(LCD failure) and if 

so
who should I contact? And if I still can?
I know I should have done it before but I couldn't imagine trying to do a
double major in Physics and Aerospace Engineering without my trusty HP. 
Even
if only for a few weeks
 
----------------------------------------------------------------------
 
Anyone know where to find a good options program to help calculate
binonmial pricing and black scholes, etc.  The exam is on Tuesday, so
dave
----------------------------------------------------------------------
 
This is the first program I have written with INFORM.  
%%HP: T(3)A(D)F(.);
<< %HELICITY ? { {
Gh222 (mdeg)  : }
{ cell path (cm):
} {
conc. in cell (M):
} {
peptide length: }
} 1 { -15 .1 .0001
20 } DUP INFORM
DROP EVAL DUP 5
ROLLD
  <<
(10*p*c*a)'
*100/(-40000*(1-
2.57/a))'
'g*100/(-40000*(1-2.57/a))' isn't and then the whole thing is
Can it be made faster and more efficient?
 
----------------------------------------------------------------------
 
well, i'm a french guy and a wonderful beginner on HP48.... Well, i've
got some programs who can only be run with XVIEW.... i've got it but, i
don't know how i can install XVIEW (like a library, i think)
Plz, can u help me?
It will super if u say me the procedure touch per touch  (because i 
won't
understand technical words... ;) )
 
----------------------------------------------------------------------
  
I'm assuming you've got XVIEW (library 1600).
Put it on your stack, that is, on the first line it should say
4:
3:
2:
1:         Library 1600:    XVI...
Then type
:0: 1600
The stack should now say
4:
3:
2:         Library 1600:    XVI...
1:                     0: 1600
Turn your calc off and then on.
Screen will flicker a little.
Check your library by pressing
 
----------------------------------------------------------------------
 
On the 48G, is the Memory label on the keyboard, the color turquoise on
all the calculators?  My son brought his G over for me to see the
difference in the KBD layout between the G and the SX.  Checking the
Memory I noticed that the  Memory label indicates the right shift
instead of the left shift.  By applying the right shift only brings up a
different option.  Are there anymore surprises like this one on the G's
keyboard?
----------------------------------------------------------------------
 
  Turquoise is right.  Right shifting on the VAR key brings up a choose box
interface for memory management while left shifting should bring up a
menu-based interface.  I don't know how this compares to the S series.
----------------------------------------------------------------------
 
Note that this is true for all the keys superlabeled only in turquoise. 
Very helpful when you get a little more advanced and want to do RECV
directly instead of going through the IO form...
 
----------------------------------------------------------------------
 
HPShell 3.12    Integrated Developing Environment
 FREEWARE           for all HP48 on DOS PCs.
* You program your HP48 and have never seen a real developing environment ?
* You look for an easy tool for file transfer ?
* You need someone managing your backups ?
* You would like somebody's doing all the stupid char conversion stuff ?
* And you want it all in one ?
Well, have a look on the HPShell !
After 5 years of continuous development it includes now a huge list of
features one would like to have while programming the HP48 or simply using 

it.
This is a (not complete) list of it's features:
-----------------------------------------------
* Multi functional Editor with Syntax highlighting, Word Wrapping,
  HP48 Displays, ...
* File Transfer (Kermit included, ANY other protocol is also useable!)
* Backup Management
* DOS and HP48 Directory Management
* Macros, Hotkeys, Toolbar
* free defineable Toolprograms
* and much more !!!
The HPShell can be found online on the Web at
Here you'll find a more complete list of features,
screen shots and a list of download-links spread all over world.
Differences to the previous verison (3.11):
-------------------------------------------
This release is a pure bug fix.
Source Code of the HPShell:
---------------------------
With this version the source code of the HPShell is released and
freely downloadable from the Internet. See
 
----------------------------------------------------------------------
 
I have Jazz library and I want to disassemble some useful commands.
Can anyone help me doing that. I have a list of the rom addresses, but I 
don't
know how to put an command in the stack without it's evaluation.
 
----------------------------------------------------------------------
 
Start with the input of these brackets: { },
While the list is open enter the desired command
(in alpha mode or directly from the selected menu).
When you have something like e.g. { UPDIR } apply
 the command HEAD on it and go on with the
desired JAZZ function.
 
----------------------------------------------------------------------
 
can anybody please tell me how I could bloc the hp's keyboard in a program
useful
commands using syseval , you you tell me also , thank you
----------------------------------------------------------------------
 
Has anyone noticed Erable making errors in limits that include undefined
structures?
I noticed Erable v3.2 made a mistake on the function
'a/(2*pi)*(((-1)^n+1)/(n^2-1))' 'n=1'.
It returns 'Infinite', when the correct answer is '-(1/4*a*i)'
This result is obtained by using l'Hospitals rule, and I thought that maybe
someone else have had trouble finding such limits?
 
----------------------------------------------------------------------
 
Is the PDF documentation for ALG48 v4.2 in pretty bad condition, or is it
just my printer that doesn't print it right (I have a HP DeskJet 820Cxi)?
The text seems very choppy/blurry and sides on tables are not complete.
It looks as if the printer is running out of toner, but this isn't the case
(many other documents - including PDF - are printing just fine).
Ice
----------------------------------------------------------------------
 
mine prints fine on a HP 5L,  PCL mode.
 
----------------------------------------------------------------------
 
Yeah, the PDF file is really bad.  However, the PS file is really good.  
It's
prints really will with ghostview.
 
----------------------------------------------------------------------
 
No, it really appears to suck. I even took it to work and ran it through
some pretty heavy duty image processing / font clean up routines - still
no joy.
Roger
 
----------------------------------------------------------------------
 
I spoke to someone about it a while back (I think it was the algorithm used
for trig functions).  I was told to check out the TI website, but I never
made it there.. You might see if you can find anything there.
-- 
----------------------------------------------------------------------
 
http://devil.ece.utexas.edu/ is a nice starting point ...
 
----------------------------------------------------------------------
I apologize for the off-track question, but does anyone out there
know how to get the HP-15C out of complex mode?
-- 
try a clear flag 8... that will do it.. if I remember correctly...
 
----------------------------------------------------------------------
 
I was wondering if anyone could tell me how to correctly write something
in User RPL that will move a grob up one pixel while you hold a button
down.
This is the Program I tried, but it didn't work.
(I don't know how to add comments properly either so I will use this as
a comment */)
   << 0 WAIT
      IF 34.1 ==
        DO
       << 'U' P STO+
              PAD {#67d #Ud} PAD GOR PICT STO PICTURE  */ PAD is the
grob name, and {#67d #Ud} are the        */ coordinates, whenever I try
to put the program on the stack it says invalid syntax when it gets to
#Ud.
I may just have the syntax wrong for calling the variable U's contents
into the y coordinates of my program.  If you could help me it would be
greatly appreciated.  I know I don't have past the starting of the loops
because I haven't gotten that far yet.
----------------------------------------------------------------------
I have an HP48G calculator, it awesome. I purchased it about 3 years ago. I
put it away for a while, now that I need it again, I got the message
Restore Memory? I said No. It worked fine, lost all my settings, but 
that
was alright. Then after that, it was really slow and lock up a lot. Now, I
can't even turn it on! I have new batteries. What could be the problem? Do
you for your help.
----------------------------------------------------------------------
 
I have an HP48G calculator, it awesome. I purchased it about 3 years ago. I
put it away for a while, now that I need it again, I got the message
Restore Memory? I said No. It worked fine, lost al...
----------------------------------------:: o
i went through a period after i got my computer, when i didn't use my 48 at
all, for like several months...and then when i started using it again, it 

was
acting really funny, and screwing up, and so on...
and then...after a while...like a couple of weeks...it stopped 'acting up' 

and
it's been working pretty well since...
and i was thinking...maybe it got lonely...!!!...and was acting out to 'test 

my
love' for it...???
i think that was it.
the other thing you might try, just to show it who's boss, is to take the
batteries out, which might actually be the more humain thing to do if 
you're
not going to use it for awhile, so that it won't be left in a 'near 
comatose'
state for several weeks...!!!
it might be more humain to just temporarily kill it, so that it won't 
notice
the passage of time...!
but anyway, as i was saying...take the batteries out for a few days, and 
maybe
the disincarnated atman that's currently in it will leave, and a new
disincarnate atman will move in, and you'll be able to start fresh with the 

new
 
 
----------------------------------------------------------------------
 
I have been having a problem with it.. It seems that when I install it
with java3.6a it seems to vanish out of memory sometimes.  I also have
alg48 and erable (different versions of them) and I always get the same
thing.  The lib vansishes but the memory is not freed up.. I warm boot
the calc and then the lib sometimes pops back and goes on working for a
bit.... but it usually wipes out pretty fast....
Any clues or just a bug?
also if there are any other things that you need to know about my calc
or software versions let me know..
I think rainEQ is an incredible program but it seems to have too many
bugs...
----------------------------------------------------------------------
 
RainEQ is nice concept but well known to be quite flaky; I contacted the
author about some of them a long time ago but nothing became of it.  So, I
don't know if the bugs will ever be fixed.
 
 
Does anyone know a (fast) FFT program for the SX???