Prev: HP41CX is looking for good home
Next: Poll: What additional software would you like to see for theHP 50G ?
From: Teuvo Yrjömäki on 14 Apr 2010 01:37
Hello

I received my 50g and 82240b yesterday. Got a few questions that I
didn't find answers for by some googling etc.

- when inputing integers, they show on the stack as, well, integers,
like '5'. But numbers rounded to integers by various means always show
up with a trailing decimal/fraction mark like '5.'. (remove quotation).
That kind of sucks as the trailing fraction marks also show up when
printed using the 82240b... So how to get rid of them? Example: RAND 5 *
0 RND

- How to print graphics from the 50g so that the stuff would really end
up being printed? With the default DELAY=1.8 it seems like printing just
freezes. DELAY=0 results in maybe one row printing, followed by a bit of
garbage and nothing more. Also the printing annunciator on the 50g
screen disappears after some time. Printing with DELAY=.1 maybe worked a
bit at least.

- The 82240b seems to drop characters randomly, how should the calc and
82240 be aligned? Same with the 48gx and 42s, eg printing a program
listing with the 42s almost worked but some linefeeds were missing. I
didn't figure out how to increase the delay on the 42s.

- I didn't manage to transfer stuff from the 48gx to the 50g using
infrared, does anyone know settings that would work?

- The different input forms for inputing the equation for plotting,
setting the window, axis ranges and stuff seem to be distributed in many
different forms accessible from the apps -> plot functions menu. So
setting the EQ and all plotting parameters requires opening lots of
those forms. Does the 50g have a similar input form for doing all the
stuff the way it is like on he 48gx? Or the text only way like on the
48gx (left shift + plot) or 48sx?

- One more thing---brute force fibonacci:
<< dup 1 > << dup 1 - fib swap 2 - fib + >> ift >> 'fib sto
Evaluating "10 fib" on the 48gx takes like 2..3 seconds, and 6..7s on
the 50g... isn't that quite strange!

Otherwise, the 50g feels quite OK, especially after setting the flag 117
"soft menus". The enter key should be somewhere else. It seems that lots
of stuff is hidden behind menus, choose boxes or something and doing
stuff requires more keypresses than on the 48gx. Anyway, I'd rather
preserve my 48gx as they don't make them any more and do my daily stuff
on the 50g... the 50g is my 8th hp calc in my modest collection.

thanks
teuvo
From: Andreas Möller on 14 Apr 2010 02:43
Hello,

> - when inputing integers, they show on the stack as, well,
integers,
> like '5'. But numbers rounded to integers by various means always
show
> up with a trailing decimal/fraction mark like '5.'. >
You can not get rid of the dot as this indicates a real number wheras
without the dot a ZINT (new object type in the 49G) is meant. If you
want to get rid of the dot for the reals consider writing a program
that turns reals into strings and then print the strings without the
quotes, or change them into a GROB and print the GROB

> - How to print graphics from the 50g so that the stuff would really end
> up being printed?
All I can say is that it worked for me when I tried it somewhere in
the past. Ensure that the batteries are fresh on the 50G and on the
printer and ensure that they are real close, not further away than a
centimeter

> - I didn't manage to transfer stuff from the 48gx to the 50g using
> infrared, does anyone know settings that would work?
This does not work due to the different protocolls that are used.

> setting the EQ and all plotting parameters requires opening lots of
> those forms. Does the 50g have a similar input form for doing all the
> stuff the way it is like on he 48gx? Or the text only way like on the
> 48gx (left shift + plot) or 48sx?
Everything is stored in PPAR (for printing to the screen) and in the
CASDIR, so you can manipulate these files directly.

> - One more thing---brute force fibonacci:
> << dup 1 > << dup 1 - fib swap 2 - fib + >> ift >> 'fib sto
> Evaluating "10 fib" on the 48gx takes like 2..3 seconds, and 6..7s on
> the 50g... isn't that quite strange!
Most likely your machine is set to exact mode (SF 105). In this case
the numbers in your program are treated as ZINTs which are always
slower than Reals. As in the 48GX a 1 is always a Real this is the
reason why it runs faster.
Change SF 105 and re-edit and save your program or edit your program
and add a dot to each number, e.g.
<< dup 1. > << dup 1. - fib swap 2. - fib + >> ift >> (note the dot)
Now your program should at least run as fast as on the 48GX. (This is
a common pitfall why migrating to the 49G but I believe it is
mentioned somewhere in the FAQ at hpcalc.org)

HTH,
Andreas
http://www.software49g.gmxhome.de
From: Teuvo Yrjömäki on 14 Apr 2010 03:10
Thanks Andreas,

Andreas M�ller schreef:
> want to get rid of the dot for the reals consider writing a program
> that turns reals into strings and then print the strings without the

In fact, that's what I did, though it feels a bit odd to do it this way.

> Everything is stored in PPAR (for printing to the screen) and in the
> CASDIR, so you can manipulate these files directly.

Ok. Maybe it would be worth to create a small lib with custom menu
shortcuts to define the plot ranges, eq, etc.

One more thing: I have rom 2.15 on the device and I can't get the
development library to show up in the APPS menu or anywhere else. Did
256 ATTACH as instructed in the AUR...?? Also, for example, typing ASM
only leaves the symbol 'ASM' on the stack.

o
From: Andreas Möller on 14 Apr 2010 03:29
Hello,

> Ok. Maybe it would be worth to create a small lib with custom menu
> shortcuts to define the plot ranges, eq, etc.
Go ahead ;-)

> One more thing: I have rom 2.15 on the device and I can't get the
> development library to show up in the APPS menu or anywhere else.
Did
> 256 ATTACH as instructed in the AUR...?? Also, for example, typing
ASM
> only leaves the symbol 'ASM' on the stack.

Quoting from an "Introduction to Saturn Assembly Language", page 27:
If you set system flag -86, the development library will automatically
attach itself after every warmstart.

So you might want to check your system flag -86.

Otherwise you can directly access the LIB (wether attached or not) by
typing
256 MENU [ENTER]. This will bring you in the LIB menu of the LIB.

HTH,
Andreas
http://www.software49g.gmxhome.de
From: John H Meyers on 14 Apr 2010 06:11
On 4/14/2010 12:37 AM, Teuvo wrote:

> When inputing integers, they show on the stack as, well, integers, like '5'.
> But numbers rounded to integers by various means
> always show up with a trailing decimal/fraction mark like '5.'

Per design, to distinguish new object type (integer) from "real"
(integer calculation, in memory, also generally takes longer
than floating point calculation, in registers)

> Example: RAND 5 * 0 RND

RAND 5 * 0 RND R\->I

However, if you want random integers 1 through 5
RAND 5 * CEIL R\->I

Random integers 0 through 4
RAND 5 * CEIL 1 - R\->I

R\->I converts real-valued integers to "integer" type (w/o decimal point),
but errors if the floating point value has a fractional part.

To convert "floating" integers but leave other real values alone:
\<< DUP FP NOT { R\->I } IFT \>>

To also ignore all other object types, without error:
\<< DUP TYPE NOT { DUP FP NOT { R\->I } IFT } IFT \>>

If you want to avoid converting anything of magnitude 1E11 or greater
(which is too big to have any fractional part, but may not really
represent an integer), then a bit more programming would be called for.


As to plot forms, in the 48G series PLOT form, you have to click OPTS
to go to a second form for more settings, so you really have
_two_ separate forms in the 48G series, collectively containing
pretty much what the 49G/50G series has in its two WIN and 2D/3D forms,
and you can switch between the all new forms
by just pressing the keys for another -- is that bad?

GRAPH goes directly to [re]drawing the graph,
without having to go through another form to get there.

The Y= form in the 49G/50G series adds something which the 48G series
does not have at all -- the ability to separately enter and manipulate
an entire _list_ of equations. TBLSET and TABLE provide even more
that the 48G doesn't have, so why shouldn't these also be separate,
and how might you design all of these to be as easily accessible?

My only little peeve is that they are all left-shifted -- in RPN mode,
if you are not careful about _holding_ the shift key,
you could end up _storing_ stack items into some variables,
instead of invoking plot forms -- this would have been
a bit less dangerous if right shifts had been used
(it's possible to re-assign all top row graph functions via a program,
if desired, to the right-shift-hold function keys, with which
any comparable key pressing slips would only _recall_ variables).


As to various command menus, all of the original 48G menus
are in the 49G/50G series, but some of these are unreachable
(or harder to reach) by keyboard navigation, sometimes the only way
being to perform a [T]MENU command with the menu number as argument.

E.g. the 48G "green" command menus on numeric keys:

SOLVE (74), PLOT (81), SYMBOLIC (93),
TIME (94), STAT (96), UNITS (59),
I/O (104), LIBRARY (110), EQLIB (113)

Alternate ways to get to some of these menus:

SOLVE (74) Right-hold NUM.SLV
TIME (94): Right-hold TIME
UNITS (59): Right-shift UNITS TOOLS

Top row 48G "green" command menus:

CHARS (62): Right-hold CHARS
MODES (63): PRG NXT MODES
MEMORY (70): PRG MEM
STACK (73): TOOL STACK

Everything in the 48G was a perfect "fit" to the keyboard,
but with the subsequent weight gain of the next series,
some of the seams became unseemly.


[r->] [OFF]
 |  Next  |  Last
Pages: 1 2
Prev: HP41CX is looking for good home
Next: Poll: What additional software would you like to see for theHP 50G ?