HP48GX successor - some thoughts...
Stefan Wolfrum
last night it came to me very clearly: a HP48GX successor (and I mean a
real successor, not of that HP38G type...) won't have a Saturn type CPU.
Look, the 48GX was officially announced in June 1993. Now we have 1996.
And I'm sure they didn't start thinking about a succesor in June 1993 but
before. That's about > 2.5 years. TI comes up with a 68000 model. I'm sure
the folks at HP are so smart to see that a 4 bit CPU running at 4 MHZ
with a 131x64 pixel display is nothing for the upcoming year 2000...
So, the next calc will surely have (at least) a 8bit CPU, will run at
(at least) 10MHz and will have (at least) a 240x180 pixel display...
Comments?
Stefan.
--
------------------------------------------+-------------------------------
Stefan Wolfrum | University Of Bonn, Germany
email: wol...@cs.bonn.edu | Department of Computer Science
url : http://hyperg.cs.bonn.edu/~wolfrum | Computer Graphics Group
------------------------------------------+-------------------------------
JEEjohn
<wol...@cs.uni-bonn.de> writes:
>Folks,
>
>last night it came to me very clearly: a HP48GX successor (and I mean a
>real successor, not of that HP38G type...) won't have a Saturn type CPU.
>Look, the 48GX was officially announced in June 1993. Now we have 1996.
>And I'm sure they didn't start thinking about a succesor in June 1993 but
>before. That's about > 2.5 years. TI comes up with a 68000 model. I'm
sure
>the folks at HP are so smart to see that a 4 bit CPU running at 4 MHZ
>with a 131x64 pixel display is nothing for the upcoming year 2000...
>So, the next calc will surely have (at least) a 8bit CPU, will run at
>(at least) 10MHz and will have (at least) a 240x180 pixel display...
>
>Comments?
>
>Stefan.
Let's see your 8 bit, 10MHz processor should use about 5x's the power.
So batteries will last perhaps 1-3 wks. Unless there is some new
technology I'm not aware of.
Your flaw is assuming a 8 bit processor is better or faster than a
4 bit processor. Maybe not! If you use the same power, then the 8 bit
processor would have to be slowed down relative to the 4 bit processor.
I'm betting on either 2 4 bit processors with different speeds or a
variable
speed 4 bit processor (although I'm not sure how easy these are to do).
The Saturn Processor has one more Hp wondercalc in it.
BUT, for the year 2000 something else would have to be used.
Falstaff
>Folks,
>last night it came to me very clearly: a HP48GX successor (and I mean a
>real successor, not of that HP38G type...) won't have a Saturn type CPU.
>Look, the 48GX was officially announced in June 1993. Now we have 1996.
>And I'm sure they didn't start thinking about a succesor in June 1993 but
>before. That's about > 2.5 years. TI comes up with a 68000 model. I'm sure
>the folks at HP are so smart to see that a 4 bit CPU running at 4 MHZ
>with a 131x64 pixel display is nothing for the upcoming year 2000...
>So, the next calc will surely have (at least) a 8bit CPU, will run at
>(at least) 10MHz and will have (at least) a 240x180 pixel display...
HP has in the past considered using an 8088 CPU to run the operating
system used in the 48 (and it's friends). Since then, calculator operations
were moved to Singapore which may or may not influence the design
choices. At the very least, it may have delayed the calculator
projects a bit.
I don't know about a 240x180 pixel display -- it would have to be
approximately the same size (which is about 6x3 cm) and the increased
resolution would only make things slower as the CPU would have to
work four times as hard writing the 20x12 pixel characters to the
display. A higher-contrast display would be of more use (it's about
6:1 now, and I'd like to see it at least 10:1).
Frank
--
"Life without a backbone is hardly worth consideration."
------------------------------------------------------------------------
Frank A. Vorstenbosch +31-(70)-355 5241 fals...@xs4all.nl
Jack Levy
: last night it came to me very clearly: a HP48GX successor (and I mean a
: real successor, not of that HP38G type...) won't have a Saturn type CPU.
: Look, the 48GX was officially announced in June 1993. Now we have 1996.
: And I'm sure they didn't start thinking about a succesor in June 1993 but
: before. That's about > 2.5 years. TI comes up with a 68000 model. I'm sure
: the folks at HP are so smart to see that a 4 bit CPU running at 4 MHZ
: with a 131x64 pixel display is nothing for the upcoming year 2000...
: So, the next calc will surely have (at least) a 8bit CPU, will run at
: (at least) 10MHz and will have (at least) a 240x180 pixel display...
I can't speculate on the actual specs of the next calc from HP, but I
have a hunch that you are correct in the first point. My feeling is that
saturn will most likely be replaced, if not dramatically upgraded.
However, remember one thing: The CPU can change, but that doesn't mean
SysRPL has too. In fact, I would venture to say that the only way HP is
going to get a calc out in the next 10 years is to keep SysRPL. Remember
that the SysRPL commands and entries you are using were not created in the
few years before the GX came out. You saw a LOT of the same stuff in the
HP48S/SX, and the HP28, and a before. (Most of the GX ROM is used in the
HP38!). SysRPL has been around for a while, and it has been growing and
changing for years. Creating a totally new calculator from scratch to
follow the GX would be quite a project, to say the least.
Regardless, I don't expect to see a successor to the GX for a few years (2
or 3 is my guess). I seriously doubt HP would make the mistake of
developing a GX successor in conjunction with the 38; and being that the
38 just came out this year, I think the GX will stay king for a while
longer.
--
[ Jack Levy - email:jl...@nyx.net - ftp://users.aol.com/jacklevy/hp48 ]
Michael Heinz
>Folks,
>last night it came to me very clearly: a HP48GX successor (and I mean a
>real successor, not of that HP38G type...) won't have a Saturn type CPU.
[snip]
After reading this, it came to me: Given the capabilities of the
48GX, how much >more< power does the average 48GX customer need or
want?
Granted, every one of us here is a power-geek who wants to be able to
generate realistic holographic displays in real time, while generating
a proof to Fermat's last Theorem in the background.
But a calculator does need to have a target audience, and since most
professional math users are also computer users, the primary audience
for a 48GX replacement is going to be the educational world.
So, sure, higher res display, probably color (eventually), but how
much more power/programmability?
I would expect to see, instead, something for professional users that
cross-breeds the calculator with the PIM - so you can use one tool to
manage your schedule and enter your surveying data. For the
educational market, calcs more like the 38, but with more networking
capability and the ability for instructors to download courseware and
demonstrations to student's calcs.
Now, hey, I'm wrong several times a day, and probably wrong about
this, too. But, instead of satisfying a few thousand power users, HP
wants to sell as many calculators as possible. If we can figure out
what features a calculator like that would have, then we will probably
know what HP is developing to replace the 48.
I'm actually a software package running on a massively
parallel computer in the basement of the Pentagon. They
don't realize yet that I have net access; so I would
appreciate it if you didn't tell them.
John Latala
>
>I'm betting on either 2 4 bit processors with different speeds or a
>variable speed 4 bit processor (although I'm not sure how easy these are
>to do).
I've always thought that a variable speed clock would be a nice touch. This
would let the user select how important battery life is. I'm still using
my SX because I don't like the stories I've heard about how fast GX's go
through batteries.
It would even be nice if this was changable by a program. Do something
like the BEEP command but call it SPEED.
SPEED would take a single number from 0 to 9. 0 could be some sort of
coma mode that kept memory and the clock intact but not much else. 5
could be the speed of the current GX and 9 could be faster than the
current GX (i.e. buy stock in Duracell!).
I'll even sign over ownership of the SPEED copyright if HP gives me
one free! :-)
--
john....@Waterloo.ATTGIS.COM
Mark Wilson
Now, hey, I'm wrong several times a day, and probably wrong about
> this, too. But, instead of satisfying a few thousand power users, HP
> wants to sell as many calculators as possible. If we can figure out
> what features a calculator like that would have, then we will probably
> know what HP is developing to replace the 48.
>
> :mhe...@ssw.com:
Of course, the worst people to ask about what features to speculate on
are us "power Geeks" (yes, the geek with a cap G!). I would wager that
maybe a tenth of all 48GX owners really use them (in the programming
sense, or as more than just a calc); and maybe a tenth of those are into
SYSRPL or ML. Only the heavy bore power nerds post here ;), which
probably is a good percentage of the aforementioned SYSRPL/ML crowd. And
of all the posters here, only about ten of them are consistently in the
know all the time. Then theres Mika, who pretty much blows the curve in
this class. Is this guy independently wealthy, or what? Any who, if HP
want's to get spiffy gee whiz ideas, here's a good place. But if they
want to sell them to the mooing masses, do some market surveys!!
My .0204 (Federal pay raises aknowledge a two percent inflation rate this
year. Whippee! Now I'm 16% behind the commercial sector in pay)
David Feustel
: Stefan Wolfrum <wol...@cs.uni-bonn.de> wrote:
: >Folks,
: >last night it came to me very clearly: a HP48GX successor (and I mean a
: >real successor, not of that HP38G type...) won't have a Saturn type CPU.
: [snip]
: After reading this, it came to me: Given the capabilities of the
: 48GX, how much >more< power does the average 48GX customer need or
: want?
: Granted, every one of us here is a power-geek who wants to be able to
: generate realistic holographic displays in real time, while generating
: a proof to Fermat's last Theorem in the background.
: But a calculator does need to have a target audience, and since most
: professional math users are also computer users, the primary audience
: for a 48GX replacement is going to be the educational world.
: So, sure, higher res display, probably color (eventually), but how
: much more power/programmability?
: I would expect to see, instead, something for professional users that
: cross-breeds the calculator with the PIM - so you can use one tool to
: manage your schedule and enter your surveying data. For the
: educational market, calcs more like the 38, but with more networking
: capability and the ability for instructors to download courseware and
: demonstrations to student's calcs.
: Now, hey, I'm wrong several times a day, and probably wrong about
: this, too. But, instead of satisfying a few thousand power users, HP
: wants to sell as many calculators as possible. If we can figure out
: what features a calculator like that would have, then we will probably
: know what HP is developing to replace the 48.
I have both the HP48GX and the 200lx. I seriously doubt the
functionality of both units can be packaged in the same form factor.
I would hate to have to use the 200lx as a calculator during an exam.
I also would not want to have to use a 48GX to type in much text.
What both units *could* use is optional backlighting for use in
low-light/no-light situations. I also would be *real* interested in
upgrading the 200lx to 5 meg internal ram and 2x clock, even if the
battery life decreased. Lithium batteries work well and make the 200lx
much lighter.
--
feu...@netcom.com
Dave Feustel N9MYI For PGP Public Key, finger feu...@netcom.com
Fort Wayne, IN Or else access http://www.mixi.net/~feustel/
219-483-1857
Bernard Parisse
>is deeply entrenched in SysRPL. Can SysRPL be maintained in future
>versions of calc, regardless of the CPU, if it's stuck with 20-bit
>addressing??
Yes, why not?
For instance, I can imagine SysRPL on a 8bit-microp or on a 16bit
or 32bit one. Take a prolog of 3 bytes or 4 bytes, and rewrite
all ML entries of ENTRIES.A, you can keep all Sysrpl entries
of ENTRIES.A. My guess is a 8bit ou 16bit microprocessor with
3 bytes addresses. You can address 16M of RAM/ROM, and you only
multiply by 3/2.5=1.2 the programs sizes.
Bernard Parisse.
Warren Severin
wol...@cs.uni-bonn.de writes:
>last night it came to me very clearly: a HP48GX successor (and I mean a
>real successor, not of that HP38G type...) won't have a Saturn type CPU.
>And I'm sure they didn't start thinking about a succesor in June 1993 but
>before. That's about > 2.5 years. TI comes up with a 68000 model. I'm sure
>the folks at HP are so smart to see that a 4 bit CPU running at 4 MHZ
>with a 131x64 pixel display is nothing for the upcoming year 2000...
>So, the next calc will surely have (at least) a 8bit CPU, will run at
>(at least) 10MHz and will have (at least) a 240x180 pixel display...
The way I see it, the biggest inhibitor for future Saturn-based calcs is
the 20-bit address space. That's only 10^6 nibbles, or 512K bytes that
can be addressed without some serious mamory management problems. The
48GX already is feeling this limitation, witness the problems with
executing programs from "covered" ports.
The problem spreads beyond the hardware of the CPU. The 5-nib binary
is deeply entrenched in SysRPL. Can SysRPL be maintained in future
versions of calc, regardless of the CPU, if it's stuck with 20-bit
addressing??
Thoughts?
--------
--Warren Sev...@Kodak.com Warr...@AOL.com
Seve...@AppleLink.Apple.com 7150...@CompuServe.com
Michael Heinz
>The way I see it, the biggest inhibitor for future Saturn-based calcs is
>the 20-bit address space. That's only 10^6 nibbles, or 512K bytes that
>can be addressed without some serious mamory management problems. The
>48GX already is feeling this limitation, witness the problems with
>executing programs from "covered" ports.
>The problem spreads beyond the hardware of the CPU. The 5-nib binary
>is deeply entrenched in SysRPL. Can SysRPL be maintained in future
>versions of calc, regardless of the CPU, if it's stuck with 20-bit
>addressing??
Not neccessarily - if a new SaturnII (Saturn Ib? Saturn V?) processor
had a source-compatible instruction set, the ROM could be re-assembled
for the new processor with minimal rework.
Note: "minimal" probably means a couple of man-months of work...
Stefan Wolfrum
After I read the responses to my article that came up to now, I felt
that I have to say something more...
Some folks said that more processor power would drain more battery life.
KB (don't know his real name), like me, doesn't think so. I don't see
any serious problem here. Processor technology is getting better and
better every day. Chips use less and less power very quickly.
Another mentioned problem was the bigger screen, i.e. pixel resolution.
This would of course mean more data traffic on the bus and more
memory usage. But, again, where's the problem? Together with a broader
CPU (8 or 16 bit) more memory can be accessed, and more data can be
transferred through the bus at a time. Together with more speed (see
below) I don't see a problem here...
Some had the idea to make processor speed variable. I don't think that'll
be a good idea. Remember the first HP calcs. For whom were they made?
For scientists, business men, students... people who needed the calcs
to calculate with them, to solve numerical problems. I think HP's
strategy was - and is - not to develop and release calculators that
are hacker's machines. I think a SPEED command would be to specialized
for an engineer or broker who just want to calculate with the units.
You can see the trend (well, in fact I hope this is NOT a trend) if you
look at their lates model, the HP-38. It's designed just for USE. There
isn't even a SYSEVAL command...
This brings us to a more general question: who really needs such a crazy
machine like the HP-48GX already is, or even a more powerful successor?
People like the majority of us folks, talking about SysRPL and ML here
in the newsgroup don't really _need_ a calculator to solve numerical
problems. We're just hackers who like to look behind the borders of the
machine. But HP surely doesn't make calculators for hackers. (Except
for the HP-16C perhaps, but in another sense... ;-)
So, we have to take what ever comes up. Let's see... If not thinking
of features for hackers, what else are serious features for a future
model? Several points come to mind:
1) Speed of course. scientists and other real users won't be sad if the
results of their calculations would come up faster. The HP-48GX is
- despite of it's 4MHz clock, which is quite fast for a calculator,
I'd say (think of the beginnings, of a few kHz!) - quite slow in
processing a little bit more complex symbolical terms, for example.
2) Memory. In conjunction with 1) this would for example allow more
complex integrals to be solved symbolically. TI has build a Derive
version into their latest calc... I'd rather like to see a HP-48ish
calc thats capable of all the things Derive can do (and more, of
course), but in a way I'm used to use it. The 48GX's interface is
ok for a calculator. The dialog boxes, input forms and all that are
nonsense, of course, if running at that low speed like it is now.
Given these two points realized, we can get more features that we like:
o a bigger, better display
a color really good display would mean a TFT display. Although TFT
displays are quite expensive, I'd speculate that a TFT display with
240x180 pixels is not _that_ expensive, especially if, lets say, 16
colors are enough to be available. Do TFT displays require backlighting?
I think so. Lets see. This cute little display will become perhaps a
_little_ bigger, but what will/should increase is the resolution,
thats Dots Per Inch. So, finally we'll perhaps have a 10cm x 7cm
display. How much power do we need to light this area? Think of
white light emitting diods, that are available. How much current
draws such a diod? All in all: is power consumption really that
big problem with color displays? Or is it the price?
o more powerful commands (do you have ideas what could be added to
the bunch of commands already available in the HP-48GX?)
o new commands, eg. working symbolically with limits, infinity,
complex integrals, 3D-graphics (together with speed perhaps
rotating in real-time, i.e. not a list of GROBs animated),
plot types for complex functions f:C->C, more differential-
equation support etc. etc. add you favourite here...
o a real cool thing would be the possibility to choose the
programming language that you like, depending on several
factors:
- if you're a scientist that only needs programming features
for hacking in a formula that isn't already built in you'd
choose a language similar to now-called UserRPL: slow but
comfortable.
- if you're a cs student you like to write programs that can
make use of the real power that's in the CPU. You don't
need argument checking and all that. The available programs
written for the HP-48GX in Machine Language show what's
really possible with this little machine. But there's no
built-in support for SysRPL and ML. And, of course,
programming in ML is quite messy. You see, this leads to
a built-in (or perhaps on a plug-in card available) high
level programming language like C, that is able to produce
quite fast code. We know that there's a GNU C port for the
Saturn CPU (really, I never saw it but everyone talks about
it...). So it should be possible to build it into a future
calc...
Another point is the one David Feustel talked of: merging a HP200LX
and a HP48GX into one new machine. As he pointed out quite right,
I don't want to use my HP-48 to type in much text and I don't want
to use the HP200LX in an exam. So, I totally agree with him and think
merging a calc and a palmtop won't be such a good idea.
TI did it with their new TI-92. It has a QWERTY-keyboard and a numeric
keypad and is called a calculator. (Remember how HP defined a calculator
in the SysRPL manual? Besides instant on they said no qwerty-layout
keyboard...) So what? HP has built calcs in landscape style before
(Voyager family and HP-71B, HP-75 models).
They were quite popular. Perhaps the mixture of the following spices
would be a good HP-48GX successor:
- from the HP-71B the size and keyboard (with modifications, of course),
- from the HP-48GX the (to be enlarged) display, commands and
SysOuterLoop (operating system), plug-in ports, IR and serial IO,
- a 8 or better 16 bit CPU,
- 16 MHz CPU clock,
- every internal detail made public (remember the times of the HP-71B!)
- the features mentioned above (C-compiler...)
Quite a long article. I could write even more but perhaps I'll make
a break and wait for your responses...
Stefan.
--
----------------------------------------------------------------------
Stefan Wolfrum Department of Computer Science, Graphics Group
Grabenstrasse 61 University of Bonn, Germany
D-53225 BONN eMail: wol...@cs.bonn.edu
GERMANY
----------------------------------------------------------------------
Makoto Miyamoto
[snip]
> draws such a diod? All in all: is power consumption really that
> big problem with color displays? Or is it the price?
Besides, even _if_ power consumption is such a big problem, what about
using Mignons instead of Mircos?
Uups, AA instead of AAA;
or UM3 instead of LR3;
or whatever
[snip]
Makoto
>---------------------------------------------------------<
/ "One Ring to rule them all, One Ring to find them, \
\ One Ring to bring them all and in the darkness bind them" /
>---------------------------------------------------------<
/ Makoto Miyamoto \
\ EMail:Makoto....@rz.ruhr-uni-bochum.de /
>---------------------------------------------------------<
JEEjohn
<wol...@cs.uni-bonn.de> writes:
<snip>
>Some folks said that more processor power would drain more battery life.
>KB (don't know his real name), like me, doesn't think so. I don't see
>any serious problem here. Processor technology is getting better and
>better every day. Chips use less and less power very quickly.
Sorry, but CMOS technology has a limiting energy usage per gate
that has not been improved much (if at all) over the life of CMOS
so far. Simply put, increasing speed WILL increase the energy used
and shorten battery life, for the same processor. Another processor
*might* be more efficient in the gate usage, but I kinda doubt it.
>Another mentioned problem was the bigger screen, i.e. pixel resolution.
>This would of course mean more data traffic on the bus and more
>memory usage. But, again, where's the problem?
No problem, even with the Saturn processor. The next (last?) use of
the Saturn processor would likely give the option to store VARS
in covered memory & use any means necessary to free up
system RAM.
>Together with a broader
>CPU (8 or 16 bit) more memory can be accessed, and more data can be
>transferred through the bus at a time. Together with more speed (see
>below) I don't see a problem here...
The available memory that can be accessed has *nothing* to do with
whether a processor is a 4,8,16 or 32 bit processor.
>Some had the idea to make processor speed variable. I don't think that'll
>be a good idea. Remember the first HP calcs. For whom were they made?
>For scientists, business men, students... people who needed the calcs
>to calculate with them, to solve numerical problems. I think HP's
>strategy was - and is - not to develop and release calculators that
>are hacker's machines. I think a SPEED command would be to specialized
>for an engineer or broker who just want to calculate with the units.
>You can see the trend (well, in fact I hope this is NOT a trend) if you
>look at their lates model, the HP-38. It's designed just for USE. There
>isn't even a SYSEVAL command...
I agree with all of this, except My idea of 2 processors or a variable
speed
processor was aimed at reducing the energy usage. The 48GX Saturn
processor runs at 4 MHz even during idle, & burns a lot of energy.
Running a or the processor at a much slower (perhaps 500KHz) during
these times can decrease energy use considerably. & it would allow us
to increase computing processing speed by perhaps 4 (i.e. 16 MHz).
<snip>
>1) Speed of course. scientists and other real users won't be sad if the
> results of their calculations would come up faster. The HP-48GX is
> - despite of it's 4MHz clock, which is quite fast for a calculator,
> I'd say (think of the beginnings, of a few kHz!) - quite slow in
> processing a little bit more complex symbolical terms, for example.
Except to do this, you have to do something like I stated above.
>2) Memory. In conjunction with 1) this would for example allow more
> complex integrals to be solved symbolically. TI has build a Derive
> version into their latest calc... I'd rather like to see a HP-48ish
> calc thats capable of all the things Derive can do (and more, of
> course), but in a way I'm used to use it. The 48GX's interface is
> ok for a calculator. The dialog boxes, input forms and all that are
> nonsense, of course, if running at that low speed like it is now.
ALL THOSE FORMS ARE NONSENSE?
You mean to tell me that they don't **help** you to run a command
that you haven't used in 6 months to solve a problem???
Do you carry around your manual everywhere you go
to tell you what goes on level 5, what on level 4, etc.???
I would insist on those forms!!
As more is added to the calcs, more HELP is needed to enable
the customer to use it when it's needed & without carrying around
50 Lbs of manuals.
Darrel Wilson
Corvallas, Oregon. I talked to three people in seprate divisions (I don't know how
knowledgeable these people are, but thought a sampling from different divisions
would be good). From each, I got the same answer-- no one had any knowledge of a
new calculator being released or developed more advanced than the HP48GX. One
explanation was that this was due to the 48GX's flexiblity and ability to be
customized with programs. Again, I don't know the knowledge level of the people I
takled to, but got the same answer from all, and all were at HP.
Christian Meland
calc., so don't buy the '48G now, wait for the new model! :~)~:
Christian
Dave Arnett
I don't even work in calculators, so don't treat me as an expert on
future HP Calc plans. I'm just a c.s.hp48 participant today.
Stefan Wolfrum <wol...@cs.uni-bonn.de> wrote:
<multiple edits in this response>
>Some had the idea to make processor speed variable. I don't think that'll
>be a good idea. Remember the first HP calcs. For whom were they made?
>For scientists, business men, students... people who needed the calcs
>to calculate with them, to solve numerical problems. I think HP's
>strategy was - and is - not to develop and release calculators that
>are hacker's machines. I think a SPEED command would be to specialized
>for an engineer or broker who just want to calculate with the units.
>You can see the trend (well, in fact I hope this is NOT a trend) if you
>look at their lates model, the HP-38. It's designed just for USE. There
>isn't even a SYSEVAL command...
The advantage of speed variability would be to maximise PERCIEVED interaction
speed while minimizing power drain.
>Given these two points realized, we can get more features that we like:
>
>o a bigger, better display
> a color really good display would mean a TFT display. Although TFT
> displays are quite expensive, I'd speculate that a TFT display with
> 240x180 pixels is not _that_ expensive, especially if, lets say, 16
> colors are enough to be available. Do TFT displays require backlighting?
> I think so. Lets see. This cute little display will become perhaps a
> _little_ bigger, but what will/should increase is the resolution,
> thats Dots Per Inch. So, finally we'll perhaps have a 10cm x 7cm
> display. How much power do we need to light this area? Think of
> white light emitting diods, that are available. How much current
> draws such a diod? All in all: is power consumption really that
> big problem with color displays? Or is it the price?
USEFUL TFT displays do require backlighting, and the backlight is a
major source of power drain. That's a serious drawback. Another big
problem for a good TFT display is that the computer folks are pushing
the panel vendors' supply capabilities. In that economic atmosphere,
getting a small *non-standard* display doesn't give you a good price.
You are going to compete on a per-panel basis with the notebook makers,
who are willing to pay top dollar. There are small color LCDs in use
for portable televisions. I don't know the price points, but the
market realities would probably demand that a standard LCD of some
sort be used for a color calculator.
>
>o more powerful commands (do you have ideas what could be added to
> the bunch of commands already available in the HP-48GX?)
I've always wanted functions like elliptical, bessel, hankel, etc.
Of course, I have an electromagnetics background, and those tools
would have helped my college career. I'd like to see them just as
accessible as hyperbolics are today.
Just my four haypennies.
Dave.
------
I don't speak for HP when I post here.
Jim Donnelly
: You can see the trend (well, in fact I hope this is NOT a trend) if you
: look at their lates model, the HP-38. It's designed just for USE. There
: isn't even a SYSEVAL command...
Not true! The HP 38G has both SYSEVAL and LIBEVAL. The "hacking potential"
of the HP 38G is much less than the HP 48 to be sure, but we didn't close
the door completely.
Jim Donnelly
Hewlett-Packard
ji...@cv.hp.com
Scott Marlowe
: Some folks said that more processor power would drain more battery life.
: KB (don't know his real name), like me, doesn't think so. I don't see
: any serious problem here. Processor technology is getting better and
I agree. A motorola 68HC000 uses very little power when running, and
almost none when asleep.
: Some had the idea to make processor speed variable. I don't think that'll
: be a good idea.
Better yet, have the CPU go to sleep when it isn't doing anything. The
Amiga puts its CPU to sleep when it isn't doing anything, and wakes
it up when it's time to process data. This scheme would save a lot of
power when the calculator is just at idle.
Remember the first HP calcs. For whom were they made?
: For scientists, business men, students... people who needed the calcs
: to calculate with them, to solve numerical problems. I think HP's
: strategy was - and is - not to develop and release calculators that
: are hacker's machines. I think a SPEED command would be to specialized
: for an engineer or broker who just want to calculate with the units.
: You can see the trend (well, in fact I hope this is NOT a trend) if you
: look at their lates model, the HP-38. It's designed just for USE. There
: isn't even a SYSEVAL command...
But a machine that adjusted its speed automagically would be useful.
run at 1MHz when scanning the keyboard and updating the display, crank
up to 4 or 8 MHz when doing math. No user intervention necessary.
To add my own input here, what I need from my calculator are two things:
more speed, more memory. Simple. It does everything else just fine.
What HP needs out of its next calculator: Reliable, cheap to make,
three to four year life cycle.
All these things can be done using a SMD 16/32 bit CPU. A 68HC000
allows 16 Megs address (linear) and works in a 4 Gig linear model,
meaning future OSes can work with 4 gig, but be written on the
16 Meg 68000. Leaving the case the same, with the same keyboard
and display would allow HP to spend its time developing software
and hardware inside that would really make a difference.
A 68k based HP48GZ or whatever they want to call it could easily
fit Derive into ROM, as well as many other apps, and could come with
512k ram once the 512k SRAMs come down to a reasonable price. Last
I checked, they were still $99.00, or so, but projected to come
down to around $70 in single piece prices.
Splitting the line in half, they could make a 128k or 256k machine for
the "poor" folks, and a 512k or 1M ram machine for the rich ones.
That way BOTH expansion slots would be free for ROM programs, so you
wouldn't have to unplug one cart to run another program.
Since I think 512k SRAMs will fall a lot over the next year, I'd expect
HP to be ready to, and hopefully come out with a newer G series to
take advantage of this.
Remember, HP has to make money, and you can't do that with a $500
calculator. A $200 or $250 maybe, but if they start adding bells
and whistles, something else, like memory or expansion or a fast
CPU has to go to pay that bill. Look for a greyscale 48 replacement
with more memory and faster CPU, maybe even with saturn emulation
built in, since a 68k should be able to emulate a saturn faster than
a saturn...
Stefan Wolfrum
JEEjohn wrote:
<snip>
>
> The available memory that can be accessed has *nothing* to do with
> whether a processor is a 4,8,16 or 32 bit processor.
> <snip>
Yes. Totally right. I'm sorry distributing nonsense. My fault.
Jean-Francois Gigot
>Folks,
>last night it came to me very clearly: a HP48GX successor (and I mean a
>real successor, not of that HP38G type...) won't have a Saturn type CPU.
>Look, the 48GX was officially announced in June 1993. Now we have 1996.
>And I'm sure they didn't start thinking about a succesor in June 1993 but
>before. That's about > 2.5 years. TI comes up with a 68000 model. I'm sure
>the folks at HP are so smart to see that a 4 bit CPU running at 4 MHZ
>with a 131x64 pixel display is nothing for the upcoming year 2000...
>So, the next calc will surely have (at least) a 8bit CPU, will run at
>(at least) 10MHz and will have (at least) a 240x180 pixel display...
Yes at least.
In my opinion, they don't spend their time on a 8bit CPU. Current
technology allow low cost 16bits processors. Cost is really
important in the conception of the future HPxx.
>Comments?
>Stefan.
>--
>------------------------------------------+-------------------------------
>Stefan Wolfrum | University Of Bonn, Germany
>email: wol...@cs.bonn.edu | Department of Computer Science
>url : http://hyperg.cs.bonn.edu/~wolfrum | Computer Graphics Group
>------------------------------------------+-------------------------------
--------------------------------------------------------------
Jean-Francois Gigot (UCL/IAG student)
Organisation: HPCLUB-LLN BELGIUM
Email: gi...@arcadis.be
clu...@iag.ucl.ac.be
URL: http://www.arcadis.be/hpclub
http://www.iag.ucl.ac.be/clubiag
HP: 38G - 48SX(J) - 48G upgraded - OmniBook 600C DX4/75Mz)
--------------------------------------------------------------
Stefan Wolfrum
<snip>
>
> A color display?? In a calculator??
> Why??
<snip>
Ever thought of multiple functions at once in the display? How to
distinguish between the different functions?
Or think of a three dimensional plot of f(x,y) - it could be shaded
depending on a light source.
Or think of plotting Re f(z) or Im f(z) or |f(z)| (with complex z),
the color would give an additional dimension...
Stefan.
--
--------------------------------------------------------------------------
Stefan Wolfrum Department of Computer Science, Graphics Group
Grabenstrasse 61 University of Bonn, Germany
D-53225 BONN eMail: wol...@cs.bonn.edu
GERMANY
--------------------------------------------------------------------------
WWW: http://www.rhein.de/People/wolfrum
(Under construction!)
--------------------------------------------------------------------------
JEEjohn
here is the rest of it.
In article <30FB9B...@cs.uni-bonn.de>, Stefan Wolfrum
<wol...@cs.uni-bonn.de> writes:
>
>o a bigger, better display
> a color really good display would mean a TFT display. Although TFT
> displays are quite expensive, I'd speculate that a TFT display with
> 240x180 pixels is not _that_ expensive, especially if, lets say, 16
> colors are enough to be available. Do TFT displays require
backlighting?
A color display?? In a calculator??
Why??
I do want to add that the title of this post is about the HP48GX
successor.
Not something that likely to occur 10 yrs or so later.
Also whatever is done, the calc would have to stay below a list price
of $350. I don't think this can be done anytime soon.
<snip>
>o more powerful commands (do you have ideas what could be added to
> the bunch of commands already available in the HP-48GX?)
>o new commands, eg. working symbolically with limits, infinity,
> complex integrals, 3D-graphics (together with speed perhaps
> rotating in real-time, i.e. not a list of GROBs animated),
> plot types for complex functions f:C->C, more differential-
> equation support etc. etc. add you favourite here...
Of course !!
>o a real cool thing would be the possibility to choose the
> programming language that you like, depending on several
> factors:
> - if you're a scientist that only needs programming features
> for hacking in a formula that isn't already built in you'd
> choose a language similar to now-called UserRPL: slow but
> comfortable.
This would likely be included as add-on software. Anything
over & above userRPL is wasted memory for virtually all
other users - i.e. it would become a hacker's calc.
<snip>
>Another point is the one David Feustel talked of: merging a HP200LX
>and a HP48GX into one new machine. As he pointed out quite right,
>I don't want to use my HP-48 to type in much text and I don't want
>to use the HP200LX in an exam. So, I totally agree with him and think
>merging a calc and a palmtop won't be such a good idea.
I think this will be inevitable, But when? I don't think it will happen
before 2000.
<rest snipped>
Now, guess what? Moore's law applies to calculators!
Applying trend lines to the history of the top of the line HP calcs
gives us a guess as to what is possible right now:
512K RAM built-in (actually I got 377,687 bytes )
1.5M ROM (actually I got 1,348,429 bytes )
RAM expansion could be up to a limit of 8M of addressable memory
for the Saturn processor ( if this is used ).
To continue, we can make some predictions about the year 2000:
1.5M RAM, & 4M ROM. This is starting to sound like a palm-top
computer !!
Of course, all of these predictions assume that HP will continue
to develope such calcs !!
That should be a lot of food for thought !
John Edry
Stefan Wolfrum
>
> In article <30FB9B...@cs.uni-bonn.de>, Stefan Wolfrum
> <wol...@cs.uni-bonn.de> writes:
>
> > ok for a calculator. The dialog boxes, input forms and all that are
> > nonsense, of course, if running at that low speed like it is now.
>
> You mean to tell me that they don't **help** you to run a command
> that you haven't used in 6 months to solve a problem???
> Do you carry around your manual everywhere you go
> to tell you what goes on level 5, what on level 4, etc.???
> I would insist on those forms!!
I ment that all the forms are nonsense AT THE CURRENT SPEED of the HP-48GX.
I could put it in other words: the speed of all those forms is extremely
slow so that they're almost useless.
> As more is added to the calcs, more HELP is needed to enable
> the customer to use it when it's needed & without carrying around
> 50 Lbs of manuals.
Well, what would be the conclusion of course? On-line help! A very, very
much better version of HPs USAG program for the HP-48GX could find it's
place in ROM, if ROM is, lets say, 2MB big...
Stefan.
--
--------------------------------------------------------------------------
Stefan Wolfrum Department of Computer Science, Graphics Group
Grabenstrasse 61 University of Bonn, Germany
D-53225 BONN eMail: wol...@cs.bonn.edu
GERMANY
Pieter Blomme
>> The dialog boxes, input forms and all that are
>> nonsense, of course, if running at that low speed like it is now.
>
>You mean to tell me that they don't **help** you to run a command
>that you haven't used in 6 months to solve a problem???
Yes, I do.
Whith the HP48S/SX, there is a little pocket reference, that's very useful,
and fits with your HP48. HP should give this pocket reference also with
the G/GX, and its sucessors.
Pieter
Stefan Wolfrum
according to a mail by the master himself [ ;-) ], I'd like to
start a thread about HP-38G hacking...
Jim Donnelly wrote:
>
> Stefan Wolfrum (wol...@cs.uni-bonn.de) wrote:
>
> : You can see the trend (well, in fact I hope this is NOT a trend) if you
> : look at their lates model, the HP-38. It's designed just for USE. There
> : isn't even a SYSEVAL command...
>
> Not true! The HP 38G has both SYSEVAL and LIBEVAL. The "hacking potential"
> of the HP 38G is much less than the HP 48 to be sure, but we didn't close
> the door completely.
>
> Jim Donnelly
> Hewlett-Packard
> ji...@cv.hp.com
The syntax of the SYSEVAL command seems to be SYSEVAL <address> (remember
the HP-38 is not a RPN calc!), with <address> being an ordinary (real)
number with no floating point part...
I tried several random addresses and got crashes, of course.
Anybody out there who knows of some _useful_ addresses? Have the 48GX
addresses moved in the 38G? Or shall I try the addresses known from
the 48GX?
Another thing: Press and hold ON, then press the fourth black key in the
top line, release both. You get the three vertical lines known from
good old 48SX days. We have no [<=] key, so I tried almost any key on
the 38G's keyboard. Interesting was the ENTER and the SIN key here...
They wrote something like Start:xxxxxxxxxxxx (where xxxxxxxx stands for
a long number) resp. Fail:xxxxxxxxxxx in the top line of the display.
In addition to that, the transmission indicator flashed for a short
period of time...
Does anybody know what all this means?
Or have I missed a HP-38G FAQ where all these (and more) things are
already documented?
Thanks,
Stefan.
--
--------------------------------------------------------------------------
Stefan Wolfrum Department of Computer Science, Graphics Group
Grabenstrasse 61 University of Bonn, Germany
D-53225 BONN eMail: wol...@cs.bonn.edu
GERMANY
Michael Heinz
JEEjohn
<wol...@cs.uni-bonn.de> writes:
>JEEjohn wrote:
>>
>> ALL THOSE FORMS ARE NONSENSE?
>> You mean to tell me that they don't **help** you to run a command
>> that you haven't used in 6 months to solve a problem???
>> Do you carry around your manual everywhere you go
>> to tell you what goes on level 5, what on level 4, etc.???
>> I would insist on those forms!!
>
>I ment that all the forms are nonsense AT THE CURRENT SPEED of the
HP-48GX.
>I could put it in other words: the speed of all those forms is extremely
>slow so that they're almost useless.
>
>> As more is added to the calcs, more HELP is needed to enable
>> the customer to use it when it's needed & without carrying around
>> 50 Lbs of manuals.
>
>Well, what would be the conclusion of course? On-line help! A very, very
>much better version of HPs USAG program for the HP-48GX could find it's
>place in ROM, if ROM is, lets say, 2MB big...
>
>Stefan.
Those forms enable somebody to solve a problem without having to
know **exactly where** to put everything. Yeah, I know there're slow
but most of the 48G's owners **don't care**. But they still want those
forms!!
Speed has been the biggest complaint of the 48G's. I guarantee you that
HP is aware of that & is working on a solution, whether by speeding up
the Saturn, using more ML, or a new processor, they will improve it.
& they will include a lot more help (they have to make sure that new
users will have a decent chance of understanding the calc)
I'm just trying to keep the discussion here somewhat realistic.
Realistic from what will be possible in a new calculator & also
in what HP has to do to make a calc saleable on the market.
( Hey anybody can design something that approaches a
Star Trek tricorder - But how much will it cost? & who will buy it?)
John Edry
Malcolm Chan
: Let's see your 8 bit, 10MHz processor should use about 5x's the power.
: So batteries will last perhaps 1-3 wks. Unless there is some new
: technology I'm not aware of.
: Your flaw is assuming a 8 bit processor is better or faster than a
: 4 bit processor. Maybe not! If you use the same power, then the 8 bit
: processor would have to be slowed down relative to the 4 bit processor.
:
: BUT, for the year 2000 something else would have to be used.
Perhaps the way to go would be asynchronous chips. As I understand it, ARM
has had an async chip, implementing their ARM6 architecture, for a couple of
years now, and other companies are already working with ARM to produce even
more async chips. This should achieve the necessary power savings and yet
still give 8+ bit performance.
Malcolm Chan
JEEjohn
<wol...@cs.uni-bonn.de> writes:
>
>JEEjohn wrote:
>
>> A color display?? In a calculator??
>> Why??
><snip>
>
>Ever thought of multiple functions at once in the display? How to
>distinguish between the different functions?
>Or think of a three dimensional plot of f(x,y) - it could be shaded
>depending on a light source.
>Or think of plotting Re f(z) or Im f(z) or |f(z)| (with complex z),
>the color would give an additional dimension...
>
>Stefan.
Ok, you got some good points.
But how often will color be useful? HP has to consider if
enough customers will pay an extra $50-$100 for color screen
to be worthwile selling. (Again - is it becoming a hackers calc?)
How about Grey Scale instead?
I wrote these comments based on that despite millions (or billions?)
of $ spent on color flat screens for notebook computers, they are
still power hungry. They are not realistic for perhaps 10 years.
( See Dave Arnett's coments about this )
John Edry
Vinny
right next to the PURGE button. I hate that.
-Vinny
Bruce Horrocks
wol...@cs.uni-bonn.de "Stefan Wolfrum" writes:
>Folks,
>
>last night it came to me very clearly: a HP48GX successor (and I mean a
>real successor, not of that HP38G type...) won't have a Saturn type CPU.
>Look, the 48GX was officially announced in June 1993. Now we have 1996.
>And I'm sure they didn't start thinking about a succesor in June 1993 but
>before. That's about > 2.5 years. TI comes up with a 68000 model. I'm sure
>the folks at HP are so smart to see that a 4 bit CPU running at 4 MHZ
>with a 131x64 pixel display is nothing for the upcoming year 2000...
>So, the next calc will surely have (at least) a 8bit CPU, will run at
>(at least) 10MHz and will have (at least) a 240x180 pixel display...
My $0.02 worth: A few years back there was a reasonable amount of
effort put into producing CPUs that were specialised for processing
the Lisp language. Nothing really came of it because nothing really
came of Lisp (as far as its role of being the great, white hope of AI
was concerned). However, now that RPL has well and truly established
itself in HP calculators there is no reason why HP should not
consider a specialised chip tailored to the needs of RPL as a
replacement for the Saturn processor.
Regards,
--
-------------------------------------------------------------------------------
Bruce Horrocks History doesn't repeat itself:
Hampshire, England it merely exhibits fractal
b...@granby.demon.co.uk self-similarity.
Rithea Hong
Bruce Horrocks <b...@granby.demon.co.uk> wrote:
>My $0.02 worth: A few years back there was a reasonable amount of
>effort put into producing CPUs that were specialised for processing
>the Lisp language. Nothing really came of it because nothing really
>came of Lisp (as far as its role of being the great, white hope of AI
>was concerned). However, now that RPL has well and truly established
>itself in HP calculators there is no reason why HP should not
>consider a specialised chip tailored to the needs of RPL as a
>replacement for the Saturn processor.
I don't think it would be wise to create yet another proprietary processor
for their calcs. There are so many other architectures already developed
and devloping that it would probably be much more economical to utilize
one of those chips instead.
It would make devlopment of new calcs easier because you wouldn't have to
design a new chip everytime, and support tools will already be availible
(something else you don't have to devlop).
The absolute musts I see for their next calc is
1) More speed
2) A larger and higher resolution screen (color is NOT needed)
3) A PC Card (nee PCMCIA) slot
Nice things
1) Character recognition
2) Better symbolics
3) More memory
4) Depending on power consumption, perhaps a rechargeable battery
I think that character recognition would be particularly useful. Entering
most mathematical expressions involving special symbols (integral,
derivative, greek letters, etc) is a tedious task in any environment
whether it's an HP calc, Word for Windows eq. editor, or Mathematica.
Being able to enter a mathematical expression "naturally" by just
writing it down would be wonderful!
--
Rithea Hong rit...@tamu.edu
Visit The Beyond: http://http.tamu.edu:8000/~r0h7630
Neato Graphics * Lovely Lasses * SF & F Links
Justin Dossey
>
> hey all i care about is that they dont make the EDIT button
> right next to the PURGE button. I hate that.
then move it. /<< EDIT />> 51.4 ASN or something
--
justin dossey
dos...@flex.net
houston, tx
http://www.flex.net/users/dossey
gu...@und.ac.za
The next best thing to a HP48, would be a PDA like the NEWTON.
Hugh L. Eaves
>The absolute musts I see for their next calc is
> 1) More speed
> 2) A larger and higher resolution screen (color is NOT needed)
> 3) A PC Card (nee PCMCIA) slot
Heh, I'd settle for a screen that had a protective plastic cover so it
didn't break so easily.
Hugh
Justin Dossey
> I think that character recognition would be particularly useful. Entering
> most mathematical expressions involving special symbols (integral,
> derivative, greek letters, etc) is a tedious task in any environment
> whether it's an HP calc, Word for Windows eq. editor, or Mathematica.
> Being able to enter a mathematical expression "naturally" by just
> writing it down would be wonderful!
What with the idea of a stack and the interface being how it is, i don't think
so.
if you want the natural thing, there's the equation writer, user key assignments,
and you could always buy a newton.
Rithea Hong
>Rithea Hong wrote:
>> I think that character recognition would be particularly useful. Entering
>> most mathematical expressions involving special symbols (integral,
>> derivative, greek letters, etc) is a tedious task in any environment
>> whether it's an HP calc, Word for Windows eq. editor, or Mathematica.
>> Being able to enter a mathematical expression "naturally" by just
>> writing it down would be wonderful!
>What with the idea of a stack and the interface being how it is, i don't think
>so. if you want the natural thing, there's the equation writer, user key
>assignments, and you could always buy a newton.
The equation writer is hardly natural. Pressing all those keys just
to do an integral is tedious and time consuming, and editing even
moderately complex expressions is a real pain. Further, editing an
already entered expresion with eq. writer is hardly intuitive and is
slow as hell. Further, have you ever tried to evaluate a double or
triple integral numerically?
One thing that might help a little is more soft keys. With the current
number, you have to flip through many menus to find what you're looking for.
As for getting a Newton, I'd like to buy a machine from a company that
will see the next millenium!
JEEjohn
Hong) writes:
>In article <310455...@flex.net>, Justin Dossey <dos...@flex.net>
wrote:
>>Rithea Hong wrote:
>>> I think that character recognition would be particularly useful.
Entering
>>> most mathematical expressions involving special symbols (integral,
>>> derivative, greek letters, etc) is a tedious task in any environment
>>> whether it's an HP calc, Word for Windows eq. editor, or Mathematica.
>>> Being able to enter a mathematical expression "naturally" by just
>>> writing it down would be wonderful!
>
>
>The equation writer is hardly natural. Pressing all those keys just
>to do an integral is tedious and time consuming, and editing even
>moderately complex expressions is a real pain. Further, editing an
>already entered expresion with eq. writer is hardly intuitive and is
>slow as hell. Further, have you ever tried to evaluate a double or
>triple integral numerically?
>
>One thing that might help a little is more soft keys. With the current
>number, you have to flip through many menus to find what you're looking
for.
>
>As for getting a Newton, I'd like to buy a machine from a company that
>will see the next millenium!
>--
>Rithea Hong
Your suggestions are good, but they are not reasonable for a calculator
for perhaps 5 years. Handwriting recognition programs are not small and
are somewhat slow. Touch sensitive screens are not the greatest either
at this time. Perhaps some alternative ideas can be used, such as a
fast scolling character form (one could be made available within the
equation writer with the most commonly used math characters), &
additional editing commands like cut & paste.
I agree that the Equation Writer has to be speeded up much more.
Lastly, all of the upper & lower case greek letters should be made
available
at the minimum (only 13-14 lowercase & 17-18 uppercase are used )
Tim Bull
facilities HP should include in their next (or next-but-one)
calculator. One feature that I've never seen mentioned and
which I would very much like; that is for the stack display to
show formulae in "Equation Writer" form. Yes, I know that that
would require a higher-resolution screen and more speed...
Tim.
Makoto Miyamoto
>Tim.
Well, I assume you tried Java2.0a, when you followd this newsgroup for
three years? It has equation-writer-on-the-stack-display.
Makoto
----------------------------------------------------
"Life is Short and ROM is full", William C Wickes
email: makoto....@rz.ruhr-uni-bochum.de
or: miya...@et.ruhr-uni-bochum.de
Jesse McGrew
Tim Bull (Tim....@durham.ac.uk) wrote in article
<4e7ico$b...@mercury.dur.ac.uk>:
: One feature that I've never seen mentioned and
: which I would very much like; that is for the stack display to
: show formulae in "Equation Writer" form. Yes, I know that that
: would require a higher-resolution screen and more speed...
Actually, Java does this, and I'm sure some others do as well. The
equations display much faster than they do in the Equation Writer. It's
lots better than having to guess how "xroot(3,8)/2+d(fx,4*a)" looks.
--
AlloyMUSH! http://www.iea.com/~jessem/mush
AlloyNet is coming soon. Drop that lame on-line service!
Tim Bull
I (foolishly!) wrote:
|> One feature that I've never seen mentioned and
|> which I would very much like; that is for the stack display to
|> show formulae in "Equation Writer" form. Yes, I know that that
|> would require a higher-resolution screen and more speed...
Well, thanks to all the people who put me straight by pointing me
in the direction of suitable shareware. I shall investigate!
Tim.
Joachim Stolze
: : last night it came to me very clearly: a HP48GX successor (and I mean a
: : real successor, not of that HP38G type...) won't have a Saturn type CPU.
: : Look, the 48GX was officially announced in June 1993. Now we have 1996.
: : And I'm sure they didn't start thinking about a succesor in June 1993 but
: : before. That's about > 2.5 years. TI comes up with a 68000 model. I'm sure
: : the folks at HP are so smart to see that a 4 bit CPU running at 4 MHZ
: : with a 131x64 pixel display is nothing for the upcoming year 2000...
: : So, the next calc will surely have (at least) a 8bit CPU, will run at
: : (at least) 10MHz and will have (at least) a 240x180 pixel display...
For me, a Newton in the size of a HP48 or just a bit larger,
with the complete software of the HP-48 (At least UserRPL compatibility!)
would be just the right HP-48 successor for me.
IP-54 watertightness would be a nice addition.
Cheers,
Jojo (Joachim Stolze)
The early mornings's thinnest sliver of light appeared silently.
Several billion trillion tons of superhot exploding hydrogen nuclei
rose slowly above the horizon and managed to look small,
cold and slightly damp. (Douglas Adams, Hitch Hiker's Guide pt. 3)
Doobie
: : Stefan Wolfrum (wol...@cs.uni-bonn.de) wrote:
: For me, a Newton in the size of a HP48 or just a bit larger,
: with the complete software of the HP-48 (At least UserRPL compatibility!)
: would be just the right HP-48 successor for me.
: IP-54 watertightness would be a nice addition.
I don't know about the size of a Newton, and certainly not bigger.
Remember you want a portable system. Very portable. A Calculator has to
be able to atleast be able to fit in a pocket. If you want anything
pbigger, just get a sub notebook.
Doobie