HP Prime too complicated

[3298]

(07-13-2018 12:24 AM)Dave Britten Wrote:  Just to stir the pot a little: the Casio Algebra FX has separate "home" and CAS modes.

Yes, it does, and it's not a very good calculator (I say that having owned one).
While we're on the topic of Casio calculators, I want to say that what I've seen about the Prime feels very Casio-like, dividing the functionality into disjointed applications accessed via an icon menu.
I once read a review that nicely summarized what I dislike about that:

Quote:The Algebra FX 2.0, like Casio's other graphing calculator products, is obviously an education-oriented device. And it shows (and no, I don't mean that as a compliment.) When you first turn on the machine, you'll be presented with a bewildering set of menu options: "RUN", STAT", "RECUR", "CONIC", "EQUA", "CAS", just to name a few examples. Selecting any one of these options activates the corresponding calculator mode, where you can perform the appropriate functions.

What is missing is functional integration. This is where HP's graphing calculators continue to excel (and with the possible exception of the TI-89 and TI-92, remain in a class of their own.) The idea is simple: a multifunction calculator should let me manipulate a symbol the same as a number. It should let me perform arithmetic functions on numbers as well as vectors, matrices, algebraic expressions, indeed any object for which the function is defined. Apart from convenience, there's a significant pedagogical aspect to this: the beauty of mathematics is not that it consists of dozens of unrelated topics, but that everything is interrelated.

In this respect, the Algebra FX 2.0 is better than most of its predecessors; for instance, complex number support is nicely integrated with the rest of the calculator's functions. But it's still nowhere close to what the HP-48 family or the TI-89 have to offer in terms of feature integration.

Source: rskey.org - I'm sure a bunch of MoHPC members know that site. Note that this review was written long before the Prime existed, so "HP's graphing calculators" obviously refers mostly to the HP 48 series.

For the record, I used a Casio fx-9750G Plus in class, which was the calculator the school recommended. That thing was my first contact with programming, but on this calculator I started to feel the limits of the language and system after about a week of messing around like any heavily math-interested pupil would. Granted, I treated the manual like a novel, reading it from page 1 to the end twice and soaking up all the information my brain could process. That's probably not quite normal behavior.
The Casio Algebra FX 2.0 Plus (its family, consisting of the FX 1.0, FX 1.0 Plus, Algebra FX 2.0, and Algebra FX 2.0 Plus, is often abbreviated as AFX in the Casio community) was meant as an upgrade, but after fixing up the used and dead-on-arrival calculator (the fix didn't last much longer than a year, though) I found that it wasn't much better. For instance, when I actually wanted the CAS to do the heavy lifting for me, it returned my input unchanged. The CAS also isn't accessible from programs at all. Apart from mostly cosmetic changes (one of which was really annoying: in the restructured function key menus the => token, used for compact If-Then structures, was missing, requiring a PC with link cable to transfer it and then copy it around from one program to the next), the disappointing CAS, and the add-in system, it was pretty much the same device as the 9750.
The AFX did introduce me to C and x86 ASM, but even that is only due to other community members having hacked their way into the add-in functionality; officially only Casio could write add-ins.
When I wanted a proper calculator as a university companion, finding the parts of a dead HP 48G inspired me to get a 50G, which eventually brought me here. Like the reviewer wrote, the HP 48 series gives you one place (the stack view) from which you access everything, allowing you to run a downloaded program, perform some arithmetic operations on its result, plug the results of that into the polynomial solver, and finally draw a graph to visualize the solutions without any mode switches inbetween. On the Casio graphing calculators you'd need to visit three to four different menus (RUN can run the program, so visiting PROG is not strictly necessary; after that, RUN for the arithmetic operations, EQUA for the solver, GRAPH for the visualization), and you'll probably need to write down or remember a few intermediate results because you cannot pull them over to the other menu while switching.

Granted, the Prime does many things better than the Casio line (multi-character variable names? Proper character string support? The list goes on), but the beauty of having all the tools you need at your disposal is missing. Even leaving aside the travesty of crippling RPN to have no access to CAS and to not be programmable, the Prime's design feels inferior to the 50G due to the segregated applications.
Yes, it's a capable calculator, but the 50G does roughly the same without forcing an artificial separation of concepts on me. What would have been more appealing to me than the Prime is a 50G with a bigger screen (I don't care for touchscreens, but I'll take color if you offer it and the contrast is good enough), a better CPU, and some improvements to the keyboard layout (get rid of the next-to-useless APPS and HIST keys; give us DEL back as non-shifted key; move + - * / back down where they belong; get the Enter key a double-width center position; keep the letters away from keys that can be important in Alpha mode like the / key, probably by adding another row or column next to the cursor keys; but for god's sake, don't reduce the amount of shift keys like on the Prime, and obviously don't add even more useless keys).

... I'd like to apologize for the fact that this post turned into another Prime-vs-50G comparison, but by comparing I can highlight what I feel is wrong. Oh, and as a disclaimer: I don't own a Prime, but I think that's not a requirement for having an opinion about it (based on information collected by reading a wide range of discussions).

tl;dr: "the beauty of mathematics is not that it consists of dozens of unrelated topics, but that everything is interrelated." Like the Casios, and unlike the HP 48 series, the Prime forgets this.

[parisse]

I disagree with this comment, at least in CAS mode, you have access to "all maths and interrelations". And you have access to more advanced maths than on the 50G.
For example, you can define a non prime finite field on the Prime and make computation in this finite field. E.g. GF(2,8) will define a field with 256 elements and will return a generator of the field, by default g, then you can work with elements of the field (0 and powers of g), polynomials with coefficients in the field, matrices with coefficients in the field, all interrelated. For example a:=ranm(2,2,g) will define a matrix with random coefficients in the field, b:=1/a compute the inverse, you can check a*b.

[Voldemar]

In CAS view one row and one column matrices can be multiplied [3]*[5]=15. In Home view can not be multiplied. Why? Looked at HP 50g, there also can not be multiplied.
Numworks online simulator can to multiply one row and one column matrices [3]*[5]=[15].

[Voldemar]

Found. In Home view should be double square brackets [[5]] Enter [[3]] Enter * [[15]]. On HP 50g same.

[grsbanks]

Yup. Single square brackets are vectors. Matrices are double square brackets (one to contain the matrix as a whole and one around each row of values).

[cyrille de brébisson]

Hello,

Serial coms are "easy". Send bytes, receive bytes...

USB is complex!

Above the low level packetized send/receive there are various layers of protocol that have to be handled by the OS. HID, MSD... being the most basic examples.

On top of this is the "user level" data exchange, which is packetized and COULD be accessed by/through user programs.

The problem is really in that OS layer. We could add some user functions to send/reveice through USB, but only for the HID protocol. And this would probably limit significantly the actuall use of the functions. This is why it has kind of fallen by the way side...

Especially as most data aquisition systems have quite complex protocols, come with their own PC software to handle all of this and therefore would be 1: a LOT of work for the user to implement in Prime while at the same time not being as good as the PC version...

Now, if one of you here is working for a company that makes/use such a device and have a real use for such USB functions (more than just: it would be cool to show that you can do it), then please do PM me and I can see if something could be arranged...

Cyrille

[DrD]

(07-16-2018 06:53 AM)cyrille de brébisson Wrote:  Hello,

Serial coms are "easy". Send bytes, receive bytes...

USB is complex!



Now, if one of you here is working for a company that makes/use such a device and have a real use for such USB functions (more than just: it would be cool to show that you can do it), then please do PM me and I can see if something could be arranged...

Cyrille

USB to serial converters exist, so if you could get an interface like that to work, you would be well along the way to success with the comm's.

I'm retired now, but I used the hp48 and hp 50 products to validate installation, calibration, and maintenance of quite a variety of data acquisition products used in the natural gas utility industry.

You're right, in that manufacturers have their own software, and licensed protocols as drivers for the core product. When I retired, most of the software was built around modbus, and some earlier products used a simple checksum on the end of an ascii command or return message. While not 'easy' necessarily, it wasn't too difficult to create a tool box with native modbus commands, or the ascii equivalents, to interrogate the various hardware.

I was quite the guy at our local distribution utility, because I could roam the system and respond to any field problem with this (hp calc) hand held terminal arrangement. It made the service for me, and the benefit to the utility, quite valuable. Laptop pc's are too big to handle, especially in inclement weather on a noisy pipeline, when the pressure is on to get telemetry, like at a strategically important, natural gas custody transfer station, in the middle of a dark night, back on line. I have had many such experiences over my career there.

In fact, I was just asked, recently, by a new manager in the engineering department, if I would be willing to come back as a contractor to fix some of the pre-existing equipment. I still have my hp50, the software I used when I retired, but I'm retired ... and it's a good thing!

So it's not about the system software provided by the manufacturer. That is something that should not be messed with in field service, the need is for a compact gadget that can be readily accessed, quickly, in confined spaces, in less than ideal environments. The prime would be sooo much better for this than the 48/50 calcs, if only ...

Much of the telemetry used radio based communication. Most of the radios used an hht for programming the radio: frequency, power, deviation, levels, etc. Yes, I used the hp calcs for that, too! I had the manufacturers terminal, but I liked my one tool hp calc approach even better.

-Dale-

[Dave Britten]

(07-13-2018 12:05 PM)3298 Wrote:  

(07-13-2018 12:24 AM)Dave Britten Wrote:  Just to stir the pot a little: the Casio Algebra FX has separate "home" and CAS modes.

Yes, it does, and it's not a very good calculator (I say that having owned one).
While we're on the topic of Casio calculators, I want to say that what I've seen about the Prime feels very Casio-like, dividing the functionality into disjointed applications accessed via an icon menu.

Yeah, the AFX isn't award winning or anything; it's got the typical Casio problem: a lot of functions separated into different modes with no interoperability between them. (I still can't figure out if there's a way to simultaneously graph functions and stat plots on a Casio.) Also, how have they managed to go 30+ years without giving their calculators a "print" command that doesn't pause and wait for you to press EXE?

The interesting thing about the AFX is that it runs on an 8086-compatible CPU, and is actually using something like Datalight ROM DOS for the OS. Technically speaking, it's an MS-DOS-compatible calculator, though none of that is ever exposed to the user.

The Prime falls prey to this kind of design to some degree. At least with the Prime, you can still access functions/data outside the current application, much like you would access files outside your current directory on Windows/Linux/etc.

[ijabbott]

(07-16-2018 10:33 AM)DrD Wrote:  

(07-16-2018 06:53 AM)cyrille de brébisson Wrote:  Hello,

Serial coms are "easy". Send bytes, receive bytes...

USB is complex!



Now, if one of you here is working for a company that makes/use such a device and have a real use for such USB functions (more than just: it would be cool to show that you can do it), then please do PM me and I can see if something could be arranged...

Cyrille

USB to serial converters exist, so if you could get an interface like that to work, you would be well along the way to success with the comm's.

Even for USB serial devices, there are quite a few different types (FTDI, Prolific, Silicon Labs, WInChipHead, to name a few) each with their own protocols for embedding data, line control and status information into USB packets. There are some devices that use standard USB serial device protocols such as USB CDC to appear as some sort of "modem", but the common ones I listed don't. It all gets very complicated!

[parisse]

(07-16-2018 11:21 AM)Dave Britten Wrote:  (I still can't figure out if there's a way to simultaneously graph functions and stat plots on a Casio.)

For the record, on the Prime, in CAS:

Code:


l:=randvector(1000,normald,0,1):;
histogram(l,-4,0.25);plotfunc(normald(x),x)

[Martin Hepperle]

(07-16-2018 01:39 PM)ijabbott Wrote:  ...
Even for USB serial devices, there are quite a few different types (FTDI, Prolific, Silicon Labs, WInChipHead, to name a few) each with their own protocols for embedding data, line control and status information into USB packets. There are some devices that use standard USB serial device protocols such as USB CDC to appear as some sort of "modem", but the common ones I listed don't. It all gets very complicated!

My 5 cents:

Let's look at the success of Arduino and similar platforms which are also used in education. This ecosystem provides a huge number of sensors and devices.
Projects typically consist of the microcontroller board which is combined with one or more external sensors and output devices.
In the majority of projects the following types of Interfaces are used:

• serial at 5V or 3.3 V level ( not true RS232) for e.g. GPS sensors or links to other computers,
• the I2C bus (TWI) to sensors and displays,
• only a few use limited USB CDC capabilities to mimic keyboards and mice.
  

A "simple" USB to TTL converter cable and associated read/write commands would enable many applications e.g. in the education world. Optionally one could also think about combining this with an I2C Interface (Tony Duell built one many years ago for the HP-48). But this would already become complex.
Such an simple serial interface could be tailored to be connected to an Arduino or similar which could then translate and forward commands (sent via serial) to e.g. I2C or other systems. A small library could be created to encourage people to tinker, using the Prime as a controller, doing calculations and displaying results. This would replace keyboard and display on the Arduino.

The old Fourier sensor thing was probably o.k., but expensive. Today it seems to be outdated and not really suitable for tinkering by kids. Maybe something to discuss with physics teachers etc. I have no idea whether this is really a market.

As this interface would contain a single USB chip (e.g. Prolific or whatever), the firmware in the Prime would only need to be adapted to talk to this chip. Not a generic USB interface with drivers for lots of different USB chips.

- rant mode on -
In general USB seems to be rather volatile technology - it has burnt many people due to e.g. Windows driver policy "arbitrary" dropping support for certain USB chips, manufacturers dropping support for older USB devices. How many different USB cables do you own (A,B,C, mini, micro, power, data, ...)?
The plain old serial (TTL or RS232) interfaces are independent of such "modern technology" and astonishingly resilient. Many POS printers still use RS232 and talk ESC/p - surviving since the 1980s.
- rant mode off -

Martin

[Jean-Baptiste Boric]

(07-16-2018 02:24 PM)Martin Hepperle Wrote:  As this interface would contain a single USB chip (e.g. Prolific or whatever), the firmware in the Prime would only need to be adapted to talk to this chip. Not a generic USB interface with drivers for lots of different USB chips.

Well, there's the TI-Innovator Hub...

On a much more serious note, there's a reason cheap, dedicated, MCU-based platforms are great at tinkering with GPIO/I2C/SPI/UART stuff. The NumWorks calculator happens to be somewhat better suited for this specific task than the HP Prime (I modded mine to expose UART/SWD/SPI pins of the STM32F4) because the platform's openness and tinker-friendly architecture at least makes this possible for an enthusiastic individual without actually requiring the manufacturer's cooperation or help.

However, just because I can mod my calculator into something it's not doesn't mean this would be a substitute for a proper, cheap MCU evaluation board. The STM32F429I-DISC1 evaluation board, which is the closest equivalent to a NumWorks architecture-wise, is €50 cheaper while still having a screen, better specs and a metric ton of I/O. If you prefer an easy-to-use platform then grab a PyBoard and program the thing in Python.

Why do people want to turn a calculator into an evaluation board or affix one to it (beyond the because I can mindset) if it'll make a really expensive, proprietary and lousy one overall? I'm not saying a calculator could not in theory be useful as an active part for experiments, but beyond very basic lab scenarios one would want the real deal anyways.

[cyrille de brébisson]

Hello,

About USB and serial.
As far as I Can tell, serial is going away as an external data path.
Serial is still used INSIDE systems, to comunicate (like SPI and I²C), but it has been a very long time since I have seen anything with an external serial com (exception for old systems of course)...

USB-Serial convertors are usually crap. Not only do they nearly all require specific drivers, but in addition, they never work well. And the reason is that USB is a milisecond, packet based system while serial is a 1/10th of a ms byte based system.

As a result, on top of serial, most systems will implement some type of packet based system (xModem for example), and this system will require ack bytes to be sent back within some time frame.

Put it all together and you get this:
On the PC, the "sender" sends a bunch of bits doing FileWrite calls.
The converter driver will wait until it has a full packet, or for a given amount of time if not and send the bytes (note the first delay there).
The receiver will reply and send bytes back to the converter.
It, itself will wait for a full packet or a delay before sending a full USB packet.
All these delays cause, at best the serial protocols that are riding on top of the converter to slow down, or even to stop working due to timeouts (depends on the protocol and how good/bad the serial converter an it's drivers are)...

I actually had to change the timeouts in the latest versions of the 49 xmodem code (at the time) to try to mitigate the problem with the USB convertors that were comming up at the time.



I like HW tinkering and, yes, it woudl be great to be able to do this on something like Prime...
However, HW tinkering can easely lead to HW failures, and then guess who would see the units comming back under waranty :-( So, if anything ever got done, it would most likey be done through a separate HW board that connects to Prime through USB or the like. With issues similar to the one that I just described.

Cyrille

[sasa]

(07-17-2018 05:41 AM)cyrille de brébisson Wrote:  USB-Serial convertors are usually crap.

BTW, there is many fake chips around in USB to Serial convertors. However, any convertor which use genuine FTDI chip worth every cent. The speed is up to 1Mbit (or more).

Of course, if precise timing is required, such use may be difficult or impossible. One of such is programming some MCUs through serial port.

[DrD]

Cyrille,

USB is important, but to ignore serial comm is to ignore the many, many devices that are STILL in use today, that require serial communication. The industrial market had many programmable products, (designed for field applications), that would become a significant capital re-investment to deploy newer USB technology, or WIFI, BT, etc. Every time a new flavor of communication comes along, a gradual phase-in or compatibility association must be worked out, if there is any hope of making a realistic transition.

Every manufacturer wants to build the ultimate mandated lock, over which it is the sole provider of the key. Industry at large, uses products from a lot of different vendors. Each product has to be integrated into the corporate mainstream SCADA system. So there are two major components that must be accommodated:

1. SCADA compatibiity (Host side)
2. Field services: Installation, callibrations, maintenance. (Client side)

There is quite a market for (2. Field services), when you consider ALL of the various devices likely to be encountered. At the utility where I did this sort of thing, we probably had at least a hundred different products that required some sort of field service. Until they were no longer serviceable, meaning no parts, and no mfr support was available AT ALL, many times they are kept in use. To see 1980 technology still in use today, alongside new products just acquired, happens all the time.

What you don't see out there, is something that bridges the gap. Every time the protocol changes, there is an expectation that the corporate world will just jump and dump in favor of the new equipment. I doubt that happens very often! This is a market opportunity, not a sole reason for building this tool. That is why I see some sort of I/O as a valuable addition to the Prime. There probably aren't that many people that would actually be able to create ppl programs, using the I/O commands, but as the need begins to be resolved with some similar approach, there is much to be gained. Some enterprising entrepreneur may find reward in this pursuit!

-Dale-

[grsbanks]

(07-17-2018 10:14 AM)DrD Wrote:  Every time a new flavor of communication comes along, a gradual phase-in or compatibility association must be worked out, if there is any hope of making a realistic transition.

I see your point and tend to agree with you. However, playing devil's advocate here, how "gradual" do you think the transition phase has to be?

USB has been around for what... 20+ years now and serial ports on computers have been phased out over the same period. Just how long does industry need to adapt?

[DrD]

Advocating for the devil? I'm not sure the devil needs an advocate!

How gradual should the transition phase be? You are looking at the wrong unit of measurement, here. The correct measurement unit is not time: rather when will the return value become less than the cost of replacement technology? (i.e., return on investment).

By way of my example, a utility using 1980's technology 24/7/365 and requires very little maintenance. The ROI is still worth it 30+ years later!

By way of hp calcs, when should you throw away your hp 48, 50, or whatever?

If the hp prime could support I/O, with the addition of a few commands, the enhanced value of the prime, over previous products, would make the ROI for this case, worth it.

-Dale-

[grsbanks]

(07-17-2018 10:57 AM)DrD Wrote:  By way of hp calcs, when should you throw away your hp 48, 50, or whatever?

That's a fair point but looking at it from a business point of view, I hold on to my 48GX, 41C, 42S, Voyagers and six 50g's for the wrong reason: nostalgia. There are some 140 calculators in my collection. Do I need them all for my job and hobbies? No. Is there anything that I would use a calculator for that can't be done by a laptop? Not really. When will I get rid of them? When my dead body has gone cold.

[Aries]

(07-11-2018 02:37 PM)JSBach Wrote:  

(07-11-2018 02:14 PM)Voldemar Wrote:  I do not think so. HP Prime is easy to master. I think HP 50g is more complicated.
P.S. I also have a daughter. She is 13 years old. And she has Casio fx-991.

Yes, it is simpler than the 50g.
But the 50g isn't a pupil's calculator - it is an engineer's calculator.

JSBach

:-)

I'm real comfortable with the Nspire CX CAS, not with the Prime.
Best,

Aries