Hypothetical WP 35S

[walter b]

(01-19-2016 04:04 AM)Paul Dale Wrote:  

(01-16-2016 11:25 AM)walter b Wrote:  And it will not include more arcane math unless there are very (!) good reasons.

Why?

d:-?

[damaltor]

beautiful.

[Joseph_21sv]

(01-19-2016 09:41 AM)walter b Wrote:  A picture tells a thousand words:



d:-)

That is an improvement over the 34S, but it still looks like copying it on another platform. Why not just dispense with emulating a seven-segment display and have a pure alphanumeric display like the HP-35s?

[walter b]

(01-19-2016 03:25 AM)Joseph_21sv Wrote:  

1. Completely
2. Perhaps Casio may license the Classwiz design
3. • Better alphanumeric display
   • Softkeys
   • Easy access to external connection port
   • Memory
   • Colored display segments/Graphical display (...)
     
4. ... any variant of the HP-17bii has the advantage that it can show softkeys and alphanumeric display simultaneously and the full two-line display of that calculator fully advances over that of the WP 34s. Moreover, a DM-35L, if it is to be programmable, just makes no earthly sense to me as a one-line calculator because the programmable HP 35s has a two-line display. But the question there is: Will Swiss Micros ever produce a two-line calculator?
   Also, what exactly is the meaning of "more arcane math"?
   

Thanks for your list. Seems you may get almost everything you request but the colored display segments.

Regarding your point 4: I admit I simply called the project 35S since its 34S + 1. No thoughts at all about HP's 35S. For the display, please see my previous post of today. And finally about math: the WP 34S contains some mathematical functions no other pocket calculator contained so far, AFAIK (e.g. Gudermannian, Lambert's W, regularized Beta, etc.); this is a subset of its function set I would not extend without need - or (as another member wrote): we don't need a canned encyclopedia. Hope this explains the matter.

d:-)

BTW, thanks for quoting my post entirely. But that's really not necessary.

Edit: Regarding your point 2: Looking at its dimensions (165.5x77x11 mm), the Casio Classwiz belongs to the battleship class of calcs IMHO. Too large for my shirt pocket.

[walter b]

(01-19-2016 03:23 PM)Joseph_21sv Wrote:  Why not just dispense with emulating a seven-segment display and have a pure alphanumeric display like the HP-35s?

We will have a pure alphanumeric display. Though within 132x16 pixels, the digit 8 (in a 5x9 matrix) will hardly look any different than shown. Feel free to propose a better digit design. Looking forward to it,

d:-)

[Joseph_21sv]

(01-19-2016 03:45 PM)walter b Wrote:  

(01-19-2016 03:23 PM)Joseph_21sv Wrote:  Why not just dispense with emulating a seven-segment display and have a pure alphanumeric display like the HP-35s?

We will have a pure alphanumeric display. Though within 132x16 pixels, the digit 8 (in a 5x9 matrix) will hardly look any different than shown. Feel free to propose a better digit design. Looking forward to it,

d:-)

The advantage of 5x9 dot matrices over a 7 segment display is that they can display alphanumeric characters more or less exactly as they appear typeset.

[Joseph_21sv]

(01-19-2016 03:40 PM)walter b Wrote:  

(01-19-2016 03:25 AM)Joseph_21sv Wrote:  

1. Completely
2. Perhaps Casio may license the Classwiz design
3. • Better alphanumeric display
   • Softkeys
   • Easy access to external connection port
   • Memory
   • Colored display segments/Graphical display (...)
     
4. ... any variant of the HP-17bii has the advantage that it can show softkeys and alphanumeric display simultaneously and the full two-line display of that calculator fully advances over that of the WP 34s. Moreover, a DM-35L, if it is to be programmable, just makes no earthly sense to me as a one-line calculator because the programmable HP 35s has a two-line display. But the question there is: Will Swiss Micros ever produce a two-line calculator?
   Also, what exactly is the meaning of "more arcane math"?
   

Thanks for your list. Seems you may get almost everything you request but the colored display segments.

Regarding your point 4: I admit I simply called the project 35S since its 34S + 1. No thoughts at all about HP's 35S. For the display, please see my previous post of today. And finally about math: the WP 34S contains some mathematical functions no other pocket calculator contained so far, AFAIK (e.g. Gudermannian, Lambert's W, regularized Beta, etc.); this is a subset of its function set I would not extend without need - or (as another member wrote): we don't need a canned encyclopedia. Hope this explains the matter.

d:-)

BTW, thanks for quoting my post entirely. But that's really not necessary.

Edit: Regarding your point 2: Looking at its dimensions (165.5x77x11 mm), the Casio Classwiz belongs to the battleship class of calcs IMHO. Too large for my shirt pocket.

Regarding the dimensions of the Casio Classwiz, there are shirt pockets large enough for it, if you do not care about being unable to close them, and it still fits without ripping them. Too thin to be a full battleship.

[walter b]

(01-20-2016 01:12 AM)Joseph_21sv Wrote:  

(01-19-2016 03:45 PM)walter b Wrote:  ... the digit 8 (in a 5x9 matrix) will hardly look any different than shown.

The advantage of 5x9 dot matrices over a 7 segment display is that they can display alphanumeric characters more or less exactly as they appear typeset.

Certainly, Sir. So how looks your 8, Sir?

d:-?

[Paul Dale]

(01-19-2016 03:40 PM)walter b Wrote:  regularized Beta

If you want the statistical distributions, you'll have this one.


- Pauli

[walter b]

(01-20-2016 05:21 AM)Paul Dale Wrote:  

(01-19-2016 03:40 PM)walter b Wrote:  regularized Beta

If you want the statistical distributions, you'll have this one.

No real problem - I just see the future X.FCN flooded by such stuff. We don't have to show everything ...

d:-)

[timofonic]

(01-18-2016 01:41 PM)Luigi Vampa Wrote:  

(01-18-2016 03:18 AM)timofonic Wrote:  Crazy idea: What about merging with newRPL project too?

RPL has its own fan club.
An RPL-programmed scientific calculator always includes an RPN stack. Nevertheless, an RPN key-stroke-programmed scientifc calculator doesn't necessarily need to include RPL. I am not aware of any such model.
That being said, I must say I do enjoy RPL programing in my HP28s & HP48sx, but WP34s, and prospect WP35s project, seem to have a very strong base of 'pure-RPN' supporters.

BR MfG +
Don Luigi

It's probably due to my ignorance, but I fail to see why both "worlds" can't be merged.

Can't RPL be used as a scripting system to add features instead code in C/C++ all time?

This would make the calculator more expandable and need less harder ti write code. Maybe the code could be compiled to make it run faster, or use some kind of Android's ART approach.

- Of course, it seems there are people allergic to LISP and derivatives.

+ But seems math nerds love it: Scheme.
* Functional programming language Haskell got influence from LISP.


There's a weird LISP dialect called Clojure that is gaining lots of popularity and used even by big corps, but it's some kind of slave language for JVM. I fail to understand the reasoning about why people are so attached to a programming language owned by Oracle, a company with a very predatory and shady reputation in a similar or worse way than Microsoft.

GoferLists (Advanced List Operations) got inspired by some Haskell libraries too.

(01-19-2016 03:45 PM)walter b Wrote:  Seems you may get almost everything you request but the colored display segments.

We will have a pure alphanumeric display. Though within 132x16 pixels, the digit 8 (in a 5x9 matrix) will hardly look any different than shown. Feel free to propose a better digit design. Looking forward to it,

Well, these days I have a FX991EX and consider that screen the minimal for a new product. Maybe I'm a snob or maybe my eyes get tired easily for different reasons ( far-sightedness + astigmatism, ADHD, using computers and soldering all day, etc)

I got tired of finding if someone did the BoM of the calculator, but it features a 192×63 dot matrix display.

Segmented-based LCD displays

I liked this historical article aboutcalculator display technology

I liked this an article comparing different display technologies(2). I did the maths and clearly EPD is the winner! I'm not sure about the different case scenarios and low volume availability , I didn't research enough for that.
JDI Announces Start of Sales of a Standard Line-up of Ultra-low Power Consumption Memory-in-Pixel Reflective-type Color LCD Modules

There's an interesting comparison of nextgen display tech at Wikipedia.

Is EPD fast-enough for a calculator?

+ It's cool to make your DIY crap and easy for novices to use them.
- Very limited graphic capabilities: What about menus or representing different symbols?
- There's issues about visibility and confusing symbols, too.

Interesting lecture: [URL=]Don’t use 7-segment displays, by Harold Thimbleby[/url]. FIT Lab — Interaction Laboratory, Swansea University
Swansea, Wales

I fail too see why segmented displays are still used:

There's great technologies that allow a lot more of definition with very low power: eINK is one great example

+ This would provide the possibility to provide graphing calculator capabilities.
*I know this project is aimed at old school style and such, that's "okay".

+ Menus would be able to be more descriptive and easy if properly designed.
+ It can be useful for more uses other than a calculator, but somewhat of a long pocket computer:
* Data acquisition/logging and visualization for research purposes.
* Symbols and numbers could be rendered in a more visually appealing way, bigger for people with vision issues or some extreme conditions.

(01-20-2016 05:21 AM)Paul Dale Wrote:  

(01-19-2016 03:40 PM)walter b Wrote:  regularized Beta

If you want the statistical distributions, you'll have this one.

- Pauli

(01-19-2016 03:40 PM)walter b Wrote:  And finally about math: the WP 34S contains some mathematical functions no other pocket calculator contained so far, AFAIK (e.g. Gudermannian, Lambert's W, regularized Beta, etc.); this is a subset of its function set I would not extend without need - or (as another member wrote): we don't need a canned encyclopedia. Hope this explains the matter.

Regarding your point 2: Looking at its dimensions (165.5x77x11 mm), the Casio Classwiz belongs to the battleship class of calcs IMHO. Too large for my shirt pocket.

Excuse me.

For us the math newbies: What do you consider to be "arcane math"?

Why aren't you going to implement it? What are the possible challenges?

What's so wrong about having a "canned encyclopaedia"? Do you know Project Euler What about using the calculator as a Inductive Chain Learning tool in order to motivate math learning by game-like software?

2. FERNÁNDEZ, María Rodríguez; CASANOVA, Eduardo Zalama; ALONSO, Ignacio González. Review of Display Technologies Focusing on Power Consumption. Sustainability, 2015, vol. 7, no 8, p. 10854-10875.

[quantalume]

(01-20-2016 02:58 PM)timofonic Wrote:  There's a weird LISP dialect called Clojure that is gaining lots of popularity and used even by big corps, but it's some kind of slave language for JVM. I fail to understand the reasoning about why people are so attached to a programming language owned by Oracle, a company with a very predatory and shady reputation in a similar or worse way than Microsoft.

Off topic, but you might want to check out Hy. There's even a kernel for the fabulous Jupyter framework.

[walter b]

@timofonic: Thanks for your contribution containing a lot of points and links which will take quite some time for checking and reading. Just allow me writing a few quick general remarks (based on my personal, mainly industrial experience) in arbitrary order:

• One of the best and most certain ways to kill a project is feature overload.
• If you want to be successful, make small steps and change one thing at a time.
  

Thus, what we will be trying with the "35S" as defined in the OP:

1. We'd like to use an existing hardware platform as is but, alas, there's none suiting our requirements AFAWK.
2. We will use an existing hardware platform as far as possible. Hence, we are not keen modifying anything unless inevitable.
3. We will use as much of the opportunities the platform offers.
4. We're not up to build a frankencalc. Hence, no manual "minor" modifications of existing mechanics. Either it fits or it doesn't.
5. We will use as much of the existing firmware as possible. Hence, no RPL.
6. We will make the UI as comfortable as possible for as many people as possible under these boundary conditions. Some complain about deep menus, others complain about their favorite functions excluded - we've to find a compromise.
   

Anybody wanting something significantly different, feel free to convince us (or start your own project).

d:-)

[d b]

my four cents:

I don't like deep menues either but i can get used to them. If there are user assignments then those deep menues aren't much of an issue anyway.

Most people here doing math for a living assemble at least 4k of user programs. Some write many k more. Think Namir here. He wakes up in program mode. I put 6k on a 42s and wanted more than the 2k left for data storage. So four times what the 34s was miraculously able to deliver in fully merged user memory would be the minimum hoped for. I (think that I) realize that it's possible to attach user programs as key assignments in that underused portion of memory of the 34s but easier is always better, for us mortals.

Easy data storage / program storage / firmware updateability through USB, or anything normal, is indespensable. HP's long lamentable history of proprietary cables were not a nice thing to saddle us with.

If the operating system shares space with the merged user memory and IF that is limited; the ability to easily delete those functions in the firmware that aren't "favorites" to create more space would be a usefull feature.

[vido]

Swissmicros is very active in producing machines, I think it would be extraordinary if there was a kind of collaboration between you and them, in order to have a custom hardware with a mature software.

[anetzer]

(01-21-2016 05:14 AM)vido Wrote:  Swissmicros is very active in producing machines, I think it would be extraordinary if there was a kind of collaboration between you and them, in order to have a custom hardware with a mature software.

I too am wondering, given the rather prominent interest in the WP 43S, whether SM would not be willing to go a few steps in that direction.

If there is a market for full size voyager clones there surely also is one for a model that is oriented to WP 43S specifications.These should be definable I guess? Plus that model could be fitted with an HP 42S or HP-41 emulator by other people who know how to implement these on a defined hardware (producing a base model with different key cap labels after all is something that SM has practice with so far).

The million dollar question is: How many of these would have to be ordered to get SM into designing appropriate hardware in collaboration with the WP 43S - team and what would they cost? Given the kind of business they are in they should be able and possibly willing to give cost estimates for different numbers of sold pieces.

I guess that most of the forum members put enough confidence in the team and if enough of us commit, alternative emulators might only be a question of time as well.

I mean, there's no strict need for SM to only produce Voyager replicas with HP ROMs...

[walter b]

Our Pixel Art design studios didn't join my "vacation":



d:-)

[walter b]

And our layouters were busy as well - assuming said hypothetical calc would feature data types then it might look like this:



d:-)

[BarryMead]

Will the RAM/FLASH sizes included in the Swiss Micros platform support a calculator with all of these functions? I was under the impression that the CPU inside the Swiss Micros calculator was even less capable than the one in the HP-30B. If the CPU can handle all of that firmware, then this is a great solution to the hardware vacuum.

If not then it is just wishful dreaming.

Does anyone know the RAM/FLASH size of the currently used CPU in the Swiss Micros family of calculators? Are there any compatible chips in the same CPU family that could easily be substituted into the Swiss Micros design that would increase the RAM/FLASH space?

[Steve Simpkin]

The LPC1115 ARM micro-controller used in the DM-41L, for example, only has 64KB of Flash and 8KB or RAM.
The Cortex-M4 type controller used in the upcoming DM-42L has much more memory. The NXP LPC54102 Cortex-M4 controller (as an example) has 512KB of Flash and 104KB of RAM. That would not only work for Free42 but it likely be a good candidate for a WP34s port and the WP43s project. The only real drawback (for some) would be the landscape layout.