DIY543 pictures



#30

Here are some quickly done pictures of one of the first prototypes done by Eric Smith and myself.


#31

I don't see any pictures, just some kind of small icon.


#32

You are right. Obviously I did something wrong. The icons do link to the pictures I posted, however.

Anyone know what I did wrong? I meant to "embed' the pictures in the post.


-- Richard


#33

It seems you didn't use the right URLs :



Nice pictures, this project seems to move closer to a product for future users :-)


Edited: 20 Oct 2013, 7:36 p.m. after one or more responses were posted


#34

Wow, that looks like a machine I could really go for. Where do I send my downpayment :-)?

One thing though: I'd really like to see an infinite stack, like the 50g etc. It's one of the few things I miss about my dumped 49g+, and it seems to me to make no sense to restrict ourselves to the same size stack as was being forced on us 40 years ago.


#35

The goal of this project has been to produce a high end "traditional RPN" calculator, since there hasn't been one on the market since November 1990, while HP is still (for now?) selling the 50g. We recognize that not everyone wants "traditional RPN", but if we try to please everyone we will end up pleasing no one. I think part of the reason that various community efforts to produce a new calculator product for sale have not been successful has been lack of focus and excessively ambitious feature plans. I'm not criticizing anyone or anything in particular; I think even my own calculator projects and plans have had those problems in the past, despite my own awareness of that problem.

It seems likely that the firmware for the DIY543 will have a selectable 8-level stack option like WP-34S, but I don't think that arbitrary/infinite depth is likely. While I'm a firmware developer for the calculator, my contribution will mostly be low level hardware-related code, so I'm not going to make promises about the user interface.

The DIY543 hardware could certainly support RPL or RPL-like calculator firmware, but no one has offered to write such a thing.

The first production hardware based on DIY543 is planned to have key legends based on what has been seen in these forums previously. If you want alternate legends for modified or alternative firmware, it seems like the best approach would be something similar to the adhesive key stickers that Eric Rechlin makes for the WP-34S.


#36

Looks fine to me; very 42-like. It looks as though all the programming-related functionality (LBL, tests, etc.) is accessed via menus, and I dare say RPL-type functionality could be accessed exactly the same way, with no need for keyboard changes - although RPL'ers would probably want to do something with that GTO key. ;)

It looks nice and clean - please keep it that way!

Best,

--- Les

[http://www.lesbell.com.au]


#37

We do not plan to offer a product using the Free42 firmware and keyboard layout. The Free42 code is not really written with limited-memory devices in mind, and there are memory fragmentation issues. I ported the Free42 code to this hardware only as a demonstration of the capabilities of the hardware.

I expect that the product will use the yet-to-be-written firmware that has been tentatively discussed here under the name WP-43S, and the corresponding keyboard layout. As with the WP-34S, the keyboard layout will look a lot "busier" than the HP-42S and Free42, because it is intended that more functions be available directly rather than requiring menu navigation.

Note that DIY543 is our designation for the prototype hardware. A product based on it will have a different name.


#38

Quote:
We do not plan to offer a product using the Free42 firmware and keyboard layout. The Free42 code is not really written with limited-memory devices in mind, and there are memory fragmentation issues. I ported the Free42 code to this hardware only as a demonstration of the capabilities of the hardware.

I expect that the product will use the yet-to-be-written firmware that has been tentatively discussed here under the name WP-43S, and the corresponding keyboard layout. As with the WP-34S, the keyboard layout will look a lot "busier" than the HP-42S and Free42, because it is intended that more functions be available directly rather than requiring menu navigation.

Note that DIY543 is our designation for the prototype hardware. A product based on it will have a different name.


Will the firmware be written in assembly? Or will you use higher level languages?


#39

I expect that the firmware will be written almost exclusively in C. In some low-level hardware related functions there are a small number of inline ARM assembly language instructions.

The prototype with Free42 code is a mix of C and C++, because Free42 was written in C++ to support different classes for binary vs. decimal floating point.

#40

Quote:
As with the WP-34S, the keyboard layout will look a lot "busier" than the HP-42S

Oh, darn. I find the 34S keyboard to be a bit too noisy; I can never find anything and after a while, I give up in frustration and pick up a 41 so I can get work done. At this rate, I'll never learn my way around it. :(

Best,

--- Les

[http://www.lesbell.com.au]


#41

Some design proposals were discussed here in the last 11 months, so I'd expect the keyboard featuring only two shift keys (f and g) - this will let the UI look cleaner than the WP 34S certainly.
With the Prime featuring an ENTER key on the right hand side, some opinions about the UI may have changed, however. Anyway, every layout will stay vapourware (please search OpenRPN for the meaning of that word) until the hardware will be running reliably. So all our hopes are with Eric and Richard!

d;-)


#42

Just a thought, but could the OpenRPL code base be ported to the Prime? Instant RPL support.


- Pauli


#43

OpenRPL??

d:-?


#44

OpenRPN then.

Late night, feeling stressed.


Pauli

#45

Quote:
every layout will stay vapourware [...] until the hardware will be running reliably.

For what it's worth, the hardware now seems to be quite reliable. Of course, what is needed is reliable hardware that is physically available to firmware developers, and we're working to make that happen soon.

#46

Quote:
I'd expect the keyboard featuring only two shift keys (f and g) - this will let the UI look cleaner than the WP 34S certainly.

Sounds good to me!

Best,

--- Les

[http://www.lesbell.com.au]


#47

Awe, less complicated looking, shucks. I want the front to look as damn complex as possible, so complicated in fact, that my co-workers are too intimidated to pick it up: at least three shift keys!

You should see the looks the WP34s gets along with, where's the equal key!

Edited: 22 Oct 2013, 2:19 a.m.


#48

In that case twenty-one shift keys is ideal.

- Pauli

#49

You can get a complex look with only 16 keys:


#50

Oh yes, but where's ENTER ??

d:-(

#51

I'm in awe.

I think we've found the ultimate 4-banger layout here.


- Pauli


#52

The up arrow key at the bottom doubling as the decimal entry key.

There is no equals at least.


We could also have a two key calculator and enter op-codes in binary :)


- Pauli


#53

Quote:

We could also have a two key calculator and enter op-codes in binary


Seems we've found the difference between the easiest and the simplest or cheapest UI.

d;->

#54

Quote:
We could also have a two key calculator and enter op-codes in binary :)

You can do even better: THIS Onion spoof shows how to do it with just a single dial!!


#55

Analogue input? Come on, we're digital these days.

- Pauli

#56

Quote:
I want the front to look as damn complex as possible

Right! Just for that, we're taking away your glass flight deck and putting you back in a steam-powered cockpit, complete with old-fashioned edge-reading DG. ;)

Best,

--- Les

[http://www.lesbell.com.au]


#57

13 years on the B737-200, all analog and gauges, fun but work. Auto pilot only held altitude! Descents, climbs and approaches were better with AP off as humans were smoother and some humans think :-)

Cheers, Geoff

But seriously, three shift keys to save menus, I think, is not too busy. Especially if the colours are not pale blue and pale low contrast day glow orange! IMHO!

Edited: 22 Oct 2013, 1:16 p.m.


#58

I'm not sure, Geoff, it's a tough call.

-41cv/cx: 39 keys incl. one shift and alpha

-15c: 39 keys incl two shifts

-42s: 37 keys incl one shift and menus

-WP43s: 37 keys incl. three shifts and catalogs (menus)

All of these managed to fit the trigs and logs/powers on the keyboard, generally as paired primary and shifted keys.

With two shifts (plus alpha) and the layout discussed in April, at least trigs and logs look to be heading to menus even though the 43 keys offer 4 to 6 more spots than these earlier examples. The keyboard will already be as loaded as the 50g (which more or less works, although I prefer the 35s color scheme). With a third shift (plus alpha), maybe that could be avoided?

As much as I hate certain functions being in menus-- for me-- I'd avoid a third shift. Like I said, though, it'd be a tough call.

-Bill


#59

Yes, but i grew up with the 65 and 67!

;-)

but yes a 42S 41C look is just fine.

All tongue in cheek with the multiple shift keys but still, i wouldn't mind!


#60

Quote:
Yes, but i grew up with the 65 and 67!

As did I, after the 45 - but in recent years, I realized that simplicity and orthogonality are the key to success.

What with all the yellow and blue labels, plus letters, below the keys, and white and green on them, the 34S is very noisy. The 41 keyboard is much easier to get around. For much the same reason, the 11C represents the "sweet spot" in the shirt-pocket back-of-an-envelope number-cruncher line - all the advanced functionality of the 15C involves so much memorization of arcane keyboard sequence trickery that unless you use it a lot, it's impossible to recall. It's remarkable that it can be done, but not that useful, especially in an era of computers with a much richer UI.

Best,

--- Les

[http://www.lesbell.com.au]


#61

Speaking of computers, my MacBook has 6 shift keys, not counting the duplicates. It's just missing the multiple legends...

#62

Speaking of noisy keyboards, the HP-32sii is a lot noisier than the HP-32s though that is preferred here. And the letters can be removed from the keyboard by a nice menu à la HP-42S though we had that discussion some years ago already ...

d;-?


#63

I've been giving it some thought - why do I find the 34S keyboard to be overloaded? It might be my advancing years. ;)

But I think that I find it easier to use the XEQ-ALPHA style of entry of the HP-41C because locating the letters to type a mnemonic is less tiring for the eyes than hunting around on top of, above and below the keys. Finding the right letter - before the keyboard becomes completely familiar - is a linear search operation, whereas locating a function in a key involves a completely different - and, I suspect, only partially controlled - eye movement up, down and back and forth across the keyboard. The result is more eye-strain.

Locating the same mnemonics in a menu is somewhere in between these two extremes.

Just tossing this out there for thought and discussion. . .

Best,

--- Les

[http://www.lesbell.com.au]


#64

So we should have randomised the alpha layout on the 34S :-)

I actually agree here. On my 28S (and the 48, 49 and 50 devices), I prefer to enter alpha mode and type than to navigate menus.


- Pauli


#65

I kind of like the idea of a randomized alpha layout ;-)

With the 43S, we can get something like 'XEQ [alpha] ...' as an alternative to 'CATALOG FCNS ...' or calling the right menu. Let the real world decide what's the best method to call operations.

d:-)


#66

Quote:
I kind of like the idea of a randomized alpha layout ;-)

Just be happy if people think you two are geniuses. There's no need to try for evil geniuses. ;)

Best,

--- Les

[http://www.lesbell.com.au]


#67

My children often call me evil :-)

- Pauli


#68

Know that, been there :-)

#69

Quote:
All of these managed to fit the trigs and logs/powers on the keyboard, generally as paired primary and shifted keys.

Sadly, this is not true for the 34s. Each and every math function (powers, trig, even roots and reciprocals) require at least one shift key. None of these common functions can be accessed with a single key stroke. This limitation is set by the used hardware: it simply lacks that additional eighth row of keys that provides, for instance on the 35s, so much better usability for common everyday calculations. Actually this additional row provides the most important functions (sin cos tan, sqrt, power and 1/x) unshifted. That's why I hope that the next calculator project is based on hardware with (at least) eight key rows.

Dieter


#70

Quote:
Each and every math function (powers, trig, even roots and reciprocals) require at least one shift key. None of these common functions can be accessed with a single key stroke ((on the WP 34S))

Not quite true. 1/x, y^x, and SQRT are primary if the keys B, C, and D are not used as global labels.
Quote:
That's why I hope that the next calculator project is based on hardware with (at least) eight key rows.

AFAICS, the 43S will feature eight rows of keys. One of them is used for softkeys - so we'll have seven rows available for labeling (i.e. 37 keys again). Speaking of primary functions, ten are required for digits, one for the radix mark, four for arithmetic operations, then EXIT, DOWN, UP, XEQ, DEL, EEX, +/-, ENTER, x<>y, Rv, STO, RCL, f, g, USER/[alpha], R/S, and complex i - a total of 32 locations. Trigs, powers, roots, and reciprocal compete for the remaining five locations. Good ideas beyond the published proposals are welcome.

d:-)


#71

Quote:

Trigs, powers, roots, and reciprocal compete for the remaining five locations.


That's the rub, isn't it? With the addition of a second shift and keys for complex i and USER/[alpha], the 37-key form factor (a la 42s) comes up just short. Compromises need to be made between the increased functionality and "usability for common everyday calculations".

Quote:

Good ideas beyond the published proposals are welcome.


My approach would be to load up the softkeys with the trigs and logs like the 42s and the A thru D keys on the WP34s. Powers, roots, and reciprocal would stay on the second row. The sixth softkey and the second-row key (freed by consolidating the TRIG and LOG menus into a single key acting like TOP-FCN on the 42s) become an embarrassment of riches.

Other than seeing past the softkey labels when using menus, I think this would be really slick and provide the best of all worlds. Since menu entries couldn't go on the softkeys, maybe they only get the key-front (gold?) shift. A user flag to suppress the TOP-FCN menu display could maximize screen real estate for "common everyday calculations".

That's my idea; its goodness can be judged by others.

-Bill


#72

Quote:
My approach would be to load up the softkeys with the trigs and logs like the 42s and the A thru D keys on the WP34s. Powers, roots, and reciprocal would stay on the second row. The sixth softkey and the second-row key (freed by consolidating the TRIG and LOG menus into a single key acting like TOP-FCN on the 42s) become an embarrassment of riches.

Other than seeing past the softkey labels when using menus, I think this would be really slick and provide the best of all worlds. Since menu entries couldn't go on the softkeys, maybe they only get the key-front (gold?) shift. A user flag to suppress the TOP-FCN menu display could maximize screen real estate for "common everyday calculations".


Bill,

Thanks for your suggestions. I'm uneasy I got everything correctly, however, so please try to submit a sketch of your layout. For comparison, the (preliminary) state of my thoughts is displayed here (labels underlined call softkey menus):

d:-)


#73

I like it! The only thing I would suggest is making the letters below and to the left of the keys more prominent.

Best,

--- Les

[http://www.lesbell.com.au]

#74

Quote:

Thanks for your suggestions. I'm uneasy I got everything correctly, however, so please try to submit a sketch of your layout.


Walter-

Here's what I tried to describe (presuming my hack of an image links correctly).

I don't have great ideas for the ??? keys. They could be HYP/HYP-1 (in keeping with earlier HP units and maybe eliminating a menu for them), but I'm not sure anybody actually uses hyperbolics. Maybe .d/DMS or HMS+/HMS- would be more useful. Whatever, better to have an extra key to work with than to be one short!

Similarly, with freed space on the second row, I thought a dedicated [sigma] key, parallel (||), and a primary [alpha] key might be welcome.

Probably there is room for optimization; I'm just trying to show how I thought the top row could be used to enhance "usability for common everyday calculations".

-Bill

Also, of course, "LOG" could be "LG" ;-) (I forgot and didn't want to edit my sketch.)


#75

Bill,

Thank you very much. You're the first person after a long time submitting a reasonable graphic layout :-)

One minor remark: P<->R and R<->P are identical IMHO. I guess you meant P->R and R->P instead.

Regarding TOP.FCN, we had a discussion and even a poll about 9 months ago IIRC. Must be in the archives. Most participants then prefered F1 ... F6, but times may change ...

d:-)

P.S.: Found the poll here. Enjoy!


Edited: 28 Oct 2013, 8:07 a.m.

#76

Walter-

Oh yes, I recall the poll now. If I found it earlier, I might have saved myself the trouble of making the sketch. At least I'm consistent in my preferences!

Regarding R->P and P-R, this was the result of a half-formed thought: are both keys needed, or is a single R<->P key adequate? Wouldn't the high-quality display make the current format clear enough that a toggle between rectangular and polar suffices?

I would still suggest making y^x primary instead of x^2 since x^2 is trivial in RPN. Maybe even lose cube and cube root to make room for gold-shifted .d and D.MS?

Actually, without the cube functions and conversions, trigs and logs could fit with a key left over for R<->P:

-Bill


#77

Looks nice. Putting trigs on shifted locations, however, doesn't save a single keystroke compared with a menu as presented on the 31st of January in said poll.

About x^2: the HP-45 featured a primary x^2 although it's trivial - a primary x^2 saves one stack level :-) Please see here also a shifted y^x and SQRT.

About P<->R: if you've got a number in X and another number in Y, how shall the calculator know those are polar or rectangular coordinates?? TIA for enlightenment.

d:-?


#78

Quote:
Putting trigs on shifted locations, however, doesn't save a single keystroke compared with a menu as presented on the 31st of January in said poll.


True, unless you're working in another menu that you'd need to go back to after doing your trig or between trig functions.

Quote:
About x^2: the HP-45 featured a primary x^2 although it's trivial - a primary x^2 saves one stack level :-) Please see here also a shifted y^x and SQRT.


I think x^2 made more sense when registers were at a premium (and an 8-level stack wasn't an option). I don't ever remember needing x^2 (although I've used it in quickie programs where I didn't want to think through the condition of the stack). I believe every calculator after the 15c (42s, 48/49 variants, 50g, 35s, plus the 41 variants) had SQRT as the primary. The Prime puts x^2 as primary because (paraphrasing Tim-- I can't find the thread) it's targeted to education and kids learn square before square root.

Quote:

About P<->R: if you've got a number in X and another number in Y, how shall the calculator know those are polar or rectangular coordinates?? TIA for enlightenment.


I was thinking that the complex number would be in X alone, where the calculator would necessarily need to know if it was rectangular or polar. My frame of reference is 48g which lets me enter (3,4) or (5[angle]53.13), displays the complex number according to the selected coordinate system, and changes between systems with the POLAR key. Other than the parenthesis, this seemed nicer than the 42s or 15c approaches. The 50g hides the [angle] key somewhere that I can never remember, and the 35s famously lacks the ability to easily convert between coordinate systems.

With the hard key for i and theta, I thought the 43s was going to perfect the 48/50g approach. I would envision being able to enter a complex number in either system, depending on whether I used i or [angle] as the separator. I could then operate on that argument and get my result back in the same system. Given the display's capabilities, I would be able to tell which system my number is in and toggle between systems with R<->P. Talk about saving a stack level ;-)!

-Bill


#79

Quote:
Quote:
Putting trigs on shifted locations, however, doesn't save a single keystroke compared with a menu as presented on the 31st of January in said poll.
True, unless you're working in another menu that you'd need to go back to after doing your trig or between trig functions.

Not true, since [TRIG] [sin] costs 2 keystrokes only when you're outside of that menu, while [f] [SIN] costs 2 keystrokes always.
Quote:
I think x^2 made more sense when registers were at a premium (and an 8-level stack wasn't an option). I don't ever remember needing x^2 (...). I believe every calculator after the 15c (42s, 48/49 variants, 50g, 35s, plus the 41 variants) had SQRT as the primary. The Prime puts x^2 as primary because (paraphrasing Tim-- I can't find the thread) it's targeted to education and kids learn square before square root.

When kids learn square they should do it mentally. In that age, I don't think anything like a prime is of advantage in the long run. Later, square roots are more challenging than squares. Anyway, I think the 43S will overcome that hurdle.
Quote:
Quote:
About P<->R: if you've got a number in X and another number in Y, how shall the calculator know those are polar or rectangular coordinates?? TIA for enlightenment.

I was thinking that the complex number would be in X alone, where the calculator would necessarily need to know if it was rectangular or polar. ... With the hard key for i and theta, I thought the 43s was going to perfect the 48/50g approach. I would envision being able to enter a complex number in either system, depending on whether I used i or [angle] as the separator. I could then operate on that argument and get my result back in the same system. Given the display's capabilities, I would be able to tell which system my number is in and toggle between systems with R<->P.

Sounds like a double use label to me. So far, I planned the traditional ->R and ->P for reals. Of course you are free to enter complex numbers in rectangular or polar notation, but that's a different cup of tea. RECT/POLAR is a global state IMO, so if RECT is set and you enter a complex number in polar notation it will be converted as soon as you close input. If you want to see all such numbers in polar notation, just set POLAR. KISS.

d:-)


#80

Walter-

Are you saying that one could equivalently do 3 [ENTER] 4 [->R] and 3 [i] 4 [ENTER]? Likewise, 5 [ENTER] 53.13 [->P] and 5 [angle] 53.13 [ENTER]?

This would be pretty wasteful of keyboard space wouldn't it? I can see both methods being retained for compatibility with old program routines, but only one need be directly keyboard accessible. I would suggest that [i] and [angle] stay on the keyboard since they
1. Require fewer keystrokes (on average)
2. Save a stack level
3. Provide direct (dare I say intuitive?) entry of complex numbers instead of combining two reals.

With the key spaces saved, direct conversion (instead of going into a mode menu and making a global conversion) between rectangular and polar display formats would address one of the major complaints about the 35s. Whether that's one key acting as a toggle or two keys...

-Bill


#81

I believe the intention is that 3 [ENTER] 4 [->R] will treat the 3 in stack-Y and 4 in stack-X as a polar-form complex number (i.e. 4/3 and convert it to the rectangular form with the imaginary component in stack-Y and the real component in stack-X. 3 [i] 4 [ENTER] will enter the complex number 3 + 4*i into stack-X. Walter, please correct me if I am mistaken.


#82

Yes, Jeff, I had multiple mistakes in my post, reversing both the stack order and the key functions. That's why I generally reach for a 48 when I'm doing this and double check the stack gymnastics with exactly this exercise when I'm using another calculator. I should have done that before typing. Ouch.

Still, my fundamental question stands. Why are there two keyboard-accessible ways to get a complex number into the x-register?

-Bill


#83

Jeff, Bill, all,

What is planned for the 43S so far is having

  1. the traditional 2D coordinate conversions ->R and ->P acting on reals in x and y,
  2. a global setting RECT / POLAR for complex number display.
Both are two different cups of tea as mentioned earlier (I hope). With the 43S, we'll have various types of objects (like on the HP-42S in a way). Thus, complex numbers will be entered immediately using [i] or [angle] instead of composing them from the stack. Hope this explains the background. Else don't hesitate to continue asking.

d:-)

Edited: 31 Oct 2013, 6:24 a.m.

#84

Hallo Walter,

I've "meditated" a while about your keyboard layout and I like it's very elaborated design very much. But I have a question---most likely already discussed in this forum---I'm sorry for asking: You have chosen "EEX" (HP 35 ... 41) instead of "E" (HP 42 ...). Good! Why didn't you use the good old label "CHS" (HP 35 ... 41) instead of "+/-" (HP 42 ...)?

For me "EEX" and "CHS" are a homage to HP's venerable R&D tools and, moreover, "CHS" seems from the mathematical point of view in my opinion a bit more well-founded: CHange Sign <= x*(-1).

Okay, peanuts... As long as we will get the WP 43s to see/buy as soon as possible! ;-)

Beste Grüße,

Martin

(Als Dinosaurier schreibe in meiner Diss. etwas würdevoller auch von der "Großrechenanlage" (CHS) und nicht vom Supercomputer (+/-), übrigens einem 'HP Superdome 2-32s' ...)

Edited: 1 Nov 2013, 8:59 a.m.


#85

Hallo Martin,

auch diese Frage ist seinerzeit im Forum diskutiert worden. In Kürze: +/- ist (wie alle mathematischen Symbole) international verständlich. Wir machen grundsätzlich Rechner für den weltweiten Gebrauch, nicht nur für die, die englisch denken. Deshalb |x| statt ABS, r statt CORR usw. Eine ähnlich schöne, kompakte Notation für EEX haben wir leider nicht gefunden. Ich war damals für E (wie bei den Pioneers), bin aber überstimmt worden.

d:-)

(For our unilingual readers: Some explanation about the history of the WP 34S key labels +/- and EEX.)


#86

Abstract: Belatedly I vote for 'E' instead of 'EEX' for the sake of consistency.

Hallo Walter,

herzlichen Dank für die Erläuterungen: |x| etc. ist mir vor längerem schon 'mal aufgefallen, diesmal sprang mir nur '+/-' wieder ins Auge. Offenbar ist da in mir eine "Voyager-Pioneer-Anomalie" ...

Auf Grundlage der überzeugenden Internationalisierungsphilosophie bin ich dann konsequenterweise auch unbedingt für 'E' statt 'EEX'! - Kippe ich damit das Abstimmungsergebnis? :-)

Martin


#87

Hallo Martin,

Thanks for your vote but - alas - it comes too late. Maybe we've got a chance with the 43S though I doubt it - our voters from the US of A are so conservative they still use the heap of units their colonial masters used 250 years ago.

d;-/

Edited: 1 Nov 2013, 11:37 a.m.


#88

I fear so, too.

Quote:
[...] still use the heap of units their colonial masters used 250 years ago.

Welch' Seitenhieb! Klasse!

Edited: 1 Nov 2013, 12:29 p.m.

#89

Quote:
. . . our voters from the US of A are so conservative they still use the heap of units their colonial masters used 250 years ago.

As a dispossessed colonial master, what pray is wrong with dividing the world into 8ths, 12ths, 20ths, 144ths, 1760ths etc.? Never could understand that metric concept.


#90

The metric system just matches the decimal notation better. If we had a duodecimal number system, your way of dividing the world in smaller units would certainly fit better, at least partly.

#91

Quote:


Sounds good to me!



We do walk extra miles to satisfy our potential customers.

d:-)

#92

Quote:
The DIY543 hardware could certainly support RPL or RPL-like calculator firmware, but no one has offered to write such a thing.

In a fit of insanity, I have started work on a 32-bit RPL system written in C++. So far I have the the compiler, the interpreter and about 300 entry points. I'm using Jim Donnelly's book as my design spec.

Dave


#93

Your insane!

I have some pics to send you, need emai address.

Edited: 22 Oct 2013, 2:20 a.m.

#94

Been there already. Not the same internally, but RPL did some rather clever stuff with the object prefix numbers also being (harmless) code to allow inline and indirect execution.


- Pauli

#95

So glad to see this project making progress... a breath of fresh air for us, traditional RPN folks (and a nice change from the PRIME threads).

Edited: 21 Oct 2013, 3:37 a.m.


#96

Quote:
So glad to see this project making progress... a breath of fresh air for us, traditional RPN folks (and a nice change from the PRIME threads).


Ditto!


Massimo

#97

Thanks for fixing my post.

-- Richard

#98

Is there a link where we can read up more on this project?


#99

There isn't really any information available beyond what's already been discussed in these forums, and the videos from the last few HHC conferences.

Jake Schwartz offers HHC conference videos, though the web page seems to be out of date. Eric Rechlin has also recorded videos of recent conferences and has put them on YouTube.


Quote:
Jake Schwartz offers HHC conference videos...

For what it is worth, I hope to have the HHC2013 video ready within a week....lots of editing of 15+ hours of material.


Thanks,
Jake

Hi Richard,

Thanks for sharing! Looks promising :-)

Looking at picture #2 (merci Didier!) in particular: The display looks a bit less high res than I expected :-? And the keyboard is a plain copy of the layout of the HP-42S the traditional Chinese way (I apologize!) - with the brains turned off ;-) as becomes obvious looking at shifted zero.

d:-)

Edited: 20 Oct 2013, 8:11 p.m.


Eric explained that the proto that you see is a repro of the 42s display font, but that the actual production display is extreme hi res.

So the font pictured is not as refned as can be displayed by that screen. I think the 41c demo was an extremely refined display.

Geoff

The display resolution is 400x240. However, running the hacked version of Free42 that I currently am using, the effective resolution is 131x80, with each pixel scaled by a factor of three on each axis.

Since Free42 is what's running on it, we needed the legends on the overlay to match for the present.


There! From the horses mouth!

Thanks Eric

P.s. I want one!


Quote:
There! From the horses mouth!

Thanks Eric

P.s. I want one!


+1 - for repurposing as agreed on many months ago (speaking for 'The Gang of Three' here).

Thanks for explaining the font. There are two fonts readily designed making use of your LCD and covering all the characters the absolute majority of mankind uses (approx. 3.7e9 people).

Eagerly looking forward to a sample.

d:-)

Quote:
P.s. I want one!

It'll still be a while. Our first test run of this prototype design identified some minor design issues. We presently have four units, one of which doesn't have a correctly wired USB port. (The error was found after the first unit was assembled, and corrected prior to assembly on the subsequent three, but is difficult to correct on an assembled unit.)

After the four units were built, we came up with a different, more efficient hardware/firmware design for support of the HP 82440 infrared printers. One unit has been modified to support that, and printing works fine with a standalone program that does nothing else. Unfortunately printing fails when I integrate the printing code with calculator code, apparently due to interrupt timing issues. I'm currently working to resolve that. This is on the critical path at the moment because it determines whether the wiring change for the infrared LED will be implemented in the next revision of the circuit board layout.

We plan to do another small test run with corrections for the known issues, and possibly the infrared LED change. We will make a very small number of units available to the rest of the firmware development team. Once that is done I expect to spend more of my time on industrial design for a manufacturable housing.


Also, remember the 34S firmware took upwards of three years to develop and this device will have more capabilities.

- Pauli

Absolutely wonderful Eric. Wish I was part of the design team.

Something I meant to ask was IR printer support; you answered that, how about a clock integrated similar to the 41cx?

Cheers, Geoff


The hardware is capable of providing clock functions. Whether they're supported in the firmware will be another matter.


Hear that boys (committee), better have a clock!!!

:-)


Quote:
Hear that boys (committee), better have a clock!!!

:-)


... and soon, too. You know what they say: time waits for no man.

;)

Best,

--- Les

[http://www.lesbell.com.au]

Just a suggestion. Offer prototypes on sale to offset some of your design costs. I would consider buying a prototype and beta testing. People that buy a prototype would have to be aware that they are NOT buying a finished product. However, that will certainly reduce some of your costs and development time to get a few extra units out there. And I am sure there are many on this site that would be happy to buy a prototype and then a finished product too.

Your prototype already looks better to ME than the current Hp Prime that I will be buying (and by the looks of it, beta testing too...).

If you were to post or contact me offline about selling me one from your next batch, I (and I suspect others)would be happy to buy. I admit you may have to be selective as some people do not realize a Prototype IS NOT a finished product.


Quote:
Offer prototypes on sale to offset some of your design costs

I agree. I'll bet that you'd do extremely well cash-wise if you put this project up on a crowd funding site. It looks like it's at the perfect stage for that.


Nice!

I like both ideas, prototypes on sale and crowdfunding.

Massimo

Quote:
Just a suggestion. Offer prototypes on sale to offset some of your design costs. I would consider buying a prototype and beta testing. People that buy a prototype would have to be aware that they are NOT buying a finished product.



I am curious...



Have you seen the prototypes of the calculator in the folded Mylar case?


Do you think that folded Mylar case is acceptable to you as a prototype worth buying?


Would you only buy a prototype with a promise of a molded case later? If so, how much later?

-- Richard


Speaking only for myself, of course:

Quote:
Have you seen the prototypes of the calculator in the folded Mylar case?

Only in the pictures posted near the top of this thread.

Quote:
Do you think that folded Mylar case is acceptable to you as a prototype worth buying?

I would buy one, both to contribute to the project, and for the curio value!

Quote:
Would you only buy a prototype with a promise of a molded case later? If so, how much later?

Not at all. When contributing to a kickstarter project, it has to be understood that there is a risk involved, even if all the ducks have been lined up. If a molded case were to become available later, I'm not even certain that I wouldn't keep the mylar case for its uniqueness and buy a separate "production" unit for actual use.

In short, sign me up!

--T


I would buy every prototype round to ensure this has sufficient funding to be completed. Even at $500+ per.

I would buy two or three completed units. Preferably not more than $400 each.

I am a sucker for this project and love everything about it.

Cheers,

Doug

I'd buy any prototype to further the project, especially if it would speed the pace up.

Doug

Eric, if you look at the IR code running on the WP 34S, you can see that I've successfully avoided such timing issues by using a hardware timer for most of the signal forming. I do not recall the exact details but interrupts are only used on a bit level, not the signal timing level.


I am using a timer. The issue seems to be interrupt latency, especially when the USB stack is active. I've got several ideas for fixing it, but haven't had time to work on it recently.

I wanted to have the timer underflow trigger a DMA transfer to reload the timer PWM threshold from the next value in an array, but unfortunately the timers that can run with only the 32.768 kHz ooscillator active are not able to trigger DMA (even though the LEUART that runs on the same oscillator can trigger DMA).

Edited: 23 Oct 2013, 3:00 p.m.

Richard & Eric:
Impressive, as always. I will "need" one.


Quote:
Richard & Eric:
Impressive, as always. I will "need" one.

IMHO almost everybody here will 'need' at least one
... plus one for backup
... plus one as a gift to a good friend
... plus one to teach math to one's kids
... plus ...

d;-)


Quote:
plus one for backup
... plus one as a gift to a good friend
... plus one to teach math to one's kids
... plus ...



plus one to accurately calculate the escape timings after the bank robbery you'll need to do, to pay for all of them... :P
:)


Possibly Related Threads...
Thread Author Replies Views Last Post
  HP-80 Teardown with pictures Max Stone 10 588 10-24-2013, 09:10 PM
Last Post: Walter B
  RPN-1200 (with pictures) Benoit Maag 12 566 09-02-2013, 06:09 PM
Last Post: Matt Agajanian
  HP Prime pictures with video ! Mic 55 1,445 06-24-2013, 02:52 PM
Last Post: CompSystems
  New pictures of HP Prime in hand Mic 52 1,992 06-15-2013, 10:15 PM
Last Post: Peter Murphy (Livermore)
  New screen on pictures of new Sharp EL-9950 Mic 4 294 06-10-2013, 10:40 AM
Last Post: Mic
  HP-65 Repair/Restoration (Pictures) :) Dan Lewis 8 458 01-12-2013, 12:41 AM
Last Post: Dan Lewis
  PCB pictures of a HP10B, HP20S or HP21S? Christoph Giesselink 1 158 12-17-2012, 10:19 PM
Last Post: Randy
  DM12CC in pictures Ingo Paschke 0 126 07-15-2012, 06:54 AM
Last Post: Ingo Paschke
  Advertisement pictures Luiz C. Vieira (Brazil) 4 227 03-07-2012, 12:52 PM
Last Post: Luiz C. Vieira (Brazil)
  Re: Richard Nelson's pictures of the 4 HP 15c versions Frido Bohn 0 130 02-27-2012, 09:44 AM
Last Post: Frido Bohn

Forum Jump: