43s status

[Tugdual]

I've seen some 43s rendered pictures, threads, youtube videos looks like the concept is extremely exciting. Now what is the 43s project? Sorry if my question is naive but I would really love to know if there is a real project, release dates, selected hardware, whatever communication (web site, blog, tweet). Who is doing that? Same people as for the 34s?
I'm excited with this project, i would love to see it, can I help? What do you guys need?

[r. pienne]

Yes, a brand-new calculator doing everything the WP34s and HP42s do and more, masterminded by the brains behind the 34s. The project web pages are at wiki4hp.com and sourceforge.net (though they look a bit out of date). A seriously impressive-looking job indeed. I'm already diverting my beer money into the 43s purchase fund. :-)

[walter b]

You're not the first one asking. Please see here for some answers.

Who are the folks engaged? HW: Eric and Richard (you may know them from the HHC conferences where they use to talk about "scaled reptiles" for some years). UI & SW: the WP 34S team as defined in the foreword of the respective manual.

Please don't hesitate asking for more information if that shouldn't be sufficient.

d:-)

[Tugdual]

Thank you walter. Looking forward to see "day X" :-)

[jwm]

I was wondering if using the 50g as a platform for the 43s (similarly to the 30b being used for the 34s) would be possible? Once you run native code on it, you can completely replace the original firmware in RAM without sacrificing the original calculator functionality.

[walter b]

(02-08-2014 01:06 AM)jwm Wrote:  I was wondering if using the 50g as a platform for the 43s (similarly to the 30b being used for the 34s) would be possible?

Maybe possible but I won't do it. The HP-50g (yes, I've got it) falls in the 'battleship' class, pocket calculators are significantly smaller IMHO. YMMV.

d:-)

[sa-penguin]

The 34S is nice, but those 3 color function keys feel like overkill. They remind me of CTRL, ALT and SHIFT on a keyboard: you could eliminate one by invoking both the other two [CTRL-ALT-DEL to reset becomes F-G-CLR]
Admittedly, that's a strange combination. Menus you tend to stay in [once selected] would be reached by this combination. Setup, alarms, I/O, Trig, statistics, finance, units, boolean logic come to mind.

For your consideration.

[Steve Simpkin]

If you consider the shift keys on the WP 34S to be overkill, consider the WP 31S.

[Paul Dale]

There was a poll many years ago about the number of shift keys on the 34S. Three was the desired result. We could have done the same functionality with fewer shift keys if pushed.


- Pauli

[sa-penguin]

(02-22-2014 12:36 PM)Paul Dale Wrote:  There was a poll many years ago about the number of shift keys on the 34S. Three was the desired result. We could have done the same functionality with fewer shift keys if pushed.

There's plenty of lively discussion about how a layout should be done. From what I see, the 43s is pretty much sold on 2 function keys. Which is fine, if only to reduce clutter on the keyboard. Having another set, reached by pressing F & G together, simply gives you... options.

To keep down keyboard clutter, you could even limit those menus to the "custom" keys at the top. Make the end buttons scroll left/right, and you'd still have 4 options per screen, reached by this method.
If I can get my head around the code in the 34s, I plan to do an example. With numbers like 34, 31 and 30 being used, I'll have to settle for the WP-33_1/3.

[Jake Schwartz]

Quote:To keep down keyboard clutter, you could even.....

Sometimes more clutter is more desirable than an excessive number of keystrokes to execute a function. It's a delicate balance, IMHO. Take a look at this investigation here and here, done back in 2010-2011.

Thanks,
Jake Schwartz

[Richard Ottosen]

(02-01-2014 10:37 PM)Tugdual Wrote:  I've seen some 43s rendered pictures, threads, youtube videos looks like the concept is extremely exciting.

Some clarifications. There are no renderings of the calculator. All of the pictures
and video you have seen are of real physical devices -- either working
calculators or, possibly, mechanical mockups. There are two likely reasons for
the confusion. First, the display has extremely good contrast. It looks too
good to be real even in photos. Second, the keyboard overlay is a graphic made
to look like real calculator keys. It is printed on photo paper.

For those of you that have not held a prototype calculator here is a quick
description...

The display has very black pixels on a slightly gray background. The calculator
is usable in dim light even though it does not have a backlight and draws so
little power that the calculator can be left on for months before the coin
cells run down.

Five of the six sides of the prototype calculator case are clear and the bottom
can be any color that is available in acrylic sheet although I have only made
black and clear bottoms. Other possible colors are opaque or translucent white,
silver or gold mirror and a large variety of common colors. I have even
considered fluorescent orange and green. :-)

The electronics is on a single printed circuit board with the battery holders
and display attached to it. The case is made up of four parts: two spacer
bars, a back plate and a clear Mylar sheet. All of these parts are cut with
an Epilog Laser.

The Mylar sheet is bent into shape to make up five of the six walls of the
case. (Actually, the fifth wall is the battery door). Openings are cut in the
Mylar for accessing the SD card and USB connectors.

The spacer bars are screwed to the PCB and then the bent Mylar sheet and back
plate are screwed to the spacer bars. The keyboard overlay is slid in through
the open battery door. It sits between push button keyboard switch tops and
the bent Mylar. The battery door is held closed by a screw in the back plate.

I hope this helps answer some of your questions.

[brouhaha]

And a further clarification. The DIY4, DIY4X, and DIY5 hardware that exists and has been shown at conferences etc. does not (yet) run WP-43s firmware, because the WP-43s firmware doesn't yet exist. Up to now, we've run microcode-level 41CX simulation (with some added features like 256 flags, and four-level stack display), and Thomas Okken's Free42 (modified for seven-level stack display).

Part of the purpose of the DIYx series of hardware is explicitly to support open-source firrmware, so that there isn't necessarily any "official firmware" for it, unlike calculators from normal vendors. (Even though HP actually encouraged development of alternate firmware for the 20b/30b, they obviously intended most buyers to use the official firmware.) That said, we expect that once the WP-43s firmware is available it will be quite popular (in so far as a specialty calculator for such a limited market can in any sense be said to be popular), and we may well ship units with that firmware preinstalled.

A major reason why the WP-43s firmware doesn't yet exist is because we (Richard Ottosen and I) have not yet provided hardware to the firmware team. We've still got a few minor issues to nail down before we make the next batch of hardware, and it is expected that units from that batch (DIY5A?) will be provided to the firmware team.

I'm very much looking forward to finalizing the design of the DIY5A electronics, because then I can spend more time on the design of the physical case. I do the 3-D design work in Solidworks, prototype on a Lulzbot TAZ 3 printer, and am trying to design for injection molding. Prototype units provided to the firmware team will be supplied in the folded-mylar-and-acrylic case that Richard described, because that's robust enough for development use, but doesn't take too long to make. (Easy for me to say, since Richard makes them.)

One further note on the LCD display. If you've seen the DIY5 runing Free42, the display looks pretty nice, but nowhere near as nice as it will when running WP-43s firmware. That's because WP-43s will use the full 400x240 display resolution, while Free42 is running at only 132x80, scaled up by a factor of 3 on each axis. In other words, the port of Free42 is only showing off 1/9 of the display resolution. With firmware that doesn't do that scaling, the display looks even better. For instance, the DIY4X running 41CX microcode simulation uses full-resolution bitmap fonts rendered from TrueType fonts, and they look really smooth.

[Richard Ottosen]

(02-27-2014 08:28 AM)brouhaha Wrote:  And a further clarification.

More status and other comments I sent to the software development team a few days ago.

I have been using a HP42s version of the DIY5 as my daily calculator hoping
(?) to find any subtle flaws in the DIY5. I haven't used my DIY4 or HP41C too
much lately. I really do dislike the HP42s having the arithmetic keys in
wrong place and in the wrong order.

Eric and I have been doing some packaging stuff -- Eric on the future
production version using injection molded parts and me on the prototype
version using LASER-cut parts. Both of these effect the mechanical design
of the printed circuit board.

Eric has been real busy with work so he has not gotten much done on the
software and I have had other projects distracting me so I have not done much
on the hardware. Once Eric gets the software cleaned up I will finish cleaning
up the hardware and we will get new PCB's made. One of our goals is to have
a local company do automated assembly of the next batch of boards.

On another note... The version of software in my DIY5 does not turn itself
off after a few minutes of inactivity. Either I forget to turn it off or it
gets turned on when I put it in my carrying case. This is really not a huge
problem since the DIY5 can stay on for months before the battery runs down.

This did get me thinking, however. I think the way the On key works is
"backwards". I think that the calculator should turn when the Shift and
On keys are pressed and turn off when just the On key is pressed. This would
make accidentally turning on the calculator less likely and more likely that
it would turn off accidentally. What do you think?

[sa-penguin]

(02-28-2014 02:14 AM)Richard Ottosen Wrote:  This did get me thinking, however. I think the way the On key works is
"backwards". I think that the calculator should turn when the Shift and
On keys are pressed and turn off when just the On key is pressed. This would
make accidentally turning on the calculator less likely and more likely that
it would turn off accidentally. What do you think?

Well - my own experience is based on the TI keyboard encoder, TCA8418. This can be programmed to generate an interrupt when a key is pressed. The interrupt can then be used to "wake" the CPU which then re-configures power mode, starts up LCD etc.

Can the chip be configured to look for a combination, vs a single key? I just.. don't know. Assembling a machine is a part-time hobby for me. I don't know how the keypad is scanned on the 43S but I'd hope the concept of a single scanning chip running while everything else hibernates is in your consideration.

In short: If you're doing everything from the CPU, do whatever you like. Document it in source code, so those who disagree can change the configuration. If you have other hardware [like me] - there may be other limits on your suggestion.

Looking forward to hearing your first run of prototypes is on Kickstarter [grin]...

[walter b]

(02-28-2014 02:14 AM)Richard Ottosen Wrote:  

(02-27-2014 08:28 AM)brouhaha Wrote:  And a further clarification.

More status and other comments I sent to the software development team a few days ago.

... I think the way the On key works is "backwards". I think that the calculator should turn when the Shift and On keys are pressed and turn off when just the On key is pressed. This would make accidentally turning on the calculator less likely and more likely that it would turn off accidentally. What do you think?

Since publishing mails seems becoming popular, here's what I responded to Richard's comment a few days ago:

I can live with almost any position of ON. I strongly vote for leaving it UNshifted. And I think having ON in the bottom row of keys will make it the easiest to protect it by some stiff cover/case. You may call me a traditionalist in that matter but there must be a reason why (IIRC) all the younger HPs have their ON keys down there, or did I miss anything?

d:-/

[Ed Wright]

Some HPs have a low profile ON key to reduce the chance of accidentally turning it on. Is that an option in this case?

[Richard Ottosen]

(02-28-2014 04:19 PM)Ed Wright Wrote:  Some HPs have a low profile ON key to reduce the chance of accidentally turning it on. Is that an option in this case?

The way the prototypes are made, the keyboard is flat. It has to be that way to keep it simple to build using the Laser-cut parts.

In the injection molded case this might be possible, maybe even practical.

[colinh]

(02-28-2014 11:22 AM)walter b Wrote:  You may call me a traditionalist in that matter but there must be a reason why (IIRC) all the younger HPs have their ON keys down there, or did I miss anything?

I'm guessing that the 4 corner keys are the easiest to find, without even looking at the device. Also, having it somewhere in the middle (like 3 in, 3 down, from the top left) would be quirky in the extreme :-)

The ON/OFF key can't be one of the top corners, because they're the soft keys / A-F.

It can't be bottom right, because that's one of the simple algebraics (+ - x /).

That just leaves bottom left :-)

[colinh]

(02-24-2014 02:55 PM)Jake Schwartz Wrote:  Sometimes more clutter is more desirable than an excessive number of keystrokes to execute a function. It's a delicate balance, IMHO. Take a look at this investigation [...]here, done back in 2010-2011.

Interesting article. In particular Richard Nelson's comment on finding/pressing keys (e.g. that a double key press is easier than two separate key presses). That amounts to a weighting of key presses.

A further modification might be that certain keys are easier to find than others: keys in corners are easier to find than keys on the edges, which are easier than keys somewhere in the middle. The Enter key is easier than others. A bright orange shift key is easier than others.

Another weighting factor which doesn't appear to have been considered is the frequency of usage of a function - ie. its importance. And here's where it gets tricky, because different people have different needs.

Ideally (theoretically) you could study a person's usage and construct a frequency plot. You'd do this for a large sample of people and perhaps simply add the results.

Perhaps the 43S (or at least the prototypes) could have a keystroke logging facility. I don't think there would be data protection issues here, unless the NSA were to say "hmmm, this guy uses PI and sin an awful lot. Let's put him on our list of suspicious persons..." :-)