A sweet little poll about the WP 31S layout

[walter b]

Hi all,

There was a recent discussion about said layout in a record-breaking thread (> 23 000 views). Apparently the WP 31S is interesting for many. I don't want to strain that thread further, thus I put the poll here:

Which layout do you prefer (regardless of the colour of the shifted labels which is an orthogonal topic):

a) or b)

TIA for your votes! This poll will run for 120 hours as usual.

d:-)

[Gene]

My vote was for "a" only because of the top two rows of keys.

I have things I don't like about the rest of "a" but the top two rows made the difference for me.

perhaps a poll for different areas of the layout?

(only have joking there)

:-)

[BobVA]

A. And I have no idea why. Perhaps it looks a little 41/45'ish to me...

[Jonathan Cameron]

(04-14-2014 11:56 PM)Gene Wrote:  My vote was for "a" only because of the top two rows of keys.

I have things I don't like about the rest of "a" but the top two rows made the difference for me.

What issues do you have with "the rest of "a""?

-Jonathan

[RMollov]

Surveyors would like 'b' so 'b' all the way!
Why:
1. ->POL & ->REC sit exactly where they should (similarly to HMS+ & HMS-), no travel time - the opposite (as on layout 'a') is unacceptable to me.
2. FILL is where it should be, exactly as on WP34S which works great.

[Jonathan Cameron]

(04-15-2014 04:04 AM)RMollov Wrote:  Surveyors would like 'b' so 'b' all the way!
Why:
1. ->POL & ->REC sit exactly where they should (similarly to HMS+ & HMS-), no travel time - the opposite (as on layout 'a') is unacceptable to me.
2. FILL is where it should be, exactly as on WP34S which works great.

Not sure if it matters to you, but the current implementation of the WP-31s does not support Grads (RAD and DEG only). I've heard that is important for some surveyors, so I'm curious if you need grads?

-Jonathan

[Craig Thomas]

a, by all means.

Overall, what a classic and classy look for your poll images. Nice.

[Craig Thomas]

By the way, is this (WP 31s) under serious consideration?

Would it be made from an existing HP or from 'scratch'?

Thanks for any thoughts.

[Jonathan Cameron]

(04-15-2014 05:05 AM)Craig Thomas Wrote:  By the way, is this (WP 31s) under serious consideration?

Would it be made from an existing HP or from 'scratch'?

Thanks for any thoughts.

This calculator already exists thanks to Sanjeev, Walter, Pauli, Marcus, Bit, and myself. You can download the flash image on the WP-31s branch of the source forge site:

http://sourceforge.net/p/wp34s/code/HEAD...hes/wp31s/

(calc.bin in the 'realbuild' directory) and flash it on a HP 30b (like the WP-34s). It current implements layout A. At this time you will have to create your own overlays. But I think Eric will create overlays for it soon---if we can ever settle on a final layout!

Here is what mine looks like (yes, the letters under the keys are partly messed up or missing):



-Jonathan

[Didier Lachieze]

Option a) for me, however there are two things that bother me in this layout:

1] While I understand the rational behind the location of x<> as a shifted function of RCL and CONV as a shifted function of x<>y, I keep looking to the x<>y key when looking for x<>, I just can’t set my mind to look to RCL instead.

2] It’s the same for FILL, in my mind if associated to the multiple ENTER it’s saving and should belong to the ENTER key as it is on the WP 34S.

For me x<> and FILL are more strongly linked to x<>y and ENTER than the current catalog functions CONV and CONST. So I would prefer to swap these 4 shifted functions:
- Move [x<>] to the [x<>y] key
- Move [FILL] to the [ENTER] key
- Move [CONST] to the [RCL] key, this has some logic as a constant is a recall of an internal value
- Move [CONV] to the [STO] key, no obvious logic here but not less than having [CLEAR] on the [ / ] key for example

I see this as a tradeoff between having all the catalogs neatly grouped together but having troubles to locate FILL and x<> , or having FILL and x<> on the same key as on the WP 34S and where my mind expect them to be, and moving CONST and CONV one row above their current location.

[Paul Dale]

Plus CONST doesn't fit well on a small key.

I like the current 31S layout and will be getting an overlay in due course.


- Pauli

[Paul Dale]

I've also got a bit of coding to do if I get time.

I want to reimplement the statistical distributions so they are native in C which will give a significant performance boost over the keystroke programs in the 34S.

I've improved the trig functions already -- the use the recently mentioned octant reductions to provide nicer results in degrees (and gradian if they were supported).


- Pauli

[walter b]

(04-15-2014 06:10 AM)Didier Lachieze Wrote:  For me x<> and FILL are more strongly linked to x<>y and ENTER than the current catalog functions CONV and CONST. So I would prefer to swap these 4 shifted functions:
- Move [x<>] to the [x<>y] key
- Move [FILL] to the [ENTER] key
- Move [CONST] to the [RCL] key, this has some logic as a constant is a recall of an internal value
- Move [CONV] to the [STO] key, no obvious logic here but not less than having [CLEAR] on the [ / ] key for example

About CLEAR: Shifted <- is taken by UNDO already, so where else should CLEAR go?

About CONST, FILL, x<>y, and CONV: I definitivly like FILL on ENTER and x<> on x<>y (as I mentioned elsewhere already) BUT

• CONST simply won't fit on a small key in the top three rows,
• I can't find a nicer place for CONV and CONST without destroying more than I'd improve.
  

Life is far from perfect

d:-)

[anetzer]

(04-14-2014 07:16 PM)walter b Wrote:  Apparently the WP 31S is interesting for many.

1.
True. Looking at the layouts they are both very clean and elegant and it really comes down to how often one uses certain function groups. I find it hard to really choose one over the other.

2.
Asked myself, why this layout has so much appeal to me. Even though the machine is non-programmable...
Only answer I came up with: So much cleaner and less confusing, almost the elegance of the HP-41.
Get me right: The 34s is a perfectly balanced and elegant machine, given the hardware constraints, but I never quite understood, how HP could go away from the downright solid approach of XEQ ALPHA Something] ALPHA. It kept the keyboard clean and you never had to think a minute, in which menu this or that instruction might hide. MODE, TEST P.FCN and X.FCN are well thought out, but I have to admit that I just have to check two catalogs far to often to find a specific thing...

3.
That said: It should even be possible - in the end - to compile and distribute BOTH layouts for the same firmware and distribute ONE overlay where people would only have to shift a few keys between two rows. I'm sure, some compiler-savvy dissident will be happy to do this, just for the heck of it. When the 34S has six different firmware packages...

What was "Rumpelstilzchen" in english again?


a.n.

[Marcus von Cube]

(04-15-2014 07:57 AM)anetzer Wrote:  Get me right: The 34s is a perfectly balanced and elegant machine, given the hardware constraints, but I never quite understood, how HP could go away from the downright solid approach of XEQ ALPHA Something] ALPHA. It kept the keyboard clean and you never had to think a minute, in which menu this or that instruction might hide.

Don't blame HP for something they aren't responsible for. It's "WP", not "HP".

As for your argument: I don't like [Alpha]something[Alpha] at all without a dedicated alpha keyboard. The best you can get is a catalog with alpha navigation. WP 34S (and WP 31S, too) have this feature built in. In WP 34S we just don't have enough memory to implement a full catalog of all functions. The latter would be even better than your beloved Alpha sequence because there is no need to type more letters than are unique for the command. As it is implemented now you just need to know the catalog a command resides in.

This is what I found out about Rumpelstilzchen: http://www.dict.cc/englisch-deutsch/Rump...tskin.html

[Paul Dale]

Marcus is correct, we didn't do the XEQ [alpha] option deliberately. Partly because of the (lack of) internal organisation I committed to very early on.

The 31S could include a complete catalogue of all functions -- there is space in flash, just not on the keyboard.

The 31S should have enough RAM to remember the position in each catalogue as well. This isn't implemented, it only remembers the position in the last catalogue. This one can be changed if there is enough desire.


- Pauli

[walter b]

Just for sake of completeness: WP 31S could feature a catalog of everything but access would become slower than with dedicated catalogs since you'd need to enter more letters to specify a command unambiguously. Right now names and distribution to catalogs are chosen in a way that you reach the overwhelming majority of commands by specifying one letter only.

d:-)

[anetzer]

(04-15-2014 08:23 AM)Marcus von Cube Wrote:  It's "WP", not "HP".

So it is, and it's HP I've got my beef with.

The WP-Team chose a perfectly sensible approach when selecting the HP-42S as starting point since the HP-41 really only makes sense as - well - the HP-41. The catalogs that exist on the HP-41 have so much to do with the ability to extend the machine via it's ports.

Had the HP-41 continued to evolve, the catalogs might have taken on an online help function, just like a pocket reference. And an HP-IL QWERT - keyboard... Well, I'll stop there; other platforms took over.


I am really looking forward to trying out the 31s...

[Manolo Sobrino]

It's odd that people are judging these layouts just by their looks. You should look at other things, more important and which you realise after some years of use. For instance, finger (and eye) travel and logical groupings/coupling of functions.

For instance, the layout a) has for it that it has the same first row of many HP calcs. Looks nice until you realise that some basic functions are more distant to ENTER and the numpad than they should/could. Which makes that some less used functions are closer. But I won't criticise a), that's not what I was/am looking after...

Layout b) for instance tries to minimise the travel from the more frequently used catalogues in numeric calculations to the arrows, that's where you're going after hitting them. And you are going to use MORE a lot, then arrows, it's optimum place is just there. There are several contiguous calculators in b) that nevertheless work together, depending on your particular use you might ignore whole areas, making it faster to operate.

I've tried to think about every key and optimise the position of all of them for global and particular use. Other example: at b) if you are entering statistical 1D data then the mean and the standard deviation are next to the key you're using all the time, then all the statistical variables are under STAT, and the parameters related to statistical dispersion are under DSTR. Legends have been shortened to 4 characters because then CLEAR, which does something different and should stand out, is the only one with 5... There are reasons for everything, it's not just aesthetics. I believe that form should follow function BTW.

I think it's nice to have a poll about them, thank you Walter. But let's not fool us. You can't design or even choose a good layout without using it first. You should test these things for some time with emulator versions and then choose/suggest alternatives.

[Manolo Sobrino]

(04-15-2014 05:37 AM)Jonathan Cameron Wrote:  Here is what mine looks like (yes, the letters under the keys are partly messed up or missing):

Hey, that's not lemon yellow! we're getting somewhere visually sensible, at last!