WP 31S available at SourceForge

[walter b]

Back to start, please. WP 31S shall be an entry level scientific (incorporating a subset of WP 34S firmware). I don't see how Julian day numbers with all the bells and whistles never seen in any calculator before shall foster that. I'd suggest taking over JG1582 and JG1752 and that's it (we must not leave it with JG1752 alone since it won't do justice to the majority).

d:-)

[Dieter]

(04-18-2014 07:20 AM)walter b Wrote:  Back to start, please. WP 31S shall be an entry level scientific (incorporating a subset of WP 34S firmware). I don't see how Julian day numbers with all the bells and whistles never seen in any calculator before shall foster that.

No bells and whistles required here. But accurate results are not optional. Never. That's why it can't stay the way it is: the 31s currently returns dubious results because the Julian day number - which is used for three date-related commands - is calculated in a way that's wrong for most users.

JG1582 (which should be the 34s default) and JG1752 are better than nothing, and obviously memory restrictions on the 34s did not allow a better, complete solution. But the 31s can have a perfect implementation for correct results under all circumstances. As already stated earlier: "accuracy is not optional".

Something like a JGSET command seems to be not the worst idea. It allows correct results (WDAY, ΔDAYS, DAYS+) for all users. And as an additional benefit it extends the functionality in that it allows calculations based on a proleptic Julian or Gregorian calendar (cf. ISO 8601). I do not see any reason why this should not get implemented. I suggest 15 Oct 1582 should be the default switch date (as usual in calendar calculations), and the user may specify any other.

Pauli, now it's up to you. ;-)

Dieter

[walter b]

FYI, we've got a problem with the commands contained in DISTR. Everything else seems to be working just fine.

d:-)

[Nigel (UK)]

I've got the latest files from Sourceforge and I can't compile the WP-31S windows emulator from Visual C++ Express (2010). The compiler tells me that the files stdafx.h and emulator.lib are missing. Am I doing something wrong?

makeflash.cmd for the WP-31S runs fine.

I can build everything for the WP-34S without any trouble.

Thanks in advance for any help!

Nigel (UK)

[Dieter]

(04-18-2014 07:12 PM)walter b Wrote:  FYI, we've got a problem with the commands contained in DISTR.

Indeed. There are also some troubles with the 34s versions that have been mentioned earlier. For instance a simple case like t ^–1 (p=0,01, df=10) runs for more than 15 seconds on a hardware 34s. So I tried a program in user code, et voilà: less than 5 seconds (resp. 6 in DP). Now imagine how fast this would run in compiled C-code. I assume there's something wrong with the initial guess for the Student quantile. In the above case the method I originally suggested starts with |t|=2,781... (result: –2,7637...), which requires just 4 resp. 5 iterations.

Quote:Everything else seems to be working just fine.

I beg to differ. Remember the calendar issue and the PRIME? test.

BTW - are there any plans on exposing DP mode to the user? If not, several time-consuming 31s functions (especially the distributions) may be rewritten for much faster execution. Remember the HP35 design goal, more than four decades ago: the user should not wait for more than a second until the result appears. ;-)

Dieter

[walter b]

(04-21-2014 01:23 PM)Dieter Wrote:  

(04-18-2014 07:12 PM)walter b Wrote:  FYI, we've got a problem with the commands contained in DISTR.

Indeed. There are also some troubles with the 34s versions that have been mentioned earlier. For instance a simple case like t ^–1 (p=0,01, df=10) runs for more than 15 seconds on a hardware 34s. So I tried a program in user code, et voilà: less than 5 seconds (resp. 6 in DP). Now imagine how fast this would run in compiled C-code. I assume there's something wrong with the initial guess for the Student quantile. In the above case the method I originally suggested starts with |t|=2,781... (result: –2,7637...), which requires just 4 resp. 5 iterations.

That's Pauli's realm.

Quote:

Quote:Everything else seems to be working just fine.

I beg to differ. Remember the calendar issue and the PRIME? test.

Thanks for the reminder. PRIME? should be done (but not built yet) and the calendar might well be solved like on the WP 34S.

Quote:BTW - are there any plans on exposing DP mode to the user? If not, several time-consuming 31s functions (especially the distributions) may be rewritten for much faster execution. Remember the HP35 design goal, more than four decades ago: the user should not wait for more than a second until the result appears. ;-)

DP will not be user-accessible on the WP 31S. Re-writing SW goes to Pauli again.

d:-)

[Marcus von Cube]

(04-20-2014 09:22 PM)Nigel (UK) Wrote:  I've got the latest files from Sourceforge and I can't compile the WP-31S windows emulator from Visual C++ Express (2010). The compiler tells me that the files stdafx.h and emulator.lib are missing. Am I doing something wrong?

Have you tried the wp31s_express.sln file? The Express versions of Visual Studio don't offer the files to compile MFC applications. I've packaged all necessary MFC stuff in a single emulator DLL (including a debug version with symbols) . If there isn't the need to update the emulator GUI, everything can be compiled with Visual Studio 2010 Express edition using the _express version of the solution file. At least this should be the case.

[Nigel (UK)]

(04-21-2014 09:20 PM)Marcus von Cube Wrote:  Have you tried the wp31s_express.sln file? The Express versions of Visual Studio don't offer the files to compile MFC applications. I've packaged all necessary MFC stuff in a single emulator DLL (including a debug version with symbols) . If there isn't the need to update the emulator GUI, everything can be compiled with Visual Studio 2010 Express edition using the _express version of the solution file. At least this should be the case.

I've got revision 3625 and I can't find a file called wp31s_express.sln in any folder in the project. (There's a wp34s_express.sln file, but that doesn't help!)

Nigel (UK)

[Marcus von Cube]

(04-22-2014 06:07 PM)Nigel (UK) Wrote:  I've got revision 3625 and I can't find a file called wp31s_express.sln in any folder in the project. (There's a wp34s_express.sln file, but that doesn't help!)

Looks like a bug. I'll add the file to the repository with the next commit. I'm in the middle of some bug tracking so it may take a few days. IN the meantime You can remove the emulator DLL from the wp31.sln file.

[Nigel (UK)]

(04-22-2014 07:19 PM)Marcus von Cube Wrote:  Looks like a bug. I'll add the file to the repository with the next commit. I'm in the middle of some bug tracking so it may take a few days. IN the meantime You can remove the emulator DLL from the wp31.sln file.

Revision 3629 lets me compile the WP-31S emulator - thanks. A couple of errors come up - after clicking on wp31s_express.sln, while Visual C++ Express is loading it complains about a missing wp31s.vcxproj file, although this doesn't seem to matter. Also, when compiling there's an error message about copying a (non-existent) manual file.

The calculator looks great, and I don't care about the keyboard layout! Thanks again for your help.

Nigel (UK)

[walter b]

The distribution functions should run ok now, too. Build 3636 can be recommended for testing in my view. Enjoy!

([:-)

[Marcus von Cube]

(04-23-2014 07:12 PM)Nigel (UK) Wrote:  A couple of errors come up - after clicking on wp31s_express.sln, while Visual C++ Express is loading it complains about a missing wp31s.vcxproj file, although this doesn't seem to matter. Also, when compiling there's an error message about copying a (non-existent) manual file.

I think I can remedy these issues too. Some missing files in the SVN repository. There isn't a WP31S PDF to copy. I may remove the step from the build process.

[walter b]

After 15 days now, we can frankly say we didn't get any bug reports anymore (nor did we detect any bugs ourselves) so everything just seems running smoothly. That's the obvious advantage of deriving a product.

OTOH, the interest in such an entry-level RPN scientific calculator seems to be ... hmmmh ... expandable so far. Since our main motivation was creating a clean, easy looking surface for a product attracting new monks to the ancient order of urrRh-Pee-anN it looks like we jumped too short. Following a good old motto, I'm interested in your opinion about the (root) causes:

• Is it too early to evaluate yet? Shall we be more patient?
• Is the WP 31S too simple? Are there any crucial features missing?
• Is our target audience traveling elsewhere? Where shall we go to reach them?
• Which other important facts did we miss?
  

TIA for your thoughts.

d:-?

[walter b]

Just adding, we've almost 26 000 views of that related thread and just 7 (seven) orders for the upcoming WP 31S manual so far. I simply can't match these two numbers. I've got no information about the amount of WP 31S downloads so said orders are the only solid figures I can get.

d:-?

[Didier Lachieze]

Just my 2 €cents: the 31S is much simpler than the WP 34S and leverages almost all functions so I don't need so much a dedicated manual for the time being. And before ordering a manual for the 31S there is one thing I need: an overlay to build a real 31S, only after that will I consider ordering a manual for the 31S.

In the mean time I'm eagerly waiting for the WP 34S spiral-bound A6 manual !

[walter b]

Merci, Didier, pour votres pensees.

(04-30-2014 11:36 AM)Didier Lachieze Wrote:  In the mean time I'm eagerly waiting for the WP 34S spiral-bound A6 manual !

A6 ?!? You'll get twice the volume: A5 !

d:-)

[Didier Lachieze]

(04-30-2014 11:54 AM)walter b Wrote:  Merci, Didier, pour votres pensees.

(04-30-2014 11:36 AM)Didier Lachieze Wrote:  In the mean time I'm eagerly waiting for the WP 34S spiral-bound A6 manual !

A6 ?!? You'll get twice the volume: A5 !

d:-)

Yes, I've been too optimistic, I'll not be able to replace the pocket reference guide by the full manual .... (just kidding )
A5 is fine and will be a significant improvement size-wise vs. current manual.

[Manolo Sobrino]

Maybe it's just too soon. The 31S project has been around for just three months and there are only 3 threads about it. It might need more time. Even young wines need a little ageing. I guess that you are busy with the 43S and want to get this side project done as soon as possible.

I can't help thinking that as a scientific it looks nice, but it lacks some features if we compare it with the current generation of scientific do-everything calculators.

a) Number bases, at least (DEC/)BIN/HEX, and logical operators.

b) A solver.

c) Numerical Integration.

b) c) Are the crux of classic RPN. You need programmability to do that. Unless you just have formula storing/evaluation, but that means writing a parser. It would be very useful though. It would provide differentiation from the 34S as well.

On the other hand, it has some overkill features such as the DISTRIB functions, that could be reduced to the essential and moved to the STAT catalogue... And you have to recognise that the LG affair has caused some friction.

[Mark Hardman]

(04-30-2014 10:31 AM)walter b Wrote:  I'm interested in your opinion about the (root) causes

I've been waiting for the availability of an overlay from Eric before flashing a WP 31S.

Mark Hardman

[Paul Dale]

I haven't been all that actively involved in the design on the 31S but the intention from the start was non-programmable which implies no solve or integrate. I think it unlikely anyone would bother adding expression support to the code base, but I might be surprised one day.

I can't really make a meaningful comment about integer mode.

However, the 31S project is far more mature than its age would suggest. The code base started from the 34S, which means lots of issues have already been ironed out. Sure, there have been some changes to the user interface and lower down but the bulk of the code is still the same.


- Pauli