Threaded Mode | Linear Mode

walter b · 02-20-2015, 08:20 AM

Just checked again: those links are ok. And the ISBNs are unambiguous. So if you can convince CreateSpace, you might get one book for free. Wink

Good luck!
d:-)

Gerald H · 02-20-2015, 01:17 PM

(02-20-2015 08:20 AM)walter b Wrote: Just checked again: those links are ok. And the ISBNs are unambiguous. So if you can convince CreateSpace, you might get one book for free.

Good luck!
d:-)

Difficult to tell whether I can convince createspace. I've now written to them 3 times & have yet to receive an answer.

Didier Lachieze · 02-20-2015, 01:56 PM

(02-19-2015 04:56 PM)Gerald H Wrote: Manual arrived today, very nice.

However, it is for Firmware 3.2, not 3.3.

Have already complained at createspace.

Has anybody ordered and received a v3.3 manual from createspace?

Terje Vallestad · 02-20-2015, 05:34 PM

(02-20-2015 01:56 PM)Didier Lachieze Wrote: Has anybody ordered and received a v3.3 manual from createspace?

I have ordered but not yet received. Looking at the order acknowledgement I am now not sure what firmware version it will be for, as they only confirm "WP 34S Owner's Manual" with no reference to revision.

We'll see in a week or two when it arrives.

Cheers, Terje

Gerald H · 02-20-2015, 06:03 PM

(02-20-2015 05:34 PM)Terje Vallestad Wrote:
(02-20-2015 01:56 PM)Didier Lachieze Wrote: Has anybody ordered and received a v3.3 manual from createspace?

I have ordered but not yet received. Looking at the order acknowledgement I am now not sure what firmware version it will be for, as they only confirm "WP 34S Owner's Manual" with no reference to revision.

We'll see in a week or two when it arrives.

Cheers, Terje

What was the info on the page you ordered on? Did it say Firmware 3.3?

Terje Vallestad · 02-20-2015, 09:24 PM

(02-20-2015 06:03 PM)Gerald H Wrote: What was the info on the page you ordered on? Did it say Firmware 3.3?

I cannot remember. I just followed the link, but if that page hasn't changed it says Firmware 3.3 on the top so presumably that is what is going to be delivered. I shall report back when I receive the book.

Cheers, Terje

BarryMead · (This post was last modified: 02-22-2015 12:19 AM by BarryMead.)

(02-20-2015 01:56 PM)Didier Lachieze Wrote: Has anybody ordered and received a v3.3 manual from createspace?

I ordered one of the "Perfect Bound Rev 3.3 WP-34S Manuals" the first day Walter posted that they were available. Mine is scheduled to arrive soon. I will let you know when it arrives if it is the right version. The page I was ordering from
showed a picture of the "Green" 3.3 manual, and I paid $39.99 for the manual plus $2.24 tax, plus $3.59 shipping for a total of $45.82, so I know I ordered the right version. We will see if the right version gets delivered. Yes!

My Version 3.3 Perfect bound Manual arrived today. It is a great size, and much quicker to thumb through than the hardcover spiral bound edition. So if I need to leave the book open to a certain page to try an example or enter a program, I can use the spiral bound version. For other quick access, or quick search uses, I can use the perfect bound soft cover version. Thanks Walter you created a very good manual!

MarkHaysHarris777 · 02-20-2015, 10:12 PM

(02-20-2015 09:30 PM)BarryMead Wrote: I ordered one of the "Perfect Bound Rev 3.3 WP-34S Manuals" the first day Walter posted that they were available. Mine is scheduled to arrive today. I will let you know when it arrives if it is the right version. The page I was ordering from
showed a picture of the "Green" 3.3 manual so I know I ordered the right version. We will see if the right version gets delivered.

I did the same ... I will also report back when it arrives. By the way, the manuals are two different prices: 3.2 $34, 3.3 $39. If your cart|order says $39 you should be getting the 3.3 manual (green cover).

marcus
Smile

Gerald H · 02-21-2015, 06:29 AM

(02-20-2015 10:12 PM)MarkHaysHarris777 Wrote:
(02-20-2015 09:30 PM)BarryMead Wrote: I ordered one of the "Perfect Bound Rev 3.3 WP-34S Manuals" the first day Walter posted that they were available. Mine is scheduled to arrive today. I will let you know when it arrives if it is the right version. The page I was ordering from
showed a picture of the "Green" 3.3 manual so I know I ordered the right version. We will see if the right version gets delivered.

I did the same ... I will also report back when it arrives. By the way, the manuals are two different prices: 3.2 $34, 3.3 $39. If your cart|order says $39 you should be getting the 3.3 manual (green cover).

marcus

I know which one I ordered - 3.3.

I know which was delivered - 3.2.

walter b · 02-21-2015, 09:00 AM

(02-21-2015 06:29 AM)Gerald H Wrote: I know which one I ordered - 3.3.

I know which was delivered - 3.2.

Then: I know who can't read ISBNs - CreateSpace.

d:-/

DMaier · 02-21-2015, 08:50 PM

(02-12-2015 08:14 PM)walter b Wrote: Merci Didier, you were faster than me. That errata sheet was published in January 2014.

In that case, I don't believe that the error on page 124 of the Firmware v3.2 manual has been identified. The example of computing $\Sigma_{k=1}^{100}\sqrt{k}$ is actually an example of calculating $\Sigma_{k=0}^{100}\sqrt{k}$. The likelihood that the reader will be misled is potentially increased by the fact that the result is, of course, the same.

Also, I believe that the documentation should warn the reader about all operations, like $\Sigma$ and $\Pi$, that clear or otherwise invalidate the stack. The warning should be similar to the one provided for f'(x), preferably improved to not assume that the stack size is 4. It also might be helpful to change the worked $\Sigma$ example to $\Sigma_{k=1}^{100}\sqrt{k} + \Sigma_{k=201}^{300}\sqrt{k}$. or something similar, as a way of conveying to the reader that the stack cannot be relied on to hold intermediate results.

MarkHaysHarris777 · (This post was last modified: 02-22-2015 01:20 AM by MarkHaysHarris777.)

Greetings folks, my v3.3 manual arrived from CreateSpace in this afternoon's snail mail (frozen solid, but otherwise in beautiful condition).

WP 34s Owner's Manual Firmware v3.3
ISBN 9 781507 891070
( v3.3 )
Page Count: 344
Size: 15.2 cm X 22.9 cm

I'll dive into this later this evening (from the arm-chair, with a mocha, and my calculator) and have some fun.

Thank you, Walter, very nice work.

PS. I prefer the size of this manual (6 X 9 inch); it fits better on the ref shelf with the rest of my handbooks.

Cheers,
marcus
Smile

walter b · (This post was last modified: 02-22-2015 04:19 AM by walter b.)

(02-21-2015 08:50 PM)DMaier Wrote: In that case, I don't believe that the error on page 124 of the Firmware v3.2 manual has been identified. The example of computing $\Sigma_{k=1}^{100}\sqrt{k}$ is actually an example of calculating $\Sigma_{k=0}^{100}\sqrt{k}$. The likelihood that the reader will be misled is potentially increased by the fact that the result is, of course, the same.

Also, I believe that the documentation should warn the reader about all operations, like $\Sigma$ and $\Pi$, that clear or otherwise invalidate the stack. The warning should be similar to the one provided for f'(x), preferably improved to not assume that the stack size is 4. It also might be helpful to change the worked $\Sigma$ example to $\Sigma_{k=1}^{100}\sqrt{k} + \Sigma_{k=201}^{300}\sqrt{k}$. or something similar, as a way of conveying to the reader that the stack cannot be relied on to hold intermediate results.

Thanks for reporting. I'll look into that later.

d:-)

Edit:
Ok, although the summation example you quote may be seen as a little test for the reader, I will change the start value to 100.001 to avoid that challenge.
And the documentation stubbornly repeats "... fills all stack levels with x before calling the routine specified." for f', f", SLV, $\Pi$, $\Sigma$, and Integrate. So I think I've fulfilled my duty there, haven't I?

DMaier · 02-22-2015, 08:26 AM

(02-21-2015 10:48 PM)walter b Wrote: Ok, although the summation example you quote may be seen as a little test for the reader, I will change the start value to 100.001 to avoid that challenge.

Changing the value to 100.001 addresses the error in the original example. However, the point of the proposed example is to make the reader aware of the issue described below. I'm not sure that it would be a valid "test", necessarily, unless the user is provided with the information required to succeed.

(02-21-2015 10:48 PM)walter b Wrote: And the documentation stubbornly repeats "... fills all stack levels with x before calling the routine specified." for f', f", SLV, $\Pi$, $\Sigma$, and Integrate. So I think I've fulfilled my duty there, haven't I?

Right. However, the issue is not the contents of the stack before calling the routine specified, which is presumably repeated because it is necessary information for the creator of the routine. The issue is that the stack is filled with zeroes when $\Sigma$ and the rest return. The point of my proposed example is to make the reader aware that the results of the first summation will no longer be on the stack after the second summation is performed.

The Firmware V3.2 documentation for f'(x) states: "On return, Y, Z, and T will be cleared...". This warning should be repeated for the others.

walter b · (This post was last modified: 02-22-2015 09:12 AM by walter b.)

Let's see: V3.3, IOP about ...

f'(x): On return, Y, Z, and T will be cleared and the position x will be in L.
SLV: It returns xroot in X, the second last x-value tested in Y, f(xroot) in Z, and 0 in T. Additionally, SLV acts as a test, so the next program step will be skipped if SLV fails to find a root. Then it points to a footnote stating: Please refer to the HP-34C Owner’s Handbook and Programming Guide (Section 8 and App. A) or the HP-15C Owner’s Handbook (Section 13 and App. D) for more information about automatic root finding and some caveats.
Integrate: ∫ returns the (approximated) integral in X and an upper limit of its uncertainty in Y. Additionally, it points to a footnote stating: Please refer to the HP-34C Owner’s Handbook and Programming Guide (Section 9 and App. B) or the HP-15C Owner’s Handbook (Section 14 and App. E) for more information about automatic integration and some caveats.
Π and Σ doesn't contain any remarks alike. I'll check what happens there. Thanks for reporting.
Addendum: Hmmmh, Π and Σ clear all stack levels except X. Though I've no idea why. Pauli?

d:-)

Paul Dale · 02-22-2015, 09:18 AM

(02-22-2015 08:59 AM)walter b Wrote: Addendum: Hmmmh, Π and Σ clear all stack levels except X. Though I've no idea why. Pauli?

They both call a user routine with the stack filled with the X value. They don't make any attempt to save the initial entry stack (we don't really have the space, especially when they nest) -- this is actually the same behaviour as f' and f", which you've documented as doing exactly this already. It is also the same as solve an integrate -- although these return extra values that hide the stack destruction better.

The general rule is that: a system function that calls a user routine exterminates the stack.

- Pauli

walter b · (This post was last modified: 02-22-2015 01:25 PM by walter b.)

Thanks for clarification. So, for using such functions within a larger calculation, precautions are necessary like:

Code:

STOS 80

Σ 'S'

STO 79

RCLS 79

(have to check that - works! :) ). Procedure for Π may look alike, for the other functions it's a bit more complicated. Couldn't we automate that?

d:-/

walter b · 02-22-2015, 09:38 AM

Local registers come into my mind, can't help it. Wink

d:-)

Marcus von Cube · 02-22-2015, 10:33 AM

(02-22-2015 09:38 AM)walter b Wrote: Local registers come into my mind, can't help it.

Local registers don't increase the amount of available memory they just make a different use of it. When these functions are nested or even used recursively we'll need a considerable amount of additional memory without any benefit. We might use a flag to indicate that it's a nested call and spare the stack saving then. Is it worth the hassle? I doubt it.

walter b · (This post was last modified: 02-22-2015 02:04 PM by walter b.)

(02-22-2015 10:33 AM)Marcus von Cube Wrote: Local registers don't increase the amount of available memory they just make a different use of it.

Know that, been there. Nevertheless, assume a function 'Σa' looking like this for a stack depth of 8 levels:

Code:

LBL 'Σa'

LocR 9

STOS .1

Σ 'S'

STO .0

RCLS .0

RTN

So 'Σa' would do the same as Σ and save the stack. The only problem to be solved is the transfer of the address of the routine containing the function to be summed up (here 'S'). See what I mean? And if 'Σa' would call itself again recursively, everything will be fine until we run out of memory - just the way it should work.

d:-)