16 replies to this topic

[DrCoyote]

So, I was fooling around with ClassPad Manager, using the Spreadheet application. I decided to create a spreadsheet that calculated the measurement uncertainty of a Hart 9131 infrared calibrator (a sort of portable blackbody), the measurement uncertainty of a Raytek Raynger infrared thermometer, and the test uncertainty ratio between the two. The spreadsheed calculated these at 1 deg. C intervals from 20 to 400 deg. C, and the result was done up as a graph so I could easily see at which temperatures the ratio was best.

I did it all as fills and sequences. It was very quick and easy, and the result was informative. I decided to transfer the spreadsheet to my ClassPad. There's where the fun stopped.

When I tried to open the spreadsheet on the Classpad, I got a message that I had insufficient system memory.

The ClassPad and the ClassPad manager were both in their initial state. I guess the emulation is somewhat wanting in accuracy. It's a shame, really.

[SoftCalc]

I don't exactly understand, but if there is a problem I'd like to see if it can be fixed....

• Did the Manager give inaccurate results or was the inaccuracy poor?
  
• I'm not sure why it couldn't transfer to the actual ClassPad. The Manager doesn't have a memory limit like the ClassPad so it is possible that the spreadsheet was too big for the handheld. Still, it doesn't sound like it was complicated enough to cause insufficient memory. Could you email me the spreadsheet (the MCS image with the spreadsheet) so I can find out what the problem is?

[DrCoyote]

I don't exactly understand, but if there is a problem I'd like to see if it can be fixed....

• Did the Manager give inaccurate results or was the inaccuracy poor?

The ClassPad Manager gave the expected results. Everything worked fine on the ClassPad Manager.

• I'm not sure why it couldn't transfer to the actual ClassPad. The Manager doesn't have a memory limit like the ClassPad so it is possible that thew spreadsheet was too big. Still, it doesn't sound like it was complicated enough to cause insufficient memory. Could you email me the spreadsheet (the MCS image with the spreadsheet) so I can find out what the problem is?

<{POST_SNAPBACK}>

The spreadsheet transferred to the ClassPad. The problem occurs when I try to open the spreadsheet on the ClassPad.


Unfortunaely, I already deleted the spreadsheet. Perhaps I'll reconstruct it some time. If I do, I'll send it along.

[SoftCalc]

The spreadsheet transferred to the ClassPad.  The problem occurs when I try to open the spreadsheet on the ClassPad.
Unfortunaely, I already deleted the spreadsheet.  Perhaps I'll reconstruct it some time.  If I do, I'll send it along.

<{POST_SNAPBACK}>

The memory limit isn't with the spreadsheet file, which is stored as a variable in MCS. The memory limit is with the RAM. The RAM usage will be greatly improved in OS 3.0 so you'll be able to create large spreadsheets without a problem.

If you ever do find a problem with the spreadsheet again please don't hesitate to let me know. I might be able to find the cause, as well as a solution.

[SoftCalc]

On that note, there are sometimes ways you can reduce the memory requirements of the spreadsheet.

The spreadsheet keeps 1 to 2 expressions for every cell in RAM. Every filled-in cell has a value (1 expression) and some cells also have a formula (a 2nd expression). You can reduce the amount of RAM needed by using "fill sequence" instead of "fill range" whenever possible.

Here is an example of two ways of creating a column of numbers 1 through 500 using "fill sequence" and "fill range". Both examples give the same result but "Fill Sequence" will use less than 1/2 the memory because it doesn't need to keep a formula in memory.

Fill Range
Put the value "1" in cell A1 and then fill cells A2:A500 with the formula =A1+1 you'll get a nice column of sequential numbers. Memorywise though, you'll have a formula AND a value in each cell. This takes up more memory.

Fill Sequence
Enter the following in the Fill Sequence dialog...Expr.: x
Var: x
Low: 1
High: 500
Step: 1
Start A1
You'll still get a nice column of sequential numbers but since each cell will only have a value (no formula needed) it will take less than 1/2 the memory.

[DrCoyote]

On that note, there are sometimes ways you can reduce the memory requirements of the spreadsheet.

The spreadsheet keeps 1 to 2 expressions for every cell in RAM. Every filled-in cell has a value (1 expression) and some cells also have a formula (a 2nd expression). You can reduce the amount of RAM needed by using "fill sequence" instead of "fill range" whenever possible.

Here is an example of two ways of creating a column of numbers 1 through 500 using "fill sequence" and "fill range". Both examples give the same result but "Fill Sequence" will use less than 1/2 the memory because it doesn't need to keep a formula in memory.

Fill Range
Put the value "1" in cell A1 and then fill cells A2:A500 with the formula =A1+1 you'll get a nice column of sequential numbers. Memorywise though, you'll have a formula AND a value in each cell. This takes up more memory.

Fill Sequence
Enter the following in the Fill Sequence dialog...Expr.: x
Var: x
Low: 1
High: 500
Step: 1
Start A1
You'll still get a nice column of sequential numbers but since each cell will only have a value (no formula needed) it will take less than 1/2 the memory.

<{POST_SNAPBACK}>

Thanks for the hints. Actually, I did it that way, not out of knowledge of how it conserved memory, but because it makes sense.

[SoftCalc]

Thanks for the hints.  Actually, I did it that way, not out of knowledge of how it conserved memory, but because it makes sense.

<{POST_SNAPBACK}>

Thanks for reporting the problem.

I am bothered by the insufficient memory error. This is the type of error that can mask a bigger problem. For example if the data wasn't transfered correctly the spreadsheet could be corrupt. If this is the case it could be trying to load thousands of cells that don't exist!

I don't think this is the problem, but whenever I see an insufficient memory error I'm always on alert. If this isn't the problem then I'm bothered that it didn't have enough memory to open the spreadsheet.

[PAP]

Speaking about the spreadsheet, I have found some serious limitations. Maybe you have something to say about these limitations, so I post them here:
Although ClassPad's spreadsheet supports the CAS (a unique feature), very useful functions that are common in every spreadsheet application are missing. I'm not speaking about rarely used functions, I'm speaking about functions that are present even in simple spreadsheets. As far as I know, functions such as countif and sumif are missing. This is a serious limitation .
Furthermore, all statistical hypothesis tests, computation of confidence intervals etc, are actually commands, not functions. This means that they simply print the results in a message window; they don't return anything, so they cannot be used in the spreadsheet, or in a basic program . It's a pity, because ClassPad has rich statistical capabilities for a calculator. I cannot see why they are implemented as commands. For example, the command OneSampleZint just prints the confidence interval (left and right edge). If you need the results for further calculations, you should write down the results in a paper . It is more than obvious that OneSampleZint should be implemented as a function returning a list containing the left and right edge of the confidence interval. This way it could be used in the spreadsheet or a basic program. As it is now, it is almost useless.
Btw, this is not the only case where ClassPad uses commands, while functions would be much more useful. For example, string manipulation (StrJoin etc) is also implemented by commands.

[DrCoyote]

Goodness. Now I don't feel quite so lonely.

[PAP]

Goodness.  Now I don't feel quite so lonely.

You are not alone, definitely. Many users have complaints about ClassPad limitations. I don't think that it is sooo difficult to remove many of these limitations (if not all of them). This makes me really angry. For the moment, I'm trying to keep a "positive attitude", because ClassPad has the potential to become a really great calculator. I don't think that I will keep this attitude forever. In fact, if the next OS version will be like OS 2.00, I will probably consider to buy V200 or even HP49g+ (you have done this already, if I remember well).
It is a pity, because many serious limitations can be removed without much effort. For example, how difficult is to implement statistical computations (and string manipulation) with functions, not with commands? How difficult is to remove that silly one-expression limitation that exists in function definitions? How difficult is to add better matrix support? (currently you cannot even compare 2 matrices with the same dimensions). The list of limitations that should be removed is unfortunately very lengthy. My patience won't last forever.

[DrCoyote]

By the way, PAP- I like your signature. I have never played games on a calculator, nor do I plan to.

Throw a bit of that programming prowess behind improving the ClassPad, Casio.

[PAP]

By the way, PAP-  I like your signature.  I have never played games on a calculator, nor do I plan to.

Thanks . There is nothing wrong with games, but I didn't bought ClassPad for this. However, If you understand french, try "Final Fantasy". I don't actually like this game, but it is interesting from a programming point of view. The author has included many features on the game, although it is written in basic. It has plenty of pictures, some kind of small "videos", and the story has a nice sense of humor. Implementing such a game in ClassPad's basic is an (useless) achievement.
Afterall, I don't see a single reason to play a "game" on a calculator while I can play StarCraft or Half Life 2 on my PC .

[DrCoyote]

I don't exactly understand, but if there is a problem I'd like to see if it can be fixed....

• Did the Manager give inaccurate results or was the inaccuracy poor?
  
  
• I'm not sure why it couldn't transfer to the actual ClassPad. The Manager doesn't have a memory limit like the ClassPad so it is possible that thew spreadsheet was too big. Still, it doesn't sound like it was complicated enough to cause insufficient memory. Could you email me the spreadsheet (the MCS image with the spreadsheet) so I can find out what the problem is?

<{POST_SNAPBACK}>

So, did you have a chance to look at that file yet? Were youable to duplicate my problem, or is it just something on my end?

[SoftCalc]

So, did you have a chance to look at that file yet?  Were youable to duplicate my problem, or is it just something on my end?

<{POST_SNAPBACK}>

It turns out it's a really big spreadsheet for the ClassPad. There are close to 2000 formulas and another 2500 values. On the PC this isn't a problem but on the handheld this is too much to store in RAM all at the same time.

It looks like the formulas in all the columns are all the same modulo the row. Having a feature like a "column formula" would make this spreadsheet more than small enough to run on the ClassPad.

[DrCoyote]

It turns out it's a really big spreadsheet for the ClassPad. There are close to 2000 formulas and another 2500 values. On the PC this isn't a problem but on the handheld this is too much to store in RAM all at the same time.

It looks like the formulas in all the columns are all the same modulo the row. Having a feature like a "column formula" would make this spreadsheet more than small enough to run on the ClassPad.

<{POST_SNAPBACK}>

So, I was correct. The emulation is not accurate. If it were, the ClassPad manager would not be able to open the spreadsheet either.

[SoftCalc]

So, I was correct.  The emulation is not accurate.  If it were, the ClassPad manager would not be able to open the spreadsheet either.

<{POST_SNAPBACK}>

You are correct. The Manager is not really an emulator, but a more like a cross-compiled, PC version of the ClassPad. It is compiled native for the PC.

It is designed to emulate (or should I say simulate) the ClassPad, but the memory configuration is different so the Manager doesn't have the same memory limitation.

There are a few problems with having it native on the PC. First, you can create a file in the Manager any happily go along thinking it will work on the ClassPad when it might not.

A second problem is with the floating point numbers. The PC uses IEEE floats for many calculations while the ClassPad uses it's own OBCD numbers. Most of the time this isn't a problem with the exception of really big or really small numbers. The ClassPad's OBCD uses an exponent of +/- E9999 which is a larger range than IEEE floats.

[DrCoyote]

You are correct. The Manager is not really an emulator, but a more like a cross-compiled, PC version of the ClassPad. It is compiled native for the PC.

It is designed to emulate (or should I say simulate) the ClassPad, but the memory configuration is different so the Manager doesn't have the same memory limitation.

There are a few problems with having it native on the PC. First, you can create a file in the Manager any happily go along thinking it will work on the ClassPad when it might not.

A second problem is with the floating point numbers. The PC uses IEEE floats for many calculations while the ClassPad uses it's own OBCD numbers. Most of the time this isn't a problem with the exception of really big or really small numbers. The ClassPad's OBCD uses an exponent of +/- E9999 which is a larger range than IEEE floats.

<{POST_SNAPBACK}>

That is good and useful information. Especially the difference in floating point numbers. That might actually come into play, should I use the calculator for work.

Thank you for the time and work.