23-Emerald V

Question

Prime 3.1 - Problem with until function?

Forum|Forum|10 years ago
July 5, 2015
3 replies
6122 views

I was looking at some of the Rosetta Code challenges for implementation in Mathcad and decided to try a couple in Prime Express (as this is what I have). The major drawback to Express is the lack of programming capability, but there's quite a bit that can be (or should be able to be done) using just range variables and functions that don't make use of programs.

I wanted to pick all of the non-zero elements out of a vector and decided to use the until function to sweep through a reverse sorted vector until it hit a zero. However, Prime keeps telling me that 'until' is undefined, even though I inserted it using the Functions Dialog. Express doesn't raise an alert saying that it's Premium Feature, so I'd expect it work. Neither can I find anything in the Migration Guide or Help to indicate it's been deprecated. Is there a known problem with the until function in Prime 3.1?

Stuart

1_test---15-07-06-until-function-01-mcdx.zip

Other

L

LucMeekes

23-Emerald IV

Mathcad 11.2 'knows' the function: if called with a single parameter it complains there's one missing, if called with 3 parameters it complains there's one too much. If called with two parameters it marks the error: "This function is obsolete. Press [F1] for Help.".

The function is listed with description,

Function for iteration: until(x,y) This function allows you to halt an iterative process based on a condition. Returns y until x is negative.

Arguments: x is a test expression (usually involving a negative, the until function halts iteration.

y is the value returned by the until function at each iteration.

but also in the list of obsolete functions.

The obsolete functions include: APPEND, READ, until, WRITE.

These functions continue to exist in Mathcad to support backward compatibility, but their use is discouraged.

So if I set the worksheet calculation mode to 'backwards compatible' it 'works' although I cannot find it to stop iteration. I tried your expression: iteration runs through to 12.

For 'until(9-k, k^2)' I would expect the iteration to run to 9 but it outputs the full range of k^2:

Ah, you cannot use it this way. It must be like:

or is the same. Too bad for the 0 at the end.

In Prime 3.0 (Express) I get the same error as you find in 3.1.

S

StuartBruff

Author

23-Emerald V

LucMeekes wrote:

Mathcad 11.2 'knows' the function: ..

but also in the list of obsolete functions.

The obsolete functions include: APPEND, READ, until, WRITE.

These functions continue to exist in Mathcad to support backward compatibility, but their use is discouraged.

So if I set the worksheet calculation mode to 'backwards compatible' it 'works' although I cannot find it to stop iteration. I tried your expression: iteration runs through to 12.

For 'until(9-k, k^2)' I would expect the iteration to run to 9 but it outputs the full range of k^2:

In Prime 3.0 (Express) I get the same error as you find in 3.1.

That's interesting. Thanks, Luc. I don't use until very often, but it has its occasional uses, particularly when I'm developing an algorithm (or, more likely, debugging a program).

I ran your example in Mathcad 15 and got this ...

So it definitely appears to have been been unbroken in M15!

Stuart

L

LucMeekes

23-Emerald IV

Hi Stuart,

Note that I changed my response after you grabbed it, see above.

Good to see that the final 0 is omitted in Mcad15, and I guess they put it back in the 'fully supported' list.

Did you try (as I initially did) to just get the result from, e.g. 'until(7-k,k)=' ?

Luc

S

StuartBruff

Author

23-Emerald V

StuartBruff wrote:

I was looking at some of the Rosetta Code challenges for implementation in Mathcad and decided to try a couple in Prime Express (as this is what I have). The major drawback to Express is the lack of programming capability, but there's quite a bit that can be (or should be able to be done) using just range variables and functions that don't make use of programs.

I wanted to pick all of the non-zero elements out of a vector and decided to use the until function to sweep through a reverse sorted vector until it hit a zero. However, Prime keeps telling me that 'until' is undefined, even though I inserted it using the Functions Dialog. Express doesn't raise an alert saying that it's Premium Feature, so I'd expect it work. Neither can I find anything in the Migration Guide or Help to indicate it's been deprecated. Is there a known problem with the until function in Prime 3.1?

Stuart

Could somebody with a full Prime 3.1 please say whether they see the same thing as I get in Express?

Thanks,

Stuart

N

nas0k

1-Visitor

Hello Stuart, I do indeed see the same thing using 3.1

S

StuartBruff

Author

23-Emerald V

Norm Schutzkus wrote:

Hello Stuart, I do indeed see the same thing using 3.1

Thanks, Norm.

As I'm not a maintenance customer, I can't raise a problem report. I'd be grateful, therefore, if an MC could flag it up to PTC. Thanks.

Stuart

F

Fred_Kohlhepp

23-Emerald I

StuartBruff wrote:

I was looking at some of the Rosetta Code challenges for implementation in Mathcad and decided to try a couple in Prime Express (as this is what I have). The major drawback to Express is the lack of programming capability, but there's quite a bit that can be (or should be able to be done) using just range variables and functions that don't make use of programs.

I wanted to pick all of the non-zero elements out of a vector and decided to use the until function to sweep through a reverse sorted vector until it hit a zero. However, Prime keeps telling me that 'until' is undefined, even though I inserted it using the Functions Dialog. Express doesn't raise an alert saying that it's Premium Feature, so I'd expect it work. Neither can I find anything in the Migration Guide or Help to indicate it's been deprecated. Is there a known problem with the until function in Prime 3.1?

Stuart

I've got Prime 3.0--we're being dragged "kicking and screaming,"

Prime does not have "until" in the programming operator menu.

S

StuartBruff

Author

23-Emerald V

Fred Kohlhepp wrote:

Prime does not have "until" in the programming operator menu.

As Norm says, it's in the Piecewise functions section. It dates back to the Good Old Days, when programming was but a distant gleam in Mathsoft's eyes and all we had was iteration. I still find it a useful exercise, now and again, to try to refashion a problem in terms of pure iteration. Usually it just makes my brain hurt and curse the limitations of Prime Express ... although not sufficiently to make me want to shell out £1000 for Prime 3.1 (not when I get Personal Use editions of Mathematica, Maple and Matlab for less than £200).

Here's the example from the Mathcad 6 User Guide (*), duly written up and working in Mathcad 15 ...

I suspect the M6 difference in N2 is to deal with the extra zero that Luc noted in M11.

I've got Prime 3.0--we're being dragged "kicking and screaming,"

You would have thought they'd at least drag you to Prime 3.1.

Stuart

(*) Yes, I've still got it! Every now and again, my wife tries to get me to throw away a book or two, on the basis that I either "don't need that many" or "you've got a newer one" or "surely you only need one book on xxxx". Ha Ha! I gloat! (but only because she's not around at the moment). This is just why you can never have too many books or dispose of old ones! It more than justifies the extra expense of wall-length book cases! But you try telling some people that ....

S

StuartBruff

Author

23-Emerald V

LucMeekes wrote:

But, apart from the extra iteration, in Mcad 6 the until function works as expected 'stand-alone':

Luc

I doubt whether we'll see it fixed in M15, but it would be nice to make it work in Prime 3.2 / Prime 4.0 (whatever, as long as it's the next iteration).

Of course, it would be really cool if they made it work with a vector ...

Even better would be if they, additionally, made better use of ranges and allowed them as arguments to functions, plus extending the concept to sequences, ie, something along the lines of (and being very loose with notation) ...

99| 3-2i | "a" | [1,2,3] | sin

1,2 | "a","c"

1,2,3 | 1,2,4 | "a","c",

<sequence>:=(<sequence>)

(1,2,4)

<range>:=<term>..<term> [<term> in this instance would be limited to types that have a natural successor or predecessor, eg numbers and characters]

1..2 | "a".."b"

1,2..5 | "a","b".."e"

1,2..5 | "a","b".."e"

(1,2,4..32),9,13 | ("a","b".."e"),5,11

<sequence>:=<sequence>,(<range>)

"a","b",(1,2..5) | 5,11,("a","b".."e")

"a","b",(1,2..5),9,13 | 5,11,("a","b".."e"),5,11

Another provisio I'd put on ranges is that the number of terms in a multiterm initial sequence determines the order of the polynomial that fits the sequence and that the final term represents the final value of the implied sequence (as it currently does). The reason for this extension I've had a few occasions when I would have liked to have non-contiguous values in a range definition, generate alphabetic sequences, generate a simple polynomial set of values. A typical example of usage is using my vec function to generate a set of values or wanting to pass multiple range definitions to a function without having to generate a corresponding set of range variables; for example: in Mathcad 15 (and M11) I can use vec(2,4..8) to create the vector [2,4,6,8] directly; I would like to call a function g(x,y) with g(0.1,0.2, 0.4..3) and have g iterate over the (quadratically implied) values 0.1,0.1,0.2,0.4,0.7,1.1,1.6,2.2 and 2.9; or have a character sequence, "a","b","d".."z" map to "a","b","c","g","k","p","v".

Whilst there is an argument that range variables can be done away and replaced with vectors, with a range definition then acting as an implied array creator, I think they still have their uses as a separate implied iterator. For one thing, they take up less memory (I've had several instances where I've run out of memory using vectors but had no problem with range variables or loops

Well, something like that, anyway.

The strange thing is that some of this (sequences of values mixed with implied linear sequence (..) range definitions) can already be done in a for loop:

And it can partially be done in M15, plus a bit of programming to show the intent (including a polynomial function generator)

There are no "optical tricks" in the above, BTW; it's all done with standard programming.

Stuart

Sign up

Please use your PTC eSupport account.

Welcome to the PTC Community

Please use your PTC eSupport account.

Scanning file for viruses.

This file cannot be downloaded