ACCEPTED SOLUTION
Accepted Solutions
Antoine Aula wrote:
Hi everyone,
I would like to know if there's a way to use a reference within a for loop in Mathcad 15. The subject has already been discussed here but the proposed solution wasn't really working well and gets too complicated with a bigger worksheet. Nevertheless, the principle is there.
To schematize, i'm trying to create a worksheet like this :
for i from 1 to n
* define variables used in reference calculation
* call reference
* get results(i) from reference
next i
show results
I hope this is technically possible, so please let me know if you know an advanced version of the pre-cited method.
In short, "No".
As you've probably found out from the previous threads, there isn't a built-in way to reliably loop over a general "reference" worksheet. All of the workarounds involve playing around with components to gain access to the automation interface.
One possibility is to rewrite your referenced worksheet in terms of range variables and define the range variables (plus any other application-dependent variables) in the calling worksheet, as per the simple example below.
Referenced Worksheet:
Calling Worksheet:
But this is rather messy and doesn't make full use of the power of Mathcad.
Needless to say (but I'll say it anyway) there is a Very Long Standing Feature Request to have this kind of capability built-in to Mathcad.
Stuart
Antoine Aula wrote:
Hi everyone,
I would like to know if there's a way to use a reference within a for loop in Mathcad 15. The subject has already been discussed here but the proposed solution wasn't really working well and gets too complicated with a bigger worksheet. Nevertheless, the principle is there.
To schematize, i'm trying to create a worksheet like this :
for i from 1 to n
* define variables used in reference calculation
* call reference
* get results(i) from reference
next i
show results
I hope this is technically possible, so please let me know if you know an advanced version of the pre-cited method.
In short, "No".
As you've probably found out from the previous threads, there isn't a built-in way to reliably loop over a general "reference" worksheet. All of the workarounds involve playing around with components to gain access to the automation interface.
One possibility is to rewrite your referenced worksheet in terms of range variables and define the range variables (plus any other application-dependent variables) in the calling worksheet, as per the simple example below.
Referenced Worksheet:
Calling Worksheet:
But this is rather messy and doesn't make full use of the power of Mathcad.
Needless to say (but I'll say it anyway) there is a Very Long Standing Feature Request to have this kind of capability built-in to Mathcad.
Stuart
Richard Jackson wrote:
The good news is that I had members of the PTC development team agree with me that adding the capability for regions to have inputs and outputs, and for them to be able to act as functions, would be a very useful addition to Mathcad. I don't know if this ever actually made it onto a to-do list though, or how many decades we will have to wait to see it even if it did.
Good Job! It's nice to know the idea's in place, even if it might take some time to get implemented.
Is there any chance of making this a multi-generational aspiration for the descendents of the development team? 
When you next speak to them, it might be worthwhile sowing the seed of making each region (more obviously) an object, with the ability to get at its attributes and methods programmatically as well as via a suitable dialog ... I think I suggested this back when Pangaea was still forming ... I think this as far back as PTC DeLorean thread chaser will go, though: Re: Mathcad worksheet as a function, https://community.ptc.com/t5/PTC-Mathcad/MathCAD-Prime-1-0-is-a-step-back/m-p/120871#159782%23159782%23159782 and Re: Format a Solve Block.
Stuart
Richard Jackson wrote:
Is there any chance of making this a multi-generational aspiration for the descendents of the development team?
Sort of like Cathedral building was in Europe in the middle ages?
That's exactly what I was thinking, or a multi-generational starship that builds up high tau factors - hmm, so 1 year of ship time could be millenia of "real" time and, of course, you get the same result in the presence of strongly curved space-time ... so maybe we're doing them a disservice and there's a black hole at the centre of PTC in which they're working round-the-clock and kicking out daily (ship-time) versions of Prime, little realizing that decades have passed in User Time?
"Son, my father started the immense project of adding inputs and outputs to regions, and I've dedicated my whole life to it. I doubt I will see it's completion in my lifetime, but you will follow after me and hopefully bring this masterwork to fruition. The experience this will bring to you will prove invaluable, as you will then embark on the even greater project of Multi-Dimensional Arrays. The magnitude and importance of the MDA project will be greater than anything in recorded history, and may take centuries to complete. You should be proud to be a founder of such a project though, because it will change the world. Just think of it. Three or more array subscripts!".
I'm just imaging the tau factor necessary to get MDAs in place before the Sun decides to expand its operations into Earth orbit ... there's a nice little question for Mathcad: How fast will such a spaceship be going when it achieves sufficient mass that it gains its own event horizon?
Actually, I think it might be more prudent if I were to adopt some diplomatic(*) skills as it appears that not only do we PTC Engineers who have taken on-board the idea of regions as objects with exposed properties and methods, but we may have an adept in the art of multi-dimensional array indexing, ie Mr Sheehan of this Parish (who also seems to share my predilection for mangling the Bard's words!). See the following article: <a href=" />
Couldn’t agree more Stuart. In a former life I did OLAP and everything was in a cube. I also agree that nested matrices is a poor substitution. Many of our customers are still using RDBMs. I think you are implying 3D databases are growing in popularity ?
So maybe we should start writing to PTC management telling them what a top notch bloke Sheehan is and why they should implement his ideas in Prime 3.2? 
Or perhaps Brutus might have the apposite message to pass ...
Stuart
(*) OTOH, maybe I ought to leave the diplomacy to those best versed in it. One of my squadron commanders wryly noted that in my version of the Oxford English Dictionary the word "diplomacy" was synomous with the word "gunboat", and "tact" meant "nailed to the wall".
Richard Jackson wrote:
No. Each package has it's own strengths. Given the choice of putting functions in a separate file (or in Matlab's case many separate files, because if you put two functions in one file you can only reference the first one) and then having a non-working script because it gets separated from the necessary files, or doing some copy and paste to put everything in one file, IMO the latter is a much better option.
I would like the option of functions in a separate file if it were implemented with one feature that Prime does have that Mathcad 15 does not. In Prime you have the option of embedding a copy of a referenced file. If the worksheet can't find the referenced file it uses the embedded copy. So Prime actually has the best of both worlds (or at least is closer to it than either MC 15 or Matlab). You can define functions in the same worksheet if you choose. You can define them in separate worksheets if you choose, and reference those worksheets. If the main sheet gets separated from the referenced sheets it still works. I don't use Prime though, because of all it's other limitations.
It is definitely good to have the referenced worksheets embedded in a worksheet. Perhaps, though, what's also required is something that Mathcad <=15 almost had (in the form of Extension packs), and that is a proper library configuration tool for adding worksheets (and user DLLs) to an appropriate "Library" folder, with the ability to add folders to the "Path". I think it would have to use a library tool, rather than just dump worksheets in a folder, to eliminate the no-doubt lengthy search through hundreds of worksheets to find a particular function,
It would also aid the user choosing the right version of function f from the hundreds of versions in the library. 
I have a vague recollection, from before the cut-off date for transferring threads from the Collaboratory to the Community, of suggesting a more object-oriented version of Mathcad, one aspect of which would be the ability to declare worksheets to be objects and then be able to pick a particular version of a function by referencing its worksheet (eg, something like wks:="Help Me.xmcd" and to use, say, function f(x,y) from this worksheet you'd write wks.f(2,3) or just f(5) if you wanted to use the single argument version in your worksheet itself).
Stuart
