Are these known bugs in the symbolic engine of Prime 4 and MC 15 M045?

[David_A]

I've been a long-time user of Mathcad (since before it was a Windows application -- I think it was version 2!) and long enjoyed the simplicity of it's interface.

PROBLEM 1 - Erroneous symbolic calculations (multiple examples)

I just bought MC Prime 4 and was testing some calculations in Prime 4 versus v15 and v13 and came across several disappointments -- errors in MC 15 and MC Prime 4 that MC 13 handles with aplomb. Perhaps I've made some mistakes, but I post this message for any helpful support understanding what's wrong from experienced folks who've been using the products regularly, hopefully to understand my mistakes, if I've made any.

Note that in EVERY symbolic calculation issue I outline below, I have no problem with the calculation in MC v13. All of these issues became apparent as I was reviewing the Applied Statistics extension pack for v13 and potentially converting some of the documents to Prime 4.

I've attached two documents to this post that detail the errors I've found, which so far are all with symbolic calculations.

Briefly, the errors are:

1) Errors in the symbolic calculation of the mean and standard deviation of the Poisson distribution. This error occurs in both MC 15 M045 and MC Prime 4.

2) Errors in the symbolic calculation of the mean and standard deviation of the Normal distribution. This error occurs in both MC 15 M045 and MC Prime 4.

3) Errors in the symbolic calculation of the mean and standard deviation of the Binomial distribution. This error occurs in both MC 15 M045 and MC Prime 4.

These oddities make me question the accuracy of the symbolic engine in both MC 15 and Prime 4, especially when I don't already know what the solution is.

PROBLEM 2 - Text formatting oddity? or misunderstood usage?

I've also noticed one other oddity. In preparing the Prime 4 document for posting, I found no way to highlight the text that precedes each example, say with blue text or a yellow highlight UNLESS the text was in a text block. Is there no way in Prime 4 to format text in a text box?

TIA

David

[LucMeekes]

David,

There are two reasons you see differences:

1. Mathcad 14, 15 and all versions of Prime use MuPad as symbolic engine, where Mathcad up to and including version 13 used Maple.

Personally I think Maple is better, only occasionally do I come across examples where the MuPad symbolic engine outperforms Maple.

As for presentation, I like MuPads better than Maple's

2. Prime has a long way to go before it becomes a real match for Mathcad 15 (let alone Mathcad 11). Fortunately customers still get Mathcad 15 for free when they buy a license for Prime. I have not seen any good reason yet to throw capital in the Prime money pit.

Hope this helps.

Luc

[David_A]

Thanks LucMeekes. What I gather is that Prime is not even as capable as MC 13. Boy do I regret buying Prime. I had MC 14 working fine and it is more capable than Prime is in every way that's important to me. The only issue was some graphics issues running on a 64-bit machine that they didn't resolve with MC14 patches.

Thanks

David

[mvenich]

Mathcad is great for design calculations and units handling (even Prime 4, I actually like this version)

But PTC's focus is not on the symbolic engine. The version of MuPad in Prime 4 is "MuPAD OEM kernel, version 3.2.1 (Mathsoft Kernel) © 2005 by SciFace Software GmbH & Co. KG." Will this version of MuPad be updated? Given that MuPad is now owned by Mathworks, it's doubtful.

[mvenich]

Do you think that PTC will update Mathcad Prime's symbolic engine?

[RichardJ]

The later versions are now owned by a competitor (The Mathworks). If PTC had any rights to them then presumably we would already have an upgrade.

[mvenich]

I love Mathcad, even Prime 4 is great for (a significant subset of) the work I do. But I'm now switching to other tools for symbolics and some specialized numeric routines. I simply cannot trust Mathcad's symbolics

[LucMeekes]

"If PTC had any rights to them then presumably we would already have an upgrade."

Can I extend that to Prime:

If PTC had any rights to it, we would already have a version that outperforms Mathcad 11...15......?

[mvenich]

Does anyone from PTC post here? If so, Could they comment? Is there a plan to improve symbolics?

[LucMeekes]

They react irregularly. No telling how often they peek.

Luc

[mvenich]

All the old regulars from PTC/Mathsoft are no longer involved in Mathcad:

John Sheehan left PTC

Mona Zetfel left PTC

Brent Edmonds has moved to another part of PTC

Lesley Bondaryk left PTC

Who at PTC is responsible for Mathcad development strategy? PTC: let us know!!

[Fred_Kohlhepp]

Martin Venich wrote:

Does anyone from PTC post here? If so, Could they comment? Is there a plan to improve symbolics?

Of course there's a plan to improve symbolics.  And Prime 4.0 will have a new plotting engine so graphs are better.  (Wait, make that later.)

And prime 3.0   3.1   4.0   5.0 will have all the functionality of Mathcad 15.

[David_A]

Based on the responses to this query and what I've seen from the little research I've done, I think we should brace ourselves - Mathcad will NEVER be the top notch "blackboard" calculation and development engine it once was nor is there ANY evidence that PTC has the intention to make it so (we're already at version 4 and it can't touch MC 15, a 5+-year old product). My guess is MC Prime will never match the symbolic capabilities of MC15, further, I suspect that the symbolics capabilities of Prime will slowly be eliminated from the program in favor of the numerics engine, which I believe serves as the calculational "glue" for other PTC products, hence all the focus on MC Prime's integration with those products. And that's a fine business decision for PTC to make; I'm sure someone got a promotion out of making the business case for this direction of MC Prime. But what doesn't seem to make sense is who they think will use MC Prime. I've always thought it's the folks coming out of college that drive new software and hardware and how they are used; isn't that how Apple grew so large? Sure you could say it's corporations that are the target audience, and that would be right, but they are hiring young engineers out of college, and if those engineers don't know MC Prime, they're likely to gravitate toward products they know and used in college - which ISN'T MC Prime. Of course I could be wrong in my assessment as I have made some significant guesses that may turn out otherwise. However, in any case it seems apparent that if you're interested in the pure math capabilities of the product once known as Mathcad, you should start looking for another solution. MC Prime is/will NOT be a significant math development or research tool; it's a big expensive calculator. It's doubtful that I will invest any more money into the sinking ship called MC Prime. Better to look toward Maple 2016 or Mathematica. Pity. What a loss. If anyone thinks I've got this all wrong, please educate me.

Thanks

David

[mvenich]

I'm using Mathcad Prime 4/Mathcad 15 and Maple 2016 side by side. A lot of my legacy work is in Mathcad, so I still use Mathcad regularly, but I'm transitioning some stuff slowly to Maple (anything that involves symbolics or intensive numerics). Still prefer Mathcad's units.

[David_A]

What do you like about Maple, and what don't you like about it? Which edition of Maple 2016 do you use?

[mvenich]

my company paid for one commercial Maple license. I love Mathcad as a quick calculation scratchpad, but we have no confidence in Mathcad's future (PTC are focused on CREO, not calculation tools), so we wanted to try something else. Evaluated Mathematica as well, but hated the syntax.

What I like about Maple

• Symbolics (goes without saying that it's far deeper than Mathcad)
• signal processing (spectrograms look absolutely beautiful) and optimization
• plots look much nicer than Mathcad (many more plot types, far more flexibility).
• fluid properties
• I had a question about a signal processing function and ended up in an email conversation with the actual developer
• there's a free global optimization addon (I used it to fit a two-diode model to experimental data, much less hassle than Mathcad)
• 2d math looks nice and crisp on a 4k laptop

What I don't like about Maple

• Maple's units an only be used for basic calculations (can't use them in numeric equation solving like in Mathcad)
• Syntax is more complex, but no more complicated than MATLAB or python
• I have to get used to Maple's way of entering math (I've used Mathcad for many years so am used to the Mathcad way of entering equations)
• no free form placement of equations or text (more like microsoft word than a whiteboard)