Talk:Mathcad

From Wikipedia, the free encyclopedia

A "history" section may be needed, including listing the version by date. File Not Found 02:31, 8 February 2006 (UTC)

This article is a bit too purple and advertising. Could someone edit it to be more professional (less marketing) and spell out history, theory, strengths, weaknesses, alternatives? 131.151.44.151 21:51, 30 May 2007 (UTC)

Look the Ukrainian version. Maybe, somebody will translate?--Leon II 18:38, 2 June 2007 (UTC)

Contents 1 a Waterloo Maple component ? 2 Criticism 3 Free Software Alternatives 4 Solve Blocks -- one of MathCAD's very beneficial features 5 Memory Management / Internal Errors

[edit] a Waterloo Maple component ?

My copy of V7 pro has a whole directory of 'maple' stuff (and without that, it doesn't do computer algebra - not under wine anyway :-)... and the bottom of the box says 'MKM (Mathsoft Kernel Maple) copyright 1994 Waterloo Maple Software' Linuxlad 11:25, 12 April 2006 (UTC)

[edit] Criticism

a) the proprietary file format, especially up to Mathcat 2001i. From my point of view, even the turn to an XML format did not change much. The same situation like with Labview. b) the need for activation. Mathcad 2001i is the LAST version of Mathcad which can be freely installed and used. hemmerling 16:51, 2 October 2007 (UTC)

[edit] Free Software Alternatives

Can anyone provide any free software alternatives to MatchCad, that have a similar feature set and that are like OpenOffice.org is to Microsoft Office? -Mardus 14:56, 4 October 2006 (UTC)

Look at PARI/GP, Maxima, and Yacas. See also List of computer algebra systems.--209.208.77.219 03:32, 4 December 2006 (UTC)

Look at scilab.

[edit] Solve Blocks -- one of MathCAD's very beneficial features

MathCAD has a feature known as solve blocks. The structure is very simple: Given ... variable-array:=minerr(variable-array) In short, and in English, "Given that I want my utopian dreamworld to be like this, please vary variables a,b,c,... until you get as close as you can to making my dreamworld come true." In less poetic language, yes this is linear programming. But such a feature was not mentioned in the article. Moreover, YOU don't have to figure out how to do linear programming, the algorithm (and syntax) is already built in.

But besides linear programming as just another feature to list, the immense scope of MathCAD's minerr solve blocks deserves mention. Absolutely ANY SET OF DESIRES WHATSOEVER -- so long as they are mathematically expressible -- can form a solve block. Great for mathematical modeling -- yes very definitely. But just to show how absolutely general the solve block is, I wanted to know how much I used each of my 9 or so lenses for my Canon SLR. I knew I used my two 24mm lenses about 3.5 times as much as my 28mm, the Phoenix 24 twice as much as my Canon 24 (because of its closer focus) and so on and so on and so on. So I said it all, along with the sum being 100% and had MathCAD solve it. Then I did the slight renormalization from the solution adding up to 99.2%. But the point is ANYTHING is ABSOLUTELY EASILY minerr solve block solvable in MathCAD. Literally as easy as typing the word "Given" then typing what you wish to be then names for the few things you were willing to vary and then those variables after the minerr( .

But yet another use of the minerr solve block is to compute inverse functions. This is trickier, but is actually very little typing. What you end up doing is DEFINING a function (the inverse function) as a solve block. Thus the inverse function is like an implicit "goto subroutine" statement. Which means if you plot that inverse function for 1,000 points, the minerr solve block "subroutine" will have to run 1,000 times.

Anyway, not having discussion of solve blocks is a serious omission and misses an important part of the MathCAD character. —Preceding unsigned comment added by 199.196.144.16 (talk) 21:17, 15 April 2008 (UTC)

An excellent point. I would add that Mathcad has several "ending functions" that can be used in solve blocks. minerr is just one of those functions. Personally, I use the "odesolve" and "find" functions quite a bit. As to the "serious omission" in the Mathcad article, be bold (WP:BB), and write something! Noca2plus (talk) 22:54, 15 April 2008 (UTC)

[edit] Memory Management / Internal Errors

I removed the following good-faith edit[1] by 81.138.36.173 (talk):

Although powerful, Mathcad suffers memory management issues as memory, once allocated, is never de-allocated. The consequences of this are that sometimes handling of very large matrices (e.g. high frequency time-series datasets) can cause non-descript errors, for example "An internal error has occured" after which there may be no observed effect through to abortion of the program without option to save or recover work. Some users of Mathcad have deemed such unexplained fatal crashes 'Mathcadisms'. [2]

In my experience, I've also seen problems with Mathcad de-allocating memory (especially after program errors). However, it seems unlikely that Mathcad lacks all ability to de-allocate memory. I think the article should point out Mathcad's apparently poor memory management, but that needs to be backed up with a verifiable citation (WP:V) and not original research (WP:NOR). The offerred reference (a group of forum postings -- WP:SPS) certainly documents that "internal errors" occur. But none of the postings suggested that this was due to memory management issues, or that the internal errors resulted in loss of data. Quite the opposite, several posters indicated that they did have the opportunity to save their work after the error. Futher, no posts mention the neologism "Mathcadisms". Noca2plus (talk) 19:27, 18 April 2008 (UTC)