Excel Replacement?

[spooky]

..

[oedipus rex]

I've checked out R, and think it would be very useful for building complex models using sophisticated numerical techniques given its wide array of statistical packages. Here's something I found in the NAAJ on using a free tool called WinBUGS (Bayesian inference using Gibbs sampling) for actuarial models. http://www.soa.org/library/naaj/1997-09/naaj0104_7.pdf

[Obi-Wan Kenobi]

For general non-MS spreadsheet work (and word processing, and presentation), try Open Office (http://www.openoffice.org).

[oedipus rex]

Or Gnumeric

[spooky]

..

[oedipus rex]

There's a GNU version of Matlab called Octave (http://www.octave.org) that you might find to be cool, especially since it's about $1,800 cheaper than Matlab (free). It's generally considered to be the best of the open source Matlabs, and it's even compatible with Matlab. To get it working on your machine, you either need a Unix/Linux box, or you can get a Unix emulation layer for windows called Cygwin (http://www.cygwin.com).

[Incredible Hulctuary]

A Windows port of Octave is available at http://sourceforge.net/projects/matlinks .

[Brad Gile]

I've been using various versions of Mathcad since 1988. Incredible tool that can be used to solve damn near anything you can throw at it.

Brad

[Brad Gile]

I've been using various versions of Mathcad since 1988. Incredible tool that can be used to solve d*mn near anything you can throw at it.

Brad
ROTFLMAO! This forum has an automatic censor?! :D

Brad

[Moderator1]

I've been using various versions of Mathcad since 1988. Incredible tool that can be used to solve d*mn near anything you can throw at it.

Brad
ROTFLMAO! This forum has an automatic censor?! :D

Brad

D*mn straight we do. Watch your language, buster. :P

[The Drunken Actuary]

Having work with Excel for years, I am getting sick of it and I don't think the program has much to offer to increasingly stat-hungry users. Plus I don't like Micro$oft.Why don't you like Microsoft?

[StephenLL]

At AmRe we have been using Matlab for over a year now. It is an amayzing modeling and prototyping language. It's geared toward operations on matrices. There are many built-in functions that would take a year and a day to program in vba. FFT, convolutions, many random number generators. There are generally two ways to run it: 1st interactively and 2nd compiled into a dll for excel. Both produce nearly the same execution time, approximately 1/50 to 1/1000 of the time it takes to run in vba. We have a model that builds the aggregate distributions using fft and then runs that distribution through our cashflow modelling to price reinsurance treaty provisions.

On a P4 2.53ghz, 1 million gamma random numbers takes about .09 to .1 seconds to run.

If anyone wants detailed examples or to discuss matlab/octave further, please feel free to let me know. This product has brought my company into the 21st century in terms of modelling. We use excel and access for front ends but the behide the scenes power comes from matlab.

Stephen

There's a GNU version of Matlab called Octave (http://www.octave.org) that you might find to be cool, especially since it's about $1,800 cheaper than Matlab (free). It's generally considered to be the best of the open source Matlabs, and it's even compatible with Matlab. To get it working on your machine, you either need a Unix/Linux box, or you can get a Unix emulation layer for windows called Cygwin (http://www.cygwin.com).

[spooky]

..

[oedipus rex]

spooky: check out Loss Models 4.7 Inversion Methods, especially 4.7.1 FFT, pp 317-320.

[MathGuy]

There's a GNU version of Matlab called Octave (http://www.octave.org) that you might find to be cool, especially since it's about $1,800 cheaper than Matlab (free). I

Thanks rex, I like it...

I don't think Microsoft is doing a very good job at making innovative products...

Wait, you don't like Microsoft because they're not innovative enough for you, but you're more than willing to use a GNU rip-off of Matlab (which screams "innovation" to me)? Sounds to me that you're more concerned about the old pocketbook. Does anyone else hear a little birdie?

[Brad Gile]

There's a GNU version of Matlab called Octave (http://www.octave.org) that you might find to be cool, especially since it's about $1,800 cheaper than Matlab (free). I

Thanks rex, I like it...

I don't think Microsoft is doing a very good job at making innovative products...

Wait, you don't like Microsoft because they're not innovative enough for you, but you're more than willing to use a GNU rip-off of Matlab (which screams "innovation" to me)? Sounds to me that you're more concerned about the old pocketbook. Does anyone else hear a little birdie?

Oh, I hear that little sucker tweeting away! :D

Brad
"Karate ni sente nashi. There is no first strike in karate."-Gichin Funakoshi
"Neither is there a second!"-Brad Gile

[oedipus rex]

For me, the issue has nothing to do with likes or dislikes hate or emotion, it's about personal choice. Do you program and want access to the source code and the freedom (and responsibility of improving it) that comes with that and don't mind the absence of a user-firiendly GUI environment (though this may change if Mono matures)? Or are you an educational professional in a Third World country that can't afford licensing costs? Then you'll probably GNU or open source, otherwise Microsoft or some other vendor will be happy to take your money, which is the value you place on not having to deal with the ugly internals. Both seem equally valid to me.

[Dr T Non-Fan]

Company pays for it, so I could care even less than I do now. (I'd show that by not writing this.)

[Incredible Hulctuary]

Company pays for it, so I could care even less than I do now. (I'd show that by not writing this.)

That works if you have a company that is responsive to purchasing new products that their employees say are useful. Large insurance companies tend to have a lengthy and bureaucratic purchasing/approval processes, such that it is usually better to grab the free alternative now than wait for what could be weeks, months, or even a year because it wasn't approved in this year's budget, or perhaps never.

[Incredible Hulctuary]

Wait, you don't like Microsoft because they're not innovative enough for you, but you're more than willing to use a GNU rip-off of Matlab (which screams "innovation" to me)? Sounds to me that you're more concerned about the old pocketbook. Does anyone else hear a little birdie?

When you're paying hundreds or thousands of dollars for a product, then yes, you do have the right to expect innovation.

[Cho Da]

When you're paying hundreds or thousands of dollars for a product, then yes, you do have the right to expect innovation.Innovation? At this point I just wish they would deliver "trustworthy" with out a side of BSOD.

[Brad Gile]

When you're paying hundreds or thousands of dollars for a product, then yes, you do have the right to expect innovation.Innovation? At this point I just wish they would deliver "trustworthy" with out a side of BSOD.

Well, then, either find better products elsewhere or create new ones yourself and kwitcherbellyachin already! :viola: :D

Brad

[oedipus rex]

When you're paying hundreds or thousands of dollars for a product, then yes, you do have the right to expect innovation.

And let's not even mention security... :wink:

[Obi-Wan Kenobi]

What about security? :P

Seriously, I'm not entirely convinced that open source software is more secure. I know the normal song and dance, and I also know that the Linux kernel needs security patches too, and that those patches don't always get applied. Nor do the Apache ones, or others.

[Cho Da]

Excel isn't too much of a security threat if you turn off macros, or at least set the macro security to medium and only allow your own code to run.
If you want to be safe, you really need to unplug everything, especially the power. :roll:

[oedipus rex]

Here's the financial engineering take on this topic: http://www.wilmott.com/310/messageview.cfm?catid=10&threadid=1931

[Brad Gile]

Here's the financial engineering take on this topic: http://www.wilmott.com/310/messageview.cfm?catid=10&threadid=1931

Very interesting, indeed. Aside from a few misconceptions about Excel (VBA is NOT a mere "scripting language"), that forum drives home a critical point: given any task, select the right tool for the task. If you need to put a screw into a piece of wood, don't use a hammer. Excel is a general tool and is "best in class" for a lot of things, but it is NOT the right tool for an awful lot of specialized tasks. Use something else if it will work better.

Brad
"Karate ni sente nashi. There is no first strike in karate."-Gichin Funakoshi
"Neither is there a second!"-Brad Gile

[oedipus rex]

I agree. VBA is a rich and powerful language that makes Excel quite extensible. Actually, if you use the powerful matrix manipulations hidden in Excel, it seems that it could be almost as powerul as any mathematical language, like Mathematica, Matlab, or Maple. I just don't know why the Excel engineers called them 'array' functions, when they really should be called matrix functions, since you can transpose matrices using them, take inverses, and do matrix multiplication.

[Wino In Training]

I agree. VBA is a rich and powerful language that makes Excel quite extensible. Actually, if you use the powerful matrix manipulations hidden in Excel, it seems that it could be almost as powerul as any mathematical language, like Mathematica, Matlab, or Maple. I just don't know why the Excel engineers called them 'array' functions, when they really should be called matrix functions, since you can transpose matrices using them, take inverses, and do matrix multiplication.

Although tools like Maple are matrix-oriented, I think you'd have one heckuva time using VBA to twist Excel into the kinds of analysis that Maple etc. are built for.

[oedipus rex]

true :P

[Brad Gile]

I agree. VBA is a rich and powerful language that makes Excel quite extensible. Actually, if you use the powerful matrix manipulations hidden in Excel, it seems that it could be almost as powerul as any mathematical language, like Mathematica, Matlab, or Maple. I just don't know why the Excel engineers called them 'array' functions, when they really should be called matrix functions, since you can transpose matrices using them, take inverses, and do matrix multiplication.

Although tools like Maple are matrix-oriented, I think you'd have one heckuva time using VBA to twist Excel into the kinds of analysis that Maple etc. are built for.
Absolutely. Like I said, use the right tool for the job. :D

Brad
"Karate ni sente nashi." There is no first strike in karate -Gichin Funakoshi
"Nor is there a second!"-Brad Gile