• C++ Programming for Financial Engineering
    Highly recommended by thousands of MFE students. Covers essential C++ topics with applications to financial engineering. Learn more Join!
    Python for Finance with Intro to Data Science
    Gain practical understanding of Python to read, understand, and write professional Python code for your first day on the job. Learn more Join!
    An Intuition-Based Options Primer for FE
    Ideal for entry level positions interviews and graduate studies, specializing in options trading arbitrage and options valuation models. Learn more Join!

The World according to Americans

In production, yes, As student, doing it from scratch is better. And it teaches you algebraic manipulation..

Now, use LU in combination with ADI solver for option pricer (btw there is hardly any ADI code on the web).
I don't see the point beyond use as a coding exercise, unless you are a researcher in the field or your job is to develop such applications (e.g. you are a developer for Intel MKL).

And even as a coding exercise, you are going to have to be more explicit with why it's better for students. What does doing it even teach? I've done it (implemented LU) myself before, and all I really remember was that I had headaches doing the bookkeeping to make sure my indices were right.
 
Much of the time, this type of superficial acquaintance is all that is needed to get a specific job done. I needed to quickly put up a web server yesterday with no frills, and I got it done by looking up tutorials.

Did I learn the ins and outs of webhosting? No. Did I need a deep understanding to get my task done? No.

Similarly, why program an LU solver when many more qualified people have spent significant amounts of time doing the same thing... and in fact have probably optimized it beyond what you would ever be capable of? I need to solve this set of equations, NOW, not a day later when I will have finished coding up and debugging a slow version of what's already out there...

I think these are 2 very different tasks. IMHO, using LU to solve a problem without knowing the ins and outs is ok since you know what the goal is and LU it's a very specific solution.

OTOH, putting a webserver up might be fine in a short run but it usually leads to greater issues later on. That's why knowing the in and outs might be a better solution.

The Playstation breach started as simply as putting a web server up.
 
OTOH, putting a webserver up might be fine in a short run but it usually leads to greater issues later on. That's why knowing the in and outs might be a better solution.

The Playstation breach started as simply as putting a web server up.
The bounds of my particular task do not require I keep a server running over the long haul. I'm not setting up a production server of any kind, so there's really no need to do production-grade learning.
 
The bounds of my particular task do not require I keep a server running over the long haul. I'm not setting up a production server of any kind, so there's really no need to do production-grade learning.
This is just the beginning. You are using it. You will show it to somebody. He/she will find value and start using it. In no time, your colleagues are using it and their boss has an idea to make visible and sell this service... you get the drill, right?
 
This is just the beginning. You are using it. You will show it to somebody. He/she will find value and start using it. In no time, your colleagues are using it and their boss has an idea to make visible and sell this service... you get the drill, right?
And pretty soon, you will find yourself learning about caching mechanism, APC, memcache, how to use CDN to speed up your pageload and tons of other things to optimize your website.
Wait, this sounds familiar.
 
Hm, while all of this is going on I'm thinking about the degree to which Americans don't even know their own country. For example, I've seen what I think of as a fair amount of the country but can count 17 states in which I've don't believe I've even spent the night:

Hawaii, Alaska, Oregon, Idaho, Montana, North Dakota, South Dakota, Colorado, Nebraska, Oklahoma, Arkansas, Iowa, Wisconsin, West Virginia, New Hampshire, Rhode Island, Delaware (although I've missed Delaware by less than 2 miles more than a dozen times in my life)

It seems like a substantial but not enormous list...until you realize that the combined total land area of just those states is nearly a million square kilometers larger than all of India....
 
Speaking of US geography, everyone should visit Mesa Verde at some point (especially if you have kids). It is the only national park that preserves"man-made objects". It is in the middle of nowhere though, about a 6 hour drive from Grand Canyon.
 
I would love to visit Mesa Verde, as I've heard it's amazing, but it's a gauge of just how incredible our national parks system is that it probably does not fall in the top 10 among those I haven't seen yet but really want to.
 
I would love to visit Mesa Verde, as I've heard it's amazing, but it's a gauge of just how incredible our national parks system is that it probably does not fall in the top 10 among those I haven't seen yet but really want to.

Take your kid when she is a little older.... I bet it will become your top 10 once you have visited it!
 
And even as a coding exercise, you are going to have to be more explicit with why it's better for students. What does doing it even teach? I've done it (implemented LU) myself before, and all I really remember was that I had headaches doing the bookkeeping to make sure my indices were right.

Exactly! if you can't do LU then ADI will be impossible.
 
Only on Quantnet that a discussion which begins with a silly map and a silly title ends up talking about LU, ADI solver, server setup, Playstation security breach.
And these are the Americans I'm most identified with ;)

The other scenario is starting with ADI and ending with a C++ vs Java war :D
 
Where do we draw the line between these activities? What if the LU code you get is buggy or you want to adapt it to Neumann boundary conditions? Then the numerical recipes break down and then what?

I know! Did you know some people just buy pre-made bikes and cars? If you never get the metal, glass, plastic, and other raw materials to build a vehicle how do you expect to be a good driver? What are you going to do when your car breaks down, find someone else to fix it? Next you'll tell me you eat food that you don't grow on your own...
 
I know! Did you know some people just buy pre-made bikes and cars? If you never get the metal, glass, plastic, and other raw materials to build a vehicle how do you expect to be a good driver? What are you going to do when your car breaks down, find someone else to fix it? Next you'll tell me you eat food that you don't grow on your own...

I got it at the 1st line;)

Unfortunatey, the evolution of software is lagging a few years behind other disciplines. Now let's move up the chain from LU. How many implementations of HW model are there in the world?

What's the difference between bicycles and software? The wheels are square:)

20 years ago there were maybe 300 s/w companies developing tools for AutoCAD in Europe. Now a lot less.
 
Only on Quantnet that a discussion which begins with a silly map and a silly title ends up talking about LU, ADI solver, server setup, Playstation security breach.
And these are the Americans I'm most identified with ;)

silly ? Come on Andy ! I know your kidding. A ton of people i work with can relate to that map.
 
I know! Did you know some people just buy pre-made bikes and cars? If you never get the metal, glass, plastic, and other raw materials to build a vehicle how do you expect to be a good driver? What are you going to do when your car breaks down, find someone else to fix it? Next you'll tell me you eat food that you don't grow on your own...

Heard on the grapevine this weekend :D

Thanks a lot for your help Cuchulainn,
I wrote a beautiful VBA code this weekend, works fantastic!!! I had to refresh my algebra - LU Decop. but it feels much better than just copying code and using it.
 
Back
Top