Anthony Scopatz

I think, therefore I amino acid.

The (Real) Grand Challenge

Having spent the past couple of days a the Fuel Cycle Research and Development meeting in Washington DC (Gaithersburg, MD really), there has been a lot of talk about the “Fuel Cycle Simulator Grand Challenge.”

On the one hand this is very, very exciting.  It is an opportunity for the DoE to replace a system that doesn’t work.  Or at the very least, it is a workflow that does not meet current and expected needs.  I have been saying this for years…

…Another thing I have been saying for years is that, ‘Good science is good software development.’  Not everyone believes me; fine.  But if you are going to build out a huge code infrastructure you may as well give it the old college try.

I am fearful though.  They are calling it a ‘Grand Challenge’ because there exists this permeating belief that:

  1. Nuclear engineers are not software developers.
  2. This problem is hard.
  3. This problem is new.

At the heart of the proposed simulator is classic framework problem that is easily solved by a Model-View-Controller (MVC) approach with good test-driven development around the physical models.  And they are far from the first scientists to have a issue with storing and visualizing large amounts of data.

I am fearful because if they choose the wrong API now, it will have effects for years and years to come.  (Thankfully, from experience, they seem to recognize this.)

I am also fearful because they are talking about a three-year time frame just to build the framework and maybe some sample physics models [1]. This is the easy part!  And every day that you spend debating whether you should be passing an array or a pointer to an array is a day that you are not doing new science and engineering [2].

So I am excited because it truly is an excellent idea.  I am just not sure that it is as VISIONary as they make it out to be.  Give me three months, a coffee-shop, and  BAM you’ll have the architecture that has been proposed.

The real grand challenge is making nuclear engineers into become software developers.

[1]: It may well take three years to put in meaningful models, but that is science not infrastructure.

[2]: I count views as new science, because without the shiny figure at the end, all you have is a bunch of numbers.

Comments