computers!!
Nov. 3rd, 2008 05:44 am![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
goneroundthebendWell...maybe not computers...maybe the person trying to write the programs.
Just as a mental exercise, I tried to re-create this computation method a former boss of mine published back in the 1965, where he outlined a "simple" method for calculating characteristic lines in supersonic flow. Let's see...at that time it could have been punch cards, but FORTRAN had been hobbled together just a few years before, there was no network, and the IBM/360 was barely getting going.
More likely it was done on a desktop calculator--and desktop was the description of the size, not the location.
Ok...so...as an intellectual exercise, I decide to try to duplicate the results. Mac G4 laptop, 1Gb memory, G95 at the ready...more power than a million of those paltry calculators...piece-o-cake.
The publication had a rough outline of the solution methodology along with some equations that I pulled together in a few hours. A couple of function calls and a 4x4 matrix solve subroutine later and I'm iterating away. Err...well...kind of...that is if one can call 1 iteration iterating.
Three days later I still couldn't get the program to produce ANYthing close to what he had published over 40 years ago. Something was going wrong on the first pass through the solution space. NaN's have a way of making life...problematic in the computational world.
Maybe it really is rocket science...oh no...I've flunked... ~_~;;
Just as a mental exercise, I tried to re-create this computation method a former boss of mine published back in the 1965, where he outlined a "simple" method for calculating characteristic lines in supersonic flow. Let's see...at that time it could have been punch cards, but FORTRAN had been hobbled together just a few years before, there was no network, and the IBM/360 was barely getting going.
More likely it was done on a desktop calculator--and desktop was the description of the size, not the location.
Ok...so...as an intellectual exercise, I decide to try to duplicate the results. Mac G4 laptop, 1Gb memory, G95 at the ready...more power than a million of those paltry calculators...piece-o-cake.
The publication had a rough outline of the solution methodology along with some equations that I pulled together in a few hours. A couple of function calls and a 4x4 matrix solve subroutine later and I'm iterating away. Err...well...kind of...that is if one can call 1 iteration iterating.
Three days later I still couldn't get the program to produce ANYthing close to what he had published over 40 years ago. Something was going wrong on the first pass through the solution space. NaN's have a way of making life...problematic in the computational world.
Maybe it really is rocket science...oh no...I've flunked... ~_~;;
