I was the in the last class of my engineering program to be offered a FORTRAN class. However, the whole aerospace industry is dominated by FORTRAN... in fact, I'm writing some code in FORTRAN 77 today... to calculate spacecraft orbits no less...
I think there is a definite need for a high level language with strong error handling. Stuff like automatic checking for buffer overflows on all input and output, automatic checking of data types on every instance where a variable is used - and built in units checking, i.e., a formula needing inputs of feet will automatically convert from meters and not allow you to input something of units of pounds or seconds (only units of length).
I'm sure there's lots more error checking that could be built into a language that would be a bitch to get through the compiler or interpreter, but would make a program more trusted in the end.
We could call it "anal"...
Re:Aerobraking and probe intelligence...
on
The Art of Aerobraking
·
· Score: 2, Interesting
The problem isn't really about computing power. The computing power needed is akin to the cruise controller on our car or the temperature controller on your refrigerator.
The difficulty is in getting a sensor that can tell the controller what the air density is. The obvious way to do this would be with an accelerometer but these are expensive and don't really work well enough anyway (with current technology). Wing-like control surfaces would also help -- but these add weight and cost.
Aerobraking and Aerocapture hold a lot of promise as most of the fuel carried by these probes is used to go into orbit at Mars (or your favorite celestial body).
Putting Hydrogen tanks on aircraft would make them very large and increase drag on the aircraft. There's been a lot of research into better storage methods for hydrogen for use in cars.
From The Hydrogen and Fuel Cell letter:
In storage, DoE funded 21 projects. Many were discontinued:
magnesium hydrides (Florida Solar Energy Center, Energy Conversion Devices, Texas A&M), fullerenes (Oak Ridge National Laboratory), chemical hydrides (NJ Transportation Department), hydride slurry (Tecogen), phase change hydride (A.D. Little), non classical polyhydrides (Hawaii Natural Energy Institute), carbon foam (Sandia), and carbon aerogels (Lawrence Livermore). Others are continuing, such as fullerene hydride (MER), nanotubes (NREL), and some which seem promising such as liquid hydrogen (Lawrence Livermore - "completing certification of a 34.5 MPa tank for both liquid and gaseous hydrogen"), composite tanks (Lawrence Livermore, Thiokol, Johns Hopkins - "successful at up to 11.3% by weight"), and alanates (Hawaii Natural Energy Institute, Sandia - "successful at up to 5% by weight").
Of course, we could also develop ways to store jet fuel in airplanes that won't leak if a plane crashes.
Why not bill chip performance in terms of flops?
Give a minimum flop rating and a maximum flop rating... seems simple enough.
C++ not useful for numerical computing
on
Why not Ruby?
·
· Score: 1
C++ is absolutely horrible if you're doing anything numerically intensive. C++ is high overhead combined with slow execution time. FORTRAN is very good for numerical stuff, but it still has quite a bit of overhead in actually getting a program to work. Perl, Python, and MATLAB are good when you're willing to sacrifice execute time to get a program written quickly... and Ruby looks interesting for this as well...
I'm waiting for wireless broadband. Now that laptops don't cost a lot more than desktops I think mobility will be the new driver. Who wants to cable up their house? --especially when standards change every few years. Let's just get some low-flying satellite contellation to link up our laptops.... that way we could use them in our flying cars:)
They don't want anyone to know the spacecraft commanding language until after the spacecraft is done with its mission... ya know, to protect the spacecraft from mischief. Generally NASA software is public domain.
Slashdot Mirror
I was the in the last class of my engineering program to be offered a FORTRAN class. However, the whole aerospace industry is dominated by FORTRAN... in fact, I'm writing some code in FORTRAN 77 today... to calculate spacecraft orbits no less...
I think there is a definite need for a high level language with strong error handling. Stuff like automatic checking for buffer overflows on all input and output, automatic checking of data types on every instance where a variable is used - and built in units checking, i.e., a formula needing inputs of feet will automatically convert from meters and not allow you to input something of units of pounds or seconds (only units of length).
I'm sure there's lots more error checking that could be built into a language that would be a bitch to get through the compiler or interpreter, but would make a program more trusted in the end.
We could call it "anal"...
The problem isn't really about computing power. The computing power needed is akin to the cruise controller on our car or the temperature controller on your refrigerator.
The difficulty is in getting a sensor that can tell the controller what the air density is. The obvious way to do this would be with an accelerometer but these are expensive and don't really work well enough anyway (with current technology). Wing-like control surfaces would also help -- but these add weight and cost.
Aerobraking and Aerocapture hold a lot of promise as most of the fuel carried by these probes is used to go into orbit at Mars (or your favorite celestial body).
Putting Hydrogen tanks on aircraft would make them very large and increase drag on the aircraft. There's been a lot of research into better storage methods for hydrogen for use in cars.
From The Hydrogen and Fuel Cell letter:
In storage, DoE funded 21 projects. Many were discontinued:
magnesium hydrides (Florida Solar Energy Center, Energy Conversion Devices, Texas A&M), fullerenes (Oak Ridge National Laboratory), chemical hydrides (NJ Transportation Department), hydride slurry (Tecogen), phase change hydride (A.D. Little), non classical polyhydrides (Hawaii Natural Energy Institute), carbon foam (Sandia), and carbon aerogels (Lawrence Livermore). Others are continuing, such as fullerene hydride (MER), nanotubes (NREL), and some which seem promising such as liquid hydrogen (Lawrence Livermore - "completing certification of a 34.5 MPa tank for both liquid and gaseous hydrogen"), composite tanks (Lawrence Livermore, Thiokol, Johns Hopkins - "successful at up to 11.3% by weight"), and alanates (Hawaii Natural Energy Institute, Sandia - "successful at up to 5% by weight").
Of course, we could also develop ways to store jet fuel in airplanes that won't leak if a plane crashes.
Why not bill chip performance in terms of flops?
Give a minimum flop rating and a maximum flop rating... seems simple enough.
C++ is absolutely horrible if you're doing anything numerically intensive. C++ is high overhead combined with slow execution time. FORTRAN is very good for numerical stuff, but it still has quite a bit of overhead in actually getting a program to work. Perl, Python, and MATLAB are good when you're willing to sacrifice execute time to get a program written quickly... and Ruby looks interesting for this as well...
But, I wonder... can anyone tell me what this "stacking" process is?
I'm waiting for wireless broadband. Now that laptops don't cost a lot more than desktops I think mobility will be the new driver. Who wants to cable up their house? --especially when standards change every few years. Let's just get some low-flying satellite contellation to link up our laptops.... that way we could use them in our flying cars :)
They don't want anyone to know the spacecraft commanding language until after the spacecraft is done with its mission... ya know, to protect the spacecraft from mischief. Generally NASA software is public domain.