| Chronology | Current Month | Current Thread | Current Date |
| [Year List] [Month List (current year)] | [Date Index] [Thread Index] | [Thread Prev] [Thread Next] | [Date Prev] [Date Next] |
At 10:37 AM -0600 12/8/98, John Trammell wrote:
Compiled versus interpreted only makes a difference in the coding
cycle. Interpreted means a faster cycle, in my opinion. Most
interpreted languages that I deal with lack sin(), log(), etc., so
that leaves Perl. Quick and Dirty.
Well, I am the Man. :-)
You are "the Man" but I think this is a little misleading. It is my
understanding that an interpreted language is one where each line of code
is changed into machine language and executed one line at a time. A
complied language is one where the entire program is turned into machine
language first before being executed. A complied program is therefore going
to execute faster than an interpreted program since the overhead of
changing into machine code is absent after the first go-round. Compliers
are also made to be 'smart'; they can pass through the program several
times looking for ways to make the code more efficient before changing into
machine language. For most applications, given the speed of current
computers, this speed up is not noticable. If you are doing any serious
number crunching, however....
Speed of execution is also one reason FORTRAN (a relatively low level,
complied language) is still around. Other languages (such as C/C++) are
writen at a "higher level" and have been developed to have more built-in
functionality (a computer science term- I didn't invent it) at the cost of
having a little slower execution time. But C/C++ complilers are catching
up- you can see a discussion of a speed comparison between FORTRAN and
C/C++ done on several platforms this past year in Computers in Physics. I
gather from the article that, if you avoid some of the special built in
features of C/C++ you can write code that executes nearly as fast as
FORTRAN.
kyle