Ace128 has asked for the wisdom of the Perl Monks concerning the following question:

Hey,

I found this on slashdot, and although he/she doesnt mention Perl, I found the question interesting and was hoping some clever feedback from this lovely community! Comments on slashdot is kinda cluttered and doesnt seem to discuss the topic much...

I have been thinking about this too lately. Specially the part where computers becomes faster and faster...

I usually say that if you wanna have low level access or really need to take the advantage of the CPU (do some advanced math computation is the usuall example), do in native language. Or well, do it in C/C++ or simular where you control every byte... But as computer becomes faster and faster one may wonder if this in the end matters that much.

http://ask.slashdot.org/article.pl?sid=06/06/12/2044245

So, what do you folks say?

/ Ace

Replies are listed 'Best First'.
Re: End of Native Code?
by Old_Gray_Bear (Bishop) on Jun 15, 2006 at 16:22 UTC
    Dead languages that I maintain code in and that still have active communities:
    • COBOL
    • FORTRAN
    • MUMPS
    • IBM Basic Assembler and Macro Assembler
    • Pascal
    • Dartmouth Basic
    Hum... This sounds a lot like the stage directions in the first act of Hamlet -- "The Ghost Walks".

    ----
    I Go Back to Sleep, Now.

    OGB

[reply]

      Hey, Fortran isn't dead!

      It's just that the bulk of its user community is engineers and physical scientists, so the CS world thinks it's dead. Check out the codes in CFD, numerical relativity, global climate models, etc.

      Now, CLISTS and JCL are languages that deserve to be killed, buried, disinterred, chopped into very small pieces, put through a blender, incinerated, and reinterred in Yucca Mountain. Maybe along with APL.

      emc

      e(π√−1) = −1
[reply]
        And FORTRAN will never die, for that reason. It is too easy for the scientists and engineers to write their own data reduction software in it. Most other languages are "richer", which means "have a bunch of features that must be learned even though they don't apply to MY problem."
[reply]

      Pascal isn't dead, Delphi is the GUI version of Pascal, and it's used quite a bit, especially with HTML editors and such. It's as easy as VB, but not as shitty.

      meh.
[reply]
Re: End of Native Code?
by vkon (Curate) on Jun 15, 2006 at 16:43 UTC
    So, what do you folks say?

    I say this discussion is so-o-o-o-o ancient, and so-o-o-o many times debated, and flame-wared, it does not even worth mentioning it even on slashdot, because this do not bring anything new.

[reply]
Re: End of Native Code?
by CountZero (Bishop) on Jun 15, 2006 at 16:22 UTC
    A typically "slashdot" issue not worth breaking your head over.

    The success of interpreted languages dates from well before the time personal computing power increased.

    I still remember the time when compiling a program could take several hours, so we jumped with joy when you could just type and "run" your BASIC program on any Z80 based computer without waiting, whereas your typical FORTRAN program had to spend the night in the mainframe and produced a stack of paper in the morning with nothing but error messages.

    IMHO the succes of interpreted languages is their short reaction times and not the strength of the processor they run on: I programmed computer games in BASIC on a TRS-80 and they ran smoother than "Oblivion" on a 2 GHz monster.

    CountZero

    "If you have four groups working on a compiler, you'll get a 4-pass compiler." - Conway's Law

[reply]
Re: End of Native Code?
by swampyankee (Parson) on Jun 15, 2006 at 17:21 UTC

    As the computers become faster and faster, the problems become disproportionately larger. The numerical relativity and n-body communities (the latter are people who model the evolution of things like galaxies) routinely have their very highly obtimized Fortran (usually) and C++ (less often) codes run for several months on very high performance or even special purpose (i.e., GRAPE) hardware.

    It's just that the really big problems don't get much attention in the popular press or even major facets of the CS world.

    emc

    e(π√−1) = −1
[reply]
Re: End of Native Code?
by Joost (Canon) on Jun 15, 2006 at 21:11 UTC
[reply]
Re: End of Native Code?
by jdporter (Paladin) on Jun 15, 2006 at 16:06 UTC
[reply]
Re: End of Native Code?
by jdtoronto (Prior) on Jun 15, 2006 at 20:30 UTC
    It doesn't matter how fast or how capable computers become, some of us still need more capability. The GRAPES project is a good example, but you should also see what is done in atmospheric modelling and semiconductor physics. Lots and lots of highly optimised code as close to the CPU as we can get. No native code might be off the radar for consumer applications, but for lots of scientific and technical stuff, it is still high on the agenda.

    jdtoronto

[reply]

Back to Seekers of Perl Wisdom