Re: Linux Kernel, Java Applications?

From: Lew Pitcher (lpitcher_at_sympatico.ca)
Date: 12/13/03

  • Next message: John Winters: "Re: Linux Mandrake 9.2 kernel headers - where?" 
    Date: Sat, 13 Dec 2003 17:24:24 -0500

    bigbinc wrote:
    > "JOUIN Romain" <jouinromain@campus.ie> wrote in message news:<3fd112ac$0$17135$626a54ce@news.free.fr>...
    >
    >>Why do you want C developpers to move to JAVA ?
    >>
    >>"bigbinc" <bigbinc@hotmail.com> a écrit dans le message de
    >>news:d1b33313.0312051411.411bf3a6@posting.google.com...
    >>
    >>>What do you think a distro based on java, not just java but jboss and
    >>>its microkernel? It may be slow on slow machines but with focus on
    >>>java, it may force developers to move away from C, it could happen(hmm
    >>>maybe even a sourceforge project).
    >
    >
    > Take me, you out of the picture. What entry-level courses are taught
    > at most major universities(2003)? java.
    >
    > I went to the book store, I saw a couple of books on C/C++, one
    > row(BORDERS), I saw almost half a section devoted to java, several
    > rows, many more books than C, not to mention the xml section.

    Can I ask you how many books on COBOL you saw? How about books on DB2? Or IMS?
    Or "Hardware design"? Or even "Operating Systems"? I bet that you didn't see
    very many of any of those.

    > I just see a trend, and linux is 'kind of' moving towards that trend.
    > Obviously Microsoft is not going with that trend.

    The problem is that there are classes of problems that are well-solved by
    certain technologies, and classes of problems that are /not/ well-solved by
    those same technologies. You tend to find books written about technologies in
    conjunction with solutions, not about technologies and problems.

    While you /might/ write a financial management system (like the banks use for
    their Demand Deposit Accounting or their Mutual Funds Management or even their
    Income Tax Accounting) in Java, you more likely would write it in COBOL. You
    /might/ write an operating system in Java, but you would more likely write it in
    C. You don't find many books on COBOL or C, because COBOL and C are well-known
    quantities. The good books have been written already, and there's no need for a
    plethora of books explaining the technology to those unfamiliar with it. OTOH,
    Java (or C++ or .NET or XML or whatever the flavour of the month is) are /not/
    well-known quantities, and there /are/ a plethora of books about them /because/
    the readers are unfamiliar with the technology.

    In other words, don't judge the worth of a technology by the number of books
    written about it. The number of books only goes to show how unfamiliar the
    technology is, not how capable it is.

    Additionally, don't write off older technologies /because/ they have fewer books
    written about them. Those technologies serve their purpose, usually better than
    the newer technologies, and you shouldn't discount them simply because they
    aren't fashionable to the bookstore crowd.

    Finally, while any Turing-complete language can be used to accomplish a
    computational task, some are better at certain tasks than others. You /can/ dig
    a tunnel with a soup spoon, but it's better and easier to use a pick and shovel.
    You /can/ build an operating system in Java (or COBOL or bash, for that matter),
    but it's better and easier to use C instead. 

    -- 
Lew Pitcher
Master Codewright and JOAT-in-training
Registered Linux User #112576 (http://counter.li.org/)
Slackware - Because I know what I'm doing.
  • Next message: John Winters: "Re: Linux Mandrake 9.2 kernel headers - where?"

    Relevant Pages