Re: Notes from the Real World

From: Greg KH (gregat_private)
Date: Tue May 01 2001 - 16:42:05 PDT

  • Next message: jmjonesat_private: "Re: Notes from the Real World"

    On Tue, May 01, 2001 at 07:13:15PM -0400, jmjonesat_private wrote:
    > Works well for coders, not software designers.  In many shops (my own 
    > included) the coders are writing to SPEC... the SPEC being human language.
    > The translation from code to human to code often results in new and better
    > results.
    
    Ah, but who wrote the implementation spec in the first place?  My odds
    are on the the people who are going to be doing the coding.  What I am
    trying to say here, is that if the coders themselves do not create the
    specs, the two trees rapidly diverge, with a lot of effort to keep them
    together.  I've worked in ISO-9000 shops, so I understand well the spec
    and paper trails needed.
    
    And if the people doing the coding are not the same people who
    wrote the implementation specs, I feel sorry for them, and you should
    tell them that there are much better jobs for programmers out there.
    
    > I do not intend to imply THIS information is not necessary.  It is.  But 
    > I think an outside evaluation of this project in HUMAN-READABLE terms may 
    > help develop a more "rounded" implementation.  I intend to undertake it,
    > and don't see it as wasted time, unless you and yours' will ignore the 
    > results?
    
    Ok, more power to you.  I will not ignore the results at all.  I just
    wanted to make sure that you understand the difficulty of it.  And since
    you do, I welcome reading the document.
    
    > Are you afraid of critical review?
    
    Not at all, I welcome it.  It's the only way people grow.  I wouldn't be
    writing open source code if I didn't like criticism :)
    
    > After all, isn't the POINT of a loadable module interface to move stuff
    > OUTSIDE the kernel in a coherent way?
    
    No, a kernel module interface is there to provide a clean interface to a
    subsystem.  It allows pluggable systems.  Think of loadable device
    drivers. It does _not_ move anything outside of the kernel.  Remember
    module code still runs within kernel space, with all the privilidges and
    pitfalls that are there.
    
    > If I (and others) can write a document that helps make this interface more
    > understandable... why do you object?
    
    I'm not objecting at all.  I just was trying to propose what I see as a
    simpler method to create the documentation, that's all.
    
    Looking forward to the documentation,
    
    greg k-h
    
    _______________________________________________
    linux-security-module mailing list
    linux-security-moduleat_private
    http://mail.wirex.com/mailman/listinfo/linux-security-module
    



    This archive was generated by hypermail 2b30 : Tue May 01 2001 - 16:42:55 PDT