Metis

[Kyle Sexton]

On 3/6/07, Jared <jared at hatwhite.com> wrote:
>
> >> The whole virtual machine concept implies that you are not
> >> to be trusted as a developer. You are expected to trust
> >> the language vendor more than your own programming skills.
>
> > This isn't the concept of the virtual machine. After all,
> > it's really just a extremely CISC CPU intended to be
> > emulated. It's not part of the language.
>
> I understand this. I still resist the notion of any limitation
> between a programming language and the machine, because it is a
> dangerous trend to limit the developer from full, assembler-level
> access to all hardware. Otherwise you incrementally stunt his
> abilities, and that is sad. I am thinking about development fifty+
> years in the future, when these issues are more mature, and from
> that perspective, the Java approach is incrementally more and
> more limiting to the developer.
>
> >> Java has cross-platform merits, but clearly, it was designed by the
> >> marketing department, not by the geek squad.
>
> > Java has cross-platform merits not because of the virtual machine,
> > but because of its extensive standard library giving abstractions
> > for most of what people need to do. C and C++ can be generally
> > cross-platform, provided you only use the standard library, but
> > their stdlibs don't cover as much ground. Qt, on the other hand,
> > builds on C++ and replaces its standard library with one that is
> > more flexible and has a broader range of support, including the
> > QtGui module which provides for graphical interfaces.
>
> Java does have cross-platform merits because of the virtual machine.
> Your observations on libraries are keen, but are not the only
> reason that Java has cross-platform merits. We're both right:
>
> I quote from http://en.wikipedia.org/wiki/Cross-platform.
>
>      As previously noted, the Java platform is an exception
>      to the general rule that an operating system is a software
>      platform. The Java language provides a virtual machine,
>      or a "virtual CPU" which runs all of the code that is
>      written for the language. This enables the same executable
>      binary to run on all systems which support the Java
>      software, through the Java Virtual Machine. Java executables
>      do not run directly on the operating system; that is,
>      neither Windows nor Linux execute Java programs directly.
>      Instead, a special computer program called the JVM, or
>      Java Virtual Machine, is run by the operating system, which
>      in turn runs the Java program.
>
> Again, I'm fine with Java as a language, if only the virtual layer
> could be removed.
>
> >> If you can't crash a server, and crash it big with buggy
> >> inline assembler, how are you ever going to learn grace
> >> under pressure, which is a necessary component of any
> >> responsible world domination plan? ;)
>
> > You can't, or at least shouldn't be able to, crash a server
> > with mere userland code. In most cases, you can't provided
> > the server has been setup properly (eg, limiting forks to
> > prevent forkbombs).
>
> The logic behind "limiting forks" is a good idea for beginners.
> But it is incrementalism in the wrong direction, not toward
> freedom, and it removes useful tinkering ground from mature
> developers. Do not think that all geeks will always have the
> freedoms they take for granted today. Classlines are already
> being drawn. Be wise. Stake a claim in the ruling class because
> you understand the underlying issues.
>
> Not far in the future, kernels will be considered "finished"
> and stable, because hierarchical power structures will have
> consolidated around them, and will prevent innovation, which
> is a pattern easily measured by studying the social dynamics
> of any large religion: Initial Freedom followed by incremental
> imprisonment of its own followers, until they are killing each
> other over absolutely trivial issues.
>
> We must recognize we are building the foundations of a future
> which will not always be as free as it is now -- unless we
> look at these deeper issues and their long term consequences.
>
> I prefer the geek to have full ability to crash the box --
> and the wisdom and ethics not to do so; not a firewall of
> technology preventing him, but a firewall of ethics. It's like
> the logic behind the 2nd Amendment to the Constitution.
>
> You seem to prefer removing power from the geek by limiting
> him, rather than relying upon the reputation economy to inspire
> geeks to do better, which is more perilous, but better in the
> end.
>
> The original rant remains standing: On a philosophical level,
> virtual machine layers limit the geek. Some geeks like that.
> This one prefers freedom.
>
> -Jared
>
> _______________________________________________
> Kclug mailing list
> Kclug at kclug.org
> http://kclug.org/mailman/listinfo/kclug
>

I think I'm the complete opposite.  I love the fact that VMWare/QEmu/etc..
are coming along and creating a virtual machine where I don't have to worry
about hardware compatibility anymore.  I'd like to eventually have
preconfigured/pretuned virtual machines that are specific to a job function
become the norm.  This makes server migrations, turn-ups, turn-downs, etc..
far easier imho.  But this is coming from a SysAdmin, not a programmer, so
that may be the difference.

-- 
Kyle Sexton
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://kclug.org/pipermail/kclug/attachments/20070309/6304d223/attachment.htm