Re: design goals vs mechanisms (was: Re: Let's do some coding :-)

[Jonathan S. Shapiro]

On Thu, 2005-10-27 at 00:11 +0200, Marcus Brinkmann wrote:
> At Wed, 26 Oct 2005 17:14:39 -0400,
> "Jonathan S. Shapiro" <address@hidden> wrote:
> > > Support for legacy applications...
> > 
> > Yes, though I think there is not yet consensus about whether a 90%
> > solution is needed or a 100% solution is needed.
> 
> I am not so sure that this is really the point of contention.  So far
> I have heard two vocal proponents for legacy support, and these are
> Alfred and Olaf.  What have they said?

Then you should add me. There is a significant body of development tools
that I want to be able to run. Most of these are part of the FSF/GNU
tool chain, and these rely on POSIX or a POSIX-like emulation.

Yes, obviously, I want to be able to run these in the POSIX environment,
but I would also like to be able to run them "natively", whatever that
turns out to mean.

> Nothing and nobody can stop Olaf from pre-installing a
> bug-by-bug compatible POSIX sandbox on top of the system and selling
> that as a product.

Marcus: I am laughing very hard at the moment. As far as I know, the
expression "bug-for-bug compatibility" first appeared in the paper "The
KeyKOS Nanokernel Architecture", in 1992. I was the primary author. The
expression was coined by Alan Bomberger, who implemented the KeyNIX
emulation for KeyKOS.

It is particularly funny to see it reinvented in exactly the same
context: POSIX emulation!



I do encourage everyone to read the two papers:

    Experiences with the Development of a Microkernel-Based,
      Multiserver Operating System, and
    The Failure of Personalities to Generalize

You can find links to both in the 3 October section of my current course
web page:

    http://srl.cs.jhu.edu/courses/600.439/2005.fall.syllabus.html

Both papers describe real-world experience. Both are very easy reads.
There were a lot of issues, but excessively ambitious integration was
definitely one of them.

Remember is that Rawson was an IBM employee, and he wanted to keep his
job. The criticisms in that paper are, if anything, understated.


shap