Linux Kernel Developer job

[Ted Gould]

--=-WFOr1znJLjA0lzHluhVg
Content-Type: text/plain
Content-Transfer-Encoding: quoted-printable

On Thu, 2004-05-13 at 18:20, Ed Skinner wrote:
>      Turning now to the posting and link above, the Land Warrior project =
is,=20
> in my opinion, a very, very bad one for Linux. Lives will be at stake and=
=20
> will depend, in part, on the reliability of the software. Linux is good, =
and=20
> it's far, far better than Windows, but neither of those operating systems=
 are=20
> appropriate in "life relies on software" settings. Please note that my is=
sues=20
> of reliability go far beyond the immediate implementation of Linux. When=20
> contemplating high-reliability systems, particularly those in the defense=
=20
> sector, I have a huge number of problems with how Linux is maintained,=20
> developed, and tested.

Well, I would have to say that on many levels I agree with this.  In a
production system I feel more comfortable using a Redhat kernel than a
kernel.org one.  Why?  Because I know that version has been tested, and
that fixes will be applied to it in a consistent way.

Likewise, I think that this is something that commercial embedded Linux
providers supply to their customers.  Yeah, as an embedded developer I
can go grab the latest kernel, and it will probably work.  But I'd
rather use a version that has a few more hours on it - like one tested
by my embedded Linux provider.

So, I guess you need to make a distinction between kernel.org, and
places like Montavista.  I would entirely agree with you about
kernel.org, but not so much about places like Montavista.

Now, I don't know of a embedded Linux supplier that is currently
providing a Linux kernel that I would consider appropriate for life and
death products.  I don't think that is a problem with Linux itself, more
that someone hasn't created a business to do it.  I think in the
decision made by the original job poster was: "It is going to be easier
and better for me to do the work to get Linux up to the level I need
than deal with a proprietary kernel" - and in many cases I agree with
them.

I currently work with VxWorks, and we have a source code license, I'm
scared that they are currently the dominant OS in the military today.=20
I'm convinced that they've tested quality in, not designed it in.  There
is no reason that Linux can't get to equal quality, if not much better.

Anyone who just blindly takes kernels of kernel.org and puts them into a
mission critical system is risking the lives of those who they serve.=20
But, I don't think that means that Linux as a whole can not be viable in
those situations.

		--Ted


--=-WFOr1znJLjA0lzHluhVg
Content-Type: application/pgp-signature; name=signature.asc
Content-Description: This is a digitally signed message part

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQBApGUbLE335pRPGp0RAmUuAKDgos0o18BXt050Krmr3vaHAOcuzQCgn6/L
VEM8tXGL48eSKuTZhIzSvf8=
=dxwE
-----END PGP SIGNATURE-----

--=-WFOr1znJLjA0lzHluhVg--