On portable C programming (was: libraw1394 struct layouts...)

[Ben Scott]

On Fri, Jan 9, 2009 at 11:35 AM,  <VirginSnow at vfemail.net> wrote:
> I have to admit I don't understand how this is supposed to work.  You
> haven't posted any URLs or project names for me to Google. :)

  Or even source code snippets.  Posting code to this list is allowed,
if that's what Jim Kuzdrall  wants to do.  I'm somewhat curious about
the techniques myself, but my curiosity is solely academic.  I'm not
math guy, so I probably wouldn't understand them, and I do not foresee
myself having a need for them.

> What would it take to apply these techniques to a real, existing
> project... for example libraw1394?

  libraw1394's problem is interfacing to the kernel, so I'm guessing
they won't help for that.  It sounds like the particular solution Jim
is describing is mainly intended for storing data to permanent storage
(disk, tape, etc.) in a format that's portable between environments
(compilers, architectures, operating systems, etc.).

  I'm sure it could be used to interface to the kernel, at least in
theory -- at the theoretical level, storage is storage, be it punched
cards or DRAM.  But as I mentioned, Linus has been against going to
extra efforts to provide uniform binary interfaces in kernel-land.  It
also may be that the algorithms are not suitable for kernel driver
interfaces, for whatever reason.

-- Ben