Re: Protocol documentation updated

On Mon, Jan 20, 2003 at 10:11:09PM +1100, Pete Ryland wrote:
> Quoting Marcus Brinkmann <[email protected]>:
> > While I am writing the
> > documentation, I am also working on a reference implementation (freemd),
> 
> Yay!  A fourth implementation, just in case the other three don't work?

Well, that is just a matter of fact.  I only know libnetmd from Marc, which
doesn't even have a license, and Pete's omd, which is quite good work, but
doesn't have the abstractions I would like to see (and which itself was
started because libnetmd was stalled to death).  Furthermore, the core of
libnetmd and omd is small, and barely the beginning of a full netmd
library anyway.  I think there are a lot of subtleties, like the empty
receive before sending commands, etc, that we have simply ignored so far,
at the potential cost of portability to (future) netmd hardware.

I also need a program to test my ideas, scan for commands, and try out
things like the get disc title command.  I am way too comfortable with the
GNU coding standard and the GNU tools in general, so it is much better for
me to work in the same environment than to try to adopt to other
environments, esp if we are only talking about a couple of dozens line of
code anyway.  The real work in this project is not in the source code (which
is canonical, and does not offer any challenges - at least as long as encoding
and encryption are left out), but in the protocol reverse engineering.

> And "reference implementation" simply does not make sense.  We are not
> *defining* a protocol, merely trying to work out an existing one.  If anything,
> the "reference implementation" in this case is Sony's proprietary one.

You are totally correct.  But another way to look at it is that we *do*
define a protocol, which we like to believe is compatible, maybe even identical
to Sonys netmd protocol.  As Sony doesn't release its specification nor a
free reference implementation, the best we can do is guessing.  And for the
protocol as I specify it in the manual, I am writing a reference manual that
implements exactly what is specified in the manual.  In other words, you
should be able to read the protocol description in the manual and the
implementation in the freemd source code, and it all should make sense and
be in perfect agreement.  I don't really care if I specify the same protocol
as Sony does internally, as long as I am compatible with all Net MD devices
(present and future).

Thanks,
Marcus


-- 
`Rhubarb is no Egyptian god.' GNU      http://www.gnu.org    [email protected]
Marcus Brinkmann              The Hurd http://www.gnu.org/software/hurd/
[email protected]
http://www.marcus-brinkmann.de/