source code documentation

[Marty Leisner]

This whole attitude bothers me.

The whole idea of  GPL is to put a good implementation out in the
public.  If someone steals ideas, the GPL version should still be
better and more commonly used.

I used Samba enormously to generate the Netbios part of Docucentre
printers (about 5 years, it was really difficult to reverse engineer
what SMB was doing -- its somewhat easier with CIFS -- but the whole
SMB architecture is so bizarre (I explain to people that printing from
copy file.ps \\host\share
and
net use lpt1: \\host\share
copy file.ps lpt1:

and click a window
are all different, not to mention the 
I would have even used samba (we would have put it 
on ftp.xerox.com and I got permission to free up the hooks (which are
useful to no one else) in it.  I found a lot of samba to be very hard
to understand (when I figured things out, I put some comments
which helped me back into samba -- I distributed them).

I also have done A LOT OF WORK on tcpdump (I need to get it intergrated to
a newer version).  I gave it back to the tcpdump guys...I didn't see it...
(I may redo it...let me know if you want to see examples of decoding
smb's I found useful in my work).

Also, to a large extent, I've spent a reasonable percentage of my work time
working on free software (to accomplish a task).  I turn out and give this
changes I find back to the community.

I just wonder about this attitude -- rms feels copyleft is a sufficient
protection -- but it seems the labor to reverse engineer SMB is incredibly
valuable (wouldn't it be less if Microsoft would definitively say how it 
worked?)

Maybe with the Linux stocks now market capped at multi-billion dollars,
maybe some money could be found to fund the samba team?

I just think there's one way to do software -- the right way.

For all practical purposes, you could put the code through an obfucator --
or a comment stripper...

marty