On Wed, May 11, 2016 at 02:55:53PM +0200, Stephane Bortzmeyer wrote: > A very interesting paper (I said "intesresting", I didn't say I > agree!) on open networks where independant nodes with independently > developed programs interoperate thanks to standards. The author claims > closed and centralized systemes are better, because they allow faster > evolution (he uses security as an example). Propietary is better if you want a walled garden, but some day you find yourself needing to let third parties use your service/protocol and then... you have the same problem, only now you're the authority for the documentation and so you have more responsibility. Walled gardens end up having to let third parties in because people end up needing integration with other systems. Imagine that Glapchat has enormous success and now businesses want to integrate it (including additional security functionality, no doubt) into their enterprise networks. Glapchat can say "nö", but maybe Glapchat can monetize better if they say "yes", but they might not be interested in making the integration investment themselves. As you can see, you can't really avoid standardization. You can choose SDO venues that have the attributes that you want (e.g., pay to play), or even create your own SDO, or act as an unofficial SDO. But you can't really avoid standardization unless you start as a walled garden and stay a walled garden with no third parties in your midst. My advice to anyone creating a protocol is to first get running code, then think about what SDO style you want, and then what SDO, and when to start the standardization effort. Some have brought their protocols to OASIS, others to IEEE, others to IETF, others have formed their own consortia, and others follow a very informal process until it no longer works -- these are all perfectly valid choices. But first, get running code, otherwise you'll get bogged down and you'll then probably blame anyone but yourself. Nico --