At this point, do we need the additional overhead of a protocol website?  I'm 
thinking here of
projects such as lucene, hibernate and others that created protocols/standards 
as part of the
reference implementation.  This would avoid unnecessary overhead and process 
while we get off the
ground.

I like the fact that everything is together.  The one change that might be 
applicable, is to split a
protocol stack project from the server project in the proposal.  This way 
implementations that are
java-based could re-use the RI easier, and we still have a protocol that 
non-java implementations
can reference (although with protobufs we might have a high level of entry 
here).

/Ian

On 11/15/2010 12:15 AM, Michael MacFadden wrote:
> All,
> 
> During the Wave Summit we discussed several topics.  The possible move
> of the WAIB project to apache incubator has been covered elsewhere:
> 
> http://groups.google.com/group/wave-protocol/browse_thread/thread/5363c635e8e950fc
> 
> However, in relation to the move to apache, the question of the Wave
> Protocol vs the implementation that is Wave In A Box was raised.
> Typically apache projects are code oriented.  Apache does not
> typically act as a standards bodies.  This is  bit of a gray area
> since new projects  applications / utilities often include XML / SQL
> schemas, APIs, and messaging formats that other entities would need to
> adopt to work with those components.  These could be considered pseudo
> standards; however, they typically dicate how to use or integrate with
> the tool in question.  They aren't really defining a public standard
> that everyone should adopt.
> 
> When we were initially discussing the the move to Apache, we assumed
> that both the protocol and the implementation would both transition to
> apache.  All content would then be migrate from waveprotocol.org to
> the new apache project.  However it was brought up that another option
> would be to leave all items relating to protocol development on
> waveprotocol.org; only the Wave in a Box project (code, docs, wiki,
> etc) would move to apache.
> 
> The gist being that WIAB becomes the reference implementation for the
> protocol, but that the two may indeed be managed differently.
> waveprotocol.org could become the public body responsible for working
> on the protocol.  Open discussion and changes to the protocol would
> happen there.  There would be a process for planning and approving
> revisions to the protocol.  Once a revision is published, the WIAB
> project would then update its codebase to maintain compliance.
> 
> A few benefits to this approach are:
> 
> - It fosters discipline on managing the protocol.  Developers can't
> simply make de facto changes to the protocol just by modifying the
> WIAB code.
> - It is possible that those that theorize about the protocol and those
> that develop the WIAB codebase might not be the same group of people
> (right now their is obviously a large if not total overlap).
> - It sets us up to move the protocol to a formal standards body later
> on if we see fit.
> - It gives the impression that we are drinking our own cool aid,
> rather than running around with scissors.
> 
> 
> I would like to open discussion on if this is what we want to do, or
> if we want to manage this all in one spot.  There are pros and cons to
> each approach.
> 
> Sincere Regards,
> 
> Michael
> 

-- 
You received this message because you are subscribed to the Google Groups "Wave 
Protocol" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/wave-protocol?hl=en.

Reply via email to