On Fri, Aug 6, 2010 at 1:11 PM, Ian Roughley <[email protected]> wrote:
>
> I wonder whether a slightly different protocol that would allow for the
> federation of existing
> non-real-time content as well as real-time content would be received better
> by the community. One
> that avoided the need for significant code changes, and one that would
> allow services to federate
> any type of content.
>
> Ian Roughley
> Pulse Platform Architect.
>
>
I agree. I think it would be pretty easy to add an abstraction layer over
the OT stack allowing for OT on different data types. That would allow us to
make the conversation manifest a first-class data structure (like
participant lists) with its own set of OT algorithms. I think on net, it
might end up reducing complexity in the code. It would also let us remove
DocumentBased[Map/List/Boolean/Etc].

For each data type, you would need to specify a network serialization and a
set of functions for transform. The transform functions would need to obey
certain constraints (TP1, invertability).

It would also be tempting to use this as an excuse to replace the OT
algorithm with something that supports TP2. If we shift to an algorithm like
TTF[1], it would let us have more uses for federation, though it might also
risk fragmenting the community. I think we need some system to decide these
things.

-Joseph

[1]
http://hal.archives-ouvertes.fr/docs/00/10/90/39/PDF/OsterCollaborateCom06.pdf

-- 
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