Hi
I think you can find answers for your questions here:
https://cwiki.apache.org/confluence/display/WAVE/Wave+Summit+Talks
I ll try to answer in short, but take in account I might have forgotten
things and the information can be somewhat incorrect. I hope someone more
knowledgeable will fix the inaccuracies.

Q: Each message or edit is a blip?
A: An edit is a delta. Blip is sort of a "message". Thread is a root blip
along with all its children. Blip contains a conversational document which
contains elements which can be text elements, annotations , gadgets or
doodads (the attachments element is a doodad). Also, take into account that
there's a UI representation of the Wave model and the Wave object model.
Those two are similar but sometimes same terms have different meaning.

Q: The action of adding a gadgets and robots is also a blip?
A: Not really. When a gadget is added to a blip, it's a delta(s). You
change the state of the blip. When a robot is added to a wave - it is
similar to adding a human user to a wave. In the sense that the root
wavelet metadata will be updated to include this user as participant. If
the user is a robot, then the changes on the wavelet will be sent to the
robot via the Robot gateway. If the user is a local (registered on the same
wave server)- then the changes will be sent to the other user's client over
websockets ( or using IO.Sockets). Id the user is remote (i.e federated) -
then the changes will be sent to the federating host. Also, when a human
user joins a wave - then a user data wavelet is created and associated with
the wave, it contains user specific data like the read/unread state of the
blips.

Q: How about the interactions inside those gadgets, e.g. adding a task to a
task list.
A: This is like editing the blip, it generates deltas.

Q: Where does wavelet concept comes in, why is it necessary?
A: A wave consists of wavelets. In WaveInABox there could be only one
conversational  wavelet per wave, while Google Wave actually supported
several conversational wavelets.
The access is granter on per wavelet basis. The wavelet is a container for
the blip threads and metadata. It is the entity that is persisted and
federated over federated wave servers.

Q: What is a document (the spec states a structured wave document?)
A: As already mentioned, document can be represented as an XML document
which has structure elements, text elements, annotations etc...

Q: Is there only one conversation manifest document for the whole wave?
A: As already mentioned, in theory there can be several conversational
wavelets per wave - as in Google Wave, but Wave in a Box implementation
supports only single conversational wavelet per wave.

Q: What's the difference between conversation and threads?
A: Not sure what you asking about. In the Wave UI - there are two options
to reply: the regular reply which creates a new blip thread and the inline
reply which creates a new child blip in the thread hierarchy.


Hope that helped a bit, but make sure to watch the Wave summit videos for
more complete and correct info.


On Sun, Oct 27, 2013 at 8:11 PM, Ricardo Mayerhofer
<ricardo....@gmail.com>wrote:

> Hi everyone,
> I'm doing some college research about Wave and have some questions about
> Wave Procotol. If this isn't the right place to post this kind of question
> please direct me and sorry :)
>
> I've read the specs but I'm having a hard to grasp some conversation model
> concepts and relate them to what in see in UI (Google Wave or Wave in Box).
>
> In wave I can create new waves, reply to messages, edit existing messages,
> add gadgets and robots.
>
> Each message or edit is a blip?

The action of adding a gadgets and robots
> is also a blip? How about the interactions inside those gadgets, e.g.
> adding a task to a task list.
>
> Where does wavelet concept comes in, why is it necessary?
>
> What is a document (the spec states a structured wave document?)
>
> Is there only one conversation manifest document for the whole wave?
>
> What's the difference between conversation and threads?
>
> Any help is appreciated.
>
> Thanks.
> --
> Ricardo Mayerhofer
>

Reply via email to