User:Daleharvey

OverView

I have decided to set up this page as my log for the google summer of code project, formalising the protocol for inkboard communication and converting the library to pedro.

Predicted TimeLine

• Deliverable 1 - Friday, 30th June

  Extend and finalise the initial draft of the protocol specification started by David Yip (/src/jabber_whiteboard/protocol/protocol.tex), intergrating as much of the standard protocol underdevelopment at http://wiki.jabber.org/index.php/Psi_Whiteboarding which must include the following

  • Specification for conflict resolution
  • Use Case for peer to peer and Multi User Chat session establishment
  • Use Case for communication of edits between sessions
  • Listing of valid message types
  • State transition diagram of Inkboard Clients

• Deliverable 2 - Friday, 14th July

  Develop and finalise the XML schema that is implemented by the inkboard clients, this will define the valid message types that may be transferred

• Deliverable 3 - Friday 28th July

  A basic inkboard client developed using the perdo XMPP client, must be stable enough to be integrated with the trunk, that contains at least the following

  • Initialisation and Destruction of multi user and peer to peer session
  • Basic communication of changes between documents

• Final Deliverable - Monday 21th August

  End of Summer of Code project, a fully stable inkboard intergrated into the inkscape trunk, ready for inclusion in the release of .45, features to include:

  • Conflict Resolution
  • Local Undo mechanism

Project Log

5/06/05

graduation next week is interfering with the progression of the project, but i have managed to develop a system for reducing the undo history stack, and have decided on session establishment protocol,. I just really need to sort out the service discovery which needs input from other, finalised the protocol document up until that point

2/06/05

Still cant decide on a protocol for MUC session establishment, I think this will need to be voted, it has quite large effects on implementation. there has been talk about limiting the undo queue needed which is rather a large implementation detail. I am late with the first deliverable, but I think it was expected with how late I was starting the project, It has also been slowed down trying to get multiple parties to agree, It will certainly be ready in time along with the XML schema

1/06/05

Discussed some of the last issues with the JEP, MUC session establishment, I think we need to have another conference. Issue include service discovery, the sandbox and actual transfer of in band images

30/06/05

I actually wont need to patch ejabberd, somone has sent the patch, hopefully it will be introduced into latest stable, darcs is being incredibly annoying to install

29/06/05

Joonas and I have agreed on most protocol issues and decided to work on the same JEP. Some of the tradeoffs, I have agreed that <new> elements should be sent as raw svg, which means I have to compile change events to new svg elements. The other major one being that we will not use the id attribute to track elements, instead a whiteboard specific id given during the creation of an object

28/06/05

the first draft of the protocol used a revised version of the old inkboard session negotiation protocol, I think CAP / disco are more suitable for the initial stages of session establishment, so I am researching this to hopefully commit to a revised protocol soon
I also followed up the issue of presence stanzas being incorrectly ordered on ejabberd servers, it look like I would have to submit my own patch to resolve this

27/06/05

Dicussed with joonas about the integration of the jabber.org and my protocol, so far there are only 2 issues, address of elements, which I need to discuss with david yip, and the sending of native svg through the protocol. which is hopefully easy to resolve

26/06/05

Started my draft of the whiteboard protocol, for now I will keep an updated copy at http://debrowse.com/protocol.html
Jabber.org hosted a discussion on whiteboarding issues, it went quite well, as both peer to peer and muc with client side synchronisation are looking to be supported, with an option for muc with a server component

25/06/05

Have had a few problems integrating a jabber_whiteboard board specific gui to the pedro client, im going to put any development work on hold until I can get the protocol drafted

24/06/05

Managed to fix the windows build, it ignored makefiles_insert and compiled old code, there are issues with compilation against older versions of gtk.

23/06/05

I spent today merging the INKBOARD_PEDRO branch with the trunk, all works well on linux, I have disabled inkboard to be built on windows for the time being as there is some rather confusing compile errors, which I will hopefully fix tomorrow

22/06/05

After discussion with David Yip and Joonas of the jabber.org SoC project, A proposal has been set for conflict resolution, I have started extending the protocol document that can be found in the pedro branch. We are setting up a time convenient for a whiteboard related conference in the next week or so

21/06/05

I have finished uni and can start working full time on the protocol / lib conversion. We have decided to integrate the whiteboard gui into the pedro gui. I am starting my work with adding the whiteboard gui into pedro, and defining groupchat and peer to peer initialisation protocol