|
m e e t i n g . a c t i v i t i e s
|
|
| |
- Results previous weeks
- Level one finished. Algorithm does take two CPP's as input.
- Algorithm cannot handle NDD's
- Algorithm does create a draft CPA document and adds concrete elements/attributes to it. Draft CPA is well formed but not valid.
- Algorithm does create more than one CPA if that is possible from the CPPs.
- Algorithm detects 'referencing' problem. Thats when elements are referenced form more than one other element.
The algorithm provides a solution by copying the problematic elements. Algorithm creates a matrix of the problem
and then does matrix magic (copy). BUT the algorithm does not backport it into the rsulting CPA. Future feature.
- Some elements are not treated (list of those elements does exist).
- Algorithm writes out conflict file (maybe the most important part). An entry has XPath expressions of element/attribute in CPP's and CPA!.
- Presentation for Tim
- Current work
- Currently implementing simple BSI (Business Service Interface) which can send receive messages (single process only atm).
This will be the negotiation architecture to send and receive negotiation messages for level 2 and three.
BSI includes a state machine which makes sure the negotiation is in a correct state and follows the negotiation schema.
Simple tui (text user interface) to administrate BSI and test system.
- The message content will be presented to a user first! Then the user can choose a reply. as level 3
- AI for a concrete negotation software for level 2 ?. Dons suggestion: choose one element only! Eg. neg. software needs
knowledge about general transport protocols. So it does need to know if a protocol is valid or not. Then it also needs
knowledge if that protocol is supported by the company, if it can be installed or if the company does not want that
protocol at all. Also the neg. software needs a strategy how do negotiate with the partner (eg present a list of
possible protocols (that would be an advantage for the other company as they could choose the one they prefer
of that list; or try to negotiate over each individual protocol). Many, too many options atm.
- Timetable: May coding, June coding, July thesis writing, August thesis writing.
- Missed to send/write a paper for the Enterprise Distributed Object Computing Conference in September in Brisbane :(
|
|
|