User Tools
This is an old revision of the document!
====== MAGIK Meetings Minutes ====== This document contains meeting minites based on the discussion about the MAGIK development. ===== 17/06/2013 ===== **Attendes**: Paramita, Ognjen **Topic**: Start of the MAGIK 3 implementation Important steps: * Develop and implement controller modul * (First attempt) We will create a external module (a web service) that will have a vector of connections. Once a new MAGIK instance request a db connection we use this vector as FIFO structure and provide a new connection by killin the oldest one. * The problem 1 is that an instance and the controller have to communicate via object reference (but how?) * The problem 2 is that even if we have the above described mechanism how we can be sure that the object create with ''connection = DriverManager.getConnection(url, props);'' we can kill? By ''connection.close()'' * **(Second attempt)** Use Hibernate for saving and loading on the level of schemas (hopefully it will do all necessary job of closing unused connections for us) - <fc #FF0000>We try this one</fc> * Implement //proper Hibernate// controller that saves on the level of schemas * Develop and implement algorithm that computes specializations of max size k for an incomplete query * Develop and implement algorithm that computes generalization for an incomplete query **Integrate VLDB and CIKM version** * Issue 1: Reasoning with one query at the time (CIKM version) or several queries at once (as in VLDB version) * Issue 2: TC Suggestion at the main page or move it to the analytic page with query approximizations * Consult Werner ---- **TC-statements (postpone for later)** * Constants in TC-statements should be represented with using single-quotes, e.g., pupil('John', '1', 'b') i.e. instead of convention that constant term starts with a lower-case letter it should be constant term are between single quotes * Note that a is different from "a" in DLV!!! (change query encoding as well to use double quotes) **Look and feel (low priority)** * Representing TC-statements in a datalog form instead of the current format (ask Werner?). For example, pupil_a(Name,Level,Code) ← pupil_i(Name,Level,Code) , learns_i(Name,'French'). * Implement JS tables that allow re-arraignment of table raws - up and down. (later) * Return coloring and style to the previous version (MAGIK CIKM)
