Date:	October 23, 2007
Time:	4 pm
Location:	Room off Unix Lab
Attendance:	Keith, Joelle, Pragna, Tim, and
Agenda:	Review activity list from last meeting Discuss MMIX manual and how to represent requirements in OO (UML model) Further discuss project requirements / definition
Description:	Swami spoke to Prof. Bolker and found: End of October project vision statement due; Mid-November build documentation (aspire for single build script); Early December Functional Specification Document (FSD). Pragna had been able to build basic C example (factorial) using Ant to build JNDI wrapper via SWIG, compile dll, compile java, and execute java command. Does not yet use SVN for checkout and does not execute JUnit testing. SWIG requires install to run and is causing problems on UMB Unix machines (wants to install in locations that we do not have access). Will work to whittle down issues and communicate via Google group. Agreed that will supported platforms will be Windows XP and Unix and that SWIG will be assumed to be installed on any development/build system. Agreed on project name: Visual MMIX Will remove risk analysis from website… serves no function. Decided we will not use Wiki utility being created by Semantic Workbench group… will continue using Google Groups as our wiki. And will use TRAC as our development page (internal, development website in addition to our external website) Will rewrite external webpage to not use Microsoft Publisher as this is not rendering well on non-Windows browser. Will use raw HTML to build the page and store in SVN to maintain revision control. Tim & Keith have experienced issues with SVN checkout… does not have access to HEAD or some similar error message. Will bring up in class to determine if other groups are experiencing the issue. Keith had MMIX manual questions which lead to following agreements on functionality: There will be no configuration file for the simulator Nothing special for trips & traps exception handling MMIX compiler does not optimize in anyway so no need to worry about this (i.e. instruction re-ordering) color code registers but place in the same pane In Editor there will only line #, label, Op code, operands, & remarks… will not determine assembly code as this requires compile on the fly No pipelining and therefore no need to support multithreading but still show all registers used for these operations In editor show (u & v) values as user completes op command and user spacing will be used, forced tabbing not in place Swami checked and the simulator supports debugger interface where parameters can be passed to the run command to run in interactive mode with defined breakpoints (will get into further detail later) Discussed writing a stand-alone JFace application or creating an Eclipse plugin and group agreed that we would prefer to create an Eclipse plugin for resume reasons. Keith & Tim will investigate feasibility / setup of creating a plugin but Swami insisted that the final project be able to run independently of Eclipse and the project description/spec should reflect. Swami stresses that we should use Google Groups!!!
Next Time:	Everyone will review the short stories and discuss on Google Groups. Perhaps add more stories. Swami will consider what to do if a user passes pipeline or multithread op-codes Swami will initiate Google group topic on short stories Continue working on setup of build system and project development system (TRAC) as well as build script (tasked to Joelle & Pragna) Keith will create basic Eclipse plugin example (text editor) that can also run standalone so that group can see the comparisons. Will work with Tim to agree on outer framework (portion of UI that will hold tabs) Fill in project vision presentation ideas on Google group discussion (describe project, why its useful, & why we think we can do it… minimize technical) Tim will work on changing website from Microsoft Publisher to general HTML pages Continue reading the MMIX manual