6.170 Progress Report: November 21 2005 Team Voltron: Ed Platt, Clayton Sims, Dave Glasser, Dennis V. Perepelitsa (TA: Vincent Yeung) New Design Issues ----------------- Our ``themes'' implementation, which allows use of arbitrary images for gizmos and balls, is nearing completion. However, we still have outstanding questions involving the appropriate way of constraining the size and shapes of the images in the themes. We especially want to work on a way of allowing our gizmos to have arbitrary polygonal boundaries to represent complex Gizmo shapes. We have not yet fully fleshed out our UI design for Edit Mode, but we have some good first steps. New Bugs -------- Absorber firing key shoots the ball even when it's not in it. The ball seems to be moving much faster than we'd expect from the animation on the project page. Rarely, balls escape from the OuterWalls. Resolved Design Issues ---------------------- We completed our initial design and implementation of Run Mode, including all of the standard gizmos and the ball. Multiball is partially implemented. Collisions is not handled for ball/ball interaction or ball/flipper interaction yet. The game is currently ``playable'', except that the only input the user can provide is firing the absorber, since flippers can't hit balls. Connections are partially implemented; key connections (from the XML) work, but gizmo/gizmo connections are not yet implemented. Rudimentary filesystem manipulation is in place (saving / loading levels). Edit Mode is essentially not implemented yet. Some issues that we went back and forth on included how much functionality to put in the Gizmo abstract base class, how much to put in the ``dumb 1x1 bumper'' class, and how much to put in the more specific triangle/square/circle classes. We also made many small refactorings, like eliminating a ``willCollide'' method that essentially duplicated all of the code of the actual collision handling method itself. Fixed Bugs --------- Over the course of this week, we of course encountered many bugs from time to time, but most of them were fixed within an hour or so (other than the currently open bugs mentioned above). Since we had no open bugs last week, we cannot report that we fixed any outstand bugs. Some of the temporary bugs that we fixed included Flippers being drawn in in the wrong absolute location; KeyPresses not getting recognized by JPanel; the XML to Level parser couldn't process unfiltered whitespaces and text; KeyPress repeating indefinitely (which we solved with the very helpful MagicKeyListener); and OuterWalls not stopping ball properly. Issues From Last Week --------------------- As last week's meeting was less formal, there were no outstanding issues. Success At Meeting Plan ----------------------- Unfortunately, when we made our plans last week, we were not fully aware of how specific the goals for the preliminary release were; we had planned to spend more time on Edit Mode and less time on implementing all of the gizmos. Thus we surpassed our goals in some areas and did not achieve them in other areas. However, we think we did a good job of meeting the requirements for the preliminary release. Next Week's Plan ---------------- Dennis will work on the internals of the Edit framework, and head the battery of unit tests. Clayton will work on the UI for the edit framework, as well as on the theming system. Dave will work on implementing new Gizmos (especially collision detection) and getting the framework for new gizmotypes implemented. Ed will work on implementing new Gizmos and getting the physics framework finetuned. We all plan to work during the Thanksgiving vacation.