ParaQ:Telecon 11/01/2005 1PM EST
Agenda
| Item | People | Description |
|---|---|---|
| 1 | Rogers | New organization for Tuesday meetings |
| 2 | Berk | Multiview update |
| 3 | Shead/McBride | ParaQ v1.0 update. |
| 4 | Shead | Undo/Redo design ( see Undo Design ) |
| 5 | Rogers | Coding Standards; Templates; Copyright |
| 6 | ETI | Pipeline Inspector |
Attendees
| SNL (NM) | Moreland, Rogers, Shead, Wylie |
| SNL (CA) | Templet, Wong |
| ETI | McBride, Richardson, Stimpson |
| Kitware | Ayachit, Geveci, Squillacote |
Notes
Multiview
Berk presented status report on Multiple View changes in ParaView. He's got a version working ( see The Multiview Page ), and by next meeting, he will check in code for others to build against. While working on this, he implemented PVSM linking, and will check that in as well.
ParaQ v0.1
We summarized the previous meeting's decisions on promoting 'Hello World' to 'ParaQ v0.1' Berk raised objections about relying on an 'organic' upgrade process, so the group agreed to specifically address the state of the current code before promoting it to v0.1. In particular, Tim and ETI are responsible for evaluating the utility, coding standard compliance, and documentation of all code by next week. At that time, we'll switch to v0.1. In addition, it was decided that we will have a 'Sandbox' or 'Application' area in the ParaQ CVS repository so that developers can manage code through CVS - in particular, code that is part of development that would break the main Client code. Developers will only check code into the Client that is compliant with ParaQ's design and coding standards.
Tim presented Testing status. Currently, tests can be 'recorded' using, for example, Python. He and Clinton are working on a solution for capturing tests that exercies widgets - the difficulty is that position of the widgets can change from build to build (changes in layout would break the tests), and the testing solution must take this into account.
Undo/Redo
Tim presented the design for Undo/Redo. This was the longest discussion of the meeting. Face it, this is a big issue, and we need to get it right. Tim argued for this capability to be placed at the ParaView Server level, and Berk advocated putting it at the PVSM level. Tim and Berk will woodshed on this in the coming week, and present a more detailed design proposal next week. Berk raised concerns about Undo/Redo on multiple connected clients being an example of feature bloat. We decided to evaluate timeline for implementing Undo/Redo - it's likely we will design a solution, then push off implementation until after the December demo is complete. Undo/Redo is not part of the current demo script, so if it's technically possible, we may push active development of Undo/Redo until after the demo is completed. The group requested that basic Undo/Redo operations be simple, even if complex requirements such as multiple undo/redo stacks are implemented.
There was some discussion that Undo/Redo across multiple servers connected to a single client is at best difficult to understand. In addition, there is a range of these complex issues that can be designed, but implementation may be delayed per the requirements of the demo schedule. Rogers will follow up with Berk about these concerns, and make certain the schedules reflects them.
Pipeline Inspector
ETI summarized the intention of the PipelineBrowser document. In short, they'd like the group to give feedback on the direction such a view should take. Brian discussed a possible relationship with Univ. of Utah, in the hopes that we could find a way to collaborate with them, and get some buy-in from them on the ParaQ project. Brian presented the Utah GL-based graph tool, which will be evaluated as a long-term solution to complex Pipeline browsing. For the short term, group agreed to pursue a simple solution that promotes the small set of operations necessary for the demo application.
Decisions
- Undo/Redo will reside at the PVSM or below. Work can now advance on the Client-side API to PVSM
- We will 'harden' the current code before transitioning from 'Hello world' to ParaQ v0.1
Action Items
| Assignee | Item | Priority |
|---|---|---|
| ETI | Evaluate Utah browser graph technology as possible Qt component. | Long term |
| ETI | Determine how to receive/display secondary video feed (from Sandia) | Short term |
| All | Offline: Settle on coding standards | Short term |
| Rogers | With ETI, determine simplist Pipeline Browser that satisfies short-term needs | Short term |
| Shead, Geveci | Refine Undo/Redo proposal; post updated design on wiki | Immediate |
| Shead, ETI | Remove all obsolete code from Client directory; Document existing code; Verify compliance with coding standards. | Immediate |
| Rogers | Create updated Microsoft Project schedule, in consultation with ETI, Kitware and SNL | Immediate |
