== October 9, 2008 == '''Attendees''': tef, Tom, Elaine, Greg, Eric, Scooter, Conrad === New Action Items === - Greg will demo OpenGL panel (item 2) - Scooter will check with Gilberto regarding classroom availability and scheduling for Library Chimera course for both Winter and Spring quarters (item 3) - Scooter and Eric will look into attachment handling in gnats and how it translates to trac (item 4) === Action Items from previous meetings === - Al will look into configurations for test machines and Greg's machine - Al will decommission SGI after the next Chimera production release - Al will install stereo-capable workstation to replace SGI before decommissioning SGIs - tef will contact Adobe for more information about embedding 3d graphics in PDF documents === Minutes === 1. Review Action Items a. Everyone agreed that installing a public, stereo-capable workstation can wait until after the upcoming production release (at the earliest). b. Everyone agreed that PDF 3D is now on the back burner. The action item for tef will remain but is very low priority. 2. Discuss OpenGL panel behavior (Greg) a. Greg passed out the third prototype for the OpenGL panel. The critique on the panel interface are: - Replace "This often results in lower performance" to "Changing settings may result in lower performance". - Remove "Save and restart Chimera to complete your changes." - Remove "Moved to saved configuration" (more below on this). - Add "Using factory defaults." - Move "Save", "Factory Defaults", "Load Saved" and "Report Config" to the same line as "Disable All", leaving a single button next to "Help" at the bottom. The single button will be labeled (something like) "Continue with Displayed Settings". - The new behavior is that the OpenGL panel will be displayed only if Chimera is started with the "--debug-opengl" flag. There are three scenarios: a. there is no saved configuration, then the factory defaults are displayed and "Save", "Load Saved" and "Report Config" will be disabled. If the user changes one or more settings, then "Save" and "Report Config" are enabled. If user actually presses save, then "Load Saved" is enabled. b. there is a saved configuration AND it matches the current environment, then the saved configuration is displayed and "Save" and "Load Saved" are disabled. If the user changes one or more settings, then "Save" and "Load Saved" are enabled. c. there is a saved configuration AND it does NOT match the current environment, then the factory defaults are displayed and "Factory Defaults" will be disabled. If the user changes one or more settings, then "Factory Defaults" will be enabled. - In all cases, when user presses "Continue...", Chimera will proceed using the DISPLAYED configuration. The saved configuration is never modified unless the user explicitly presses "Save". b. Tom suggested that the OpenGL panel is being designed and implemented with "what's easiest" rather than "what's best for the user" criteria. Perhaps we should step back and consider whether we should just make a production release (so Windows fixes are in a production release rather than in a daily build) without this feature. We can then spend a little time (not expected to be much effort by Tom) investigating whether it is possible to make the OpenGL completely interactive and not require Chimera restarts. The OpenGL panel can go out in the next production release. [[BR]] The counter-arguments are that (a) it's more work, and (b) it's a debugging and not a tuning tool and therefore can be a bit rough around the edges since most users should not need to use it. These arguments effectively agree with Tom's first point, but carry weight because they also address man-power availability issues. As Scooter pointed out, there are many other "real" features that need to be prioritized along with the OpenGL panel if we were to delay until the next release. Conrad wants the tool out with the upcoming release because it's a useful tool and, if it's not perfect, users who are debugging driver issues would prefer to have something rather than nothing. [[BR]] In the end, it was decided that we should continue with the current course to put the OpenGL panel into the upcoming release. c. Eric suggested that a flag be placed in preferences so that users can set it in one session and have the OpenGL panel show up in subsequent sessions. This obviates the need for the --debug-opengl command line flag, which can present a problem for less savvy users, particularly on the Mac. Conrad suggested that, if a user changes settings but does not save them, a dialog is presented when the session ends to offer a chance to save the current settings. Conrad further suggested that these be refinements to be implemented after the upcoming release (but perhaps Greg can look to see if these are relatively easy to add). 3. Schedule Chimera library course for Winter quarter (Scooter) a. Scooter reports that the Library course series will include both the Chimera and Cytoscape sessions, making it a nine-session series. Elaine and Tom both suggested that the "Publication-quality images and movies with UCSF Chimera" session be split into two sessions since there is so much to cover. That makes it a ten-session series, which matches well with the ten-week quarter. b. tef asked whether we will then continue will a Fall-Winter quarter schedule next academic year. Scooter responded that he preferred a Fall-Spring schedule, which suggested that perhaps we should delay this year's series until the Spring quarter. Scooter will check with Gilberto on scheduling possibilities for both quarters. The preferred time slot for the course seems to be 10am-noon. 4. Discuss gnats->trac transition, in particular issue categories (Scooter) a. Scooter has imported the Chimera gnats reports into trac (http://socrates2.cgl.ucsf.edu/trac/chimera) using the category mapping (hastily) supplied by Conrad. Suggestions for alternative categorization methods are welcome. There are 4000+ bug reports, with 600+ that are not closed. Everyone should check out the trac interface in preparation for Monday's group meeting. b. Scooter discovered that all the gnats attachment directories are empty. Thus, there are no attachments in trac. Both Elaine and Eric said that there are attachments in bug reports. Eric and Scooter agreed that the attachments may be in the e-mail message rather than explicitly in gnats. Conrad suggests that the attachments can be extracted easily using standard Python modules. Scooter and Eric will look into where the attachments are and how easy it is to put them in the right place in trac. Meeting ended well after 4pm. Following item was deferred until next meeting. 5. Continue evaluation of features to remove or improve (All)