search
attachments
weblink
advanced
Overview
Content Tools
Work Completed
- Medici
- While waiting for comments/feedback on suggested UI changes I tweaked my wiki review to clarify parts and add more comments
- Traced UI code to better understand how Medici works and to prepare for tweaking the UI
- Started sketching new/improved UI's. At Luigi's suggestion, I started posting them to another wiki page. I will continue that this week. I will also start working on improvements that won't require feedback (e.g. adding remove tag comand/handler to desktop client) and creating wizards for dialogs that should be handled by wizards (e.g. import dataset).
- Attended staff meeting via skype
Work Planned
- Medici
- Review and add any additional comments/clarifications to UI review
- Study Medici desktop UI code to better understand how Medici works
- Begin sketching new UI's and layouts to get a baseline for making additional tweaks once others have commented on UI review
- Attend staff meeting at 10am on Monday November 29th.
Comments
This week went mostly as planned. I thought I might receive feedback on my UI review, but time has prevented this from happening. I was able to start the UI sketches to work from once I get feedback and will enter Jira entries and start working on obvious improvements (see above).
Work Completed
- KISTI
- Pulled together a list of high-level tasks for year two of the KISTI project so Jong and I can begin scoping the work for next year and present a budget to KISTI requirement
- Medici
- Attended Monday 10am Staff Meeting via Skype
- Began preliminary work on desktop redesign by sending email to Rob, Luigi, Joel to provide feedback on the wiki document that discusses my review of the desktop client. After they review the document and make suggested changes to the document, I want to pull together a list of high level improvements/features so we can prioritize for 1.2 and I can begin sketching new UI's for further feedback. I also suggested a possible 1/2 - full day meeting to brain storm ideas to get the design started.
- Found and reported a minor bug with the about screen for the desktop client (version not updated)
Work Planned
- KISTI
- List high-level tasks for KISTI Year 2 and provide estimates on time requirements for each task
- Medici
- Attend Monday 10am Staff Meeting
- Next major task will be to lead redesign of Desktop User Interface
Comments
This week went as planned. I took half a personal day on Wednesday afternoon and Thursday/Friday were holidays.
Work Completed
- Medici
- Attended Monday staff meeting via skype
- Resolved MMDB-1024, Arrows to preview elements of collection should not launch dataset selection event
- Continued evaluating Medici desktop and determining potential improvements, organized my notes and put them on the wiki at: https://opensource.ncsa.illinois.edu/confluence/display/MMDB/Desktop+Features. These notes will need to be evaluated and we will need to open features in Jira for the enhancements we determine should be made to Medici Desktop.
- KISTI
- Discussed briefly the issues Jong is having with running PTPFlow on KISTI's machines and followed his and Albert Rossi's email exchange.
- Jong did not ask me to evaluate any time requirements so this might be bumped to next week since he is still debugging the KISTI PTPFlow installation
Work Planned
- Work on Medici Project
- Attend Medici Staff Meeting on November 15 at 10am
- Finished assigned Medici bugs for the version 1.1 release
- Finish evaluating Medici Desktop client and update wiki with comments
- KISTI
- Provide feedback on time requirements for future KISTI work for Jong
Comments
This week went as planned. Next week I might spend some time on evaluating KISTI feature time requirements since that didn't happen this past week.
Work Completed
- Worked on Medici
- Resolved MMDB-978, gallery view null pointer exception caused by incorrectly displayed popup (nothing was selected so a popup menu should not have displayed). This issue is fixed.
- Resolved MMDB-1010, file upload through the desktop client shows file size of -1. This issue was linked with MMDB-972 and resolved by switching to the REST endpoint for file upload.
- Resolved MMDB-1011, add tag in Metadata page does not check for empty selection. This was fixed.
- Resolved MMDB-1016, preview of a collection does not display when selecting a collection in the collection view. This was fixed by updating the plug-ins in the build since only the build displayed this problem.
- Resolved MMDB-1020, check that selecting a dataset in the gallery view does not download the original file. This was verified.
- Resolved MMDB-1027, clear cache does not actually clear the preview cache. This is fixed.
- Worked on MMDB-972, change file upload to use REST endpoint. There was one slight difference between the web applet and the desktop client that I could not resolve, but I believe should be completed today (Monday November 15th) after Joel provided some further updates.
- Attended Medici staff meeting/bug session via Skype on Tuesday November 9, 2010
Work Planned
- Work on Medici
- Attend Medici Staff Meeting/Bug session using Skype
- Work on desktop bugs for Medici 1.1 release
- Time permitting, finish software recommendations and post on wiki
Comments
This week went as planned. I took a personal day on Friday, November 12. This coming week I will finish working on bugs until Medici 1.1 is released, then I will finish my software recommendations and post them on the wiki.
Work Completed
- Worked on MAEviz
- Resolved MAE-1221, MAEviz latest is not building properly and SYNER-G developer cannot launch MAEviz from product file, this is now fixed
- Tagged new build 3.1.2
- Created working set that will simplify the setup process for new developers
- Updated the MAEviz developer wiki with instructions on how to setup MAEviz using the project set
- Nightly build is now building successfully
- Worked on Medici
- Met with Medici development team using Skype to work through a debugging session and determine where in the short term I can help them the most
- Checked out Medici to Eclipse workspace and setup my development environment
- Ran Medici and based on my experiences I filed bug reports for bugs found. I also started documenting improvements to the software which I will continue doing this week and then transfer them to a wiki for review to determine my next role in the project.
- I filed and fixed MMDB-975, AnnotationView had the wrong ID in the Perspective so it wasn't loading that view correctly.
Work Planned
- Work on MAEviz build, determine why SYNER-G developer is receiving errors launching MAEviz from Eclipse
- Work on Medici project, determine what role I can play and how I can help the project most.
Comments
This week went as planned. Next week I will continue working with Medici to find bugs and potential software improvements and new features.
Work Completed
- Completed ethics training
- Worked on KISTI Project and released final version 1.0
- Resolved KNSG-95, update wiki tutorial for 1.0 release.
- Closed issue [KNSG-97|http://jira.ncsa.uiuc.edu/browse/KNSG-97, create team project set for eAIRS, this was finished for last weeks tutorials.
- Resolved KNSG-98, moved a common method to superclass Analysis wizard.
- Resolved KNSG-100, allow users to create eAIRS input files within application
- Resolved KNSG-101, eAIRS mesh parameter does not vary, this was a bug in the UI
- Resolved KNSG-102, added a preference to allow user to specify remote location of eAIRS executable and auto-populate UI field for this
- Resolved KNSG-103, add sections on creating inputs and using the new preference to the eAIRS tutorial. This issue was opened to address the new functionality we were going to add to the final release after our meeting with KISTI. This work probably could have been added to the work on KNSG-95, but I didn't see the issue before this was opened.
- Resolved KNSG-104, convert all error stacktraces to use log4j.
- Resolved KNSG-105, move the top level Security menu to be before the Help menu.
- Finalized wiki documentation for eAIRS tutorial and architecture documentation.
- Reviewed final report to KISTI that Jong prepared from our wiki documents on the project.
- Attended Friday's all staff meeting via webex
Work Planned
- Work on KISTI Project for final release on October 29, 2010
- Finish ethics training
Comments
This week went as planned, the final release was published and final report submitted to KISTI.
Work Completed
- Arrived in Champaign on Monday at 2pm, met with KISTI, presented KNSG tutorial, eAIRS, and hosted seminars on Tuesday and Wednesday
- Worked on KISTI Project. There was a new feature request that came out of the meeting and a software bug
- KNSG-100, allow users to create input files from within eAIRS application. This is nearly finished, the interface just needs to be tweaked.
- KNSG-101, for eAIRS, the mesh does not vary in the parameterized runs, only the input file. This is nearly fixed and just needs to be debugged.
Work Planned
- Travel to Champaign Monday to meet with KISIT representatives (Dr. Nam and Dr. Moon) over the next 2 days (Tuesday/Wednesday)
- Work on KISTI Project, continue wrapup
Comments
This week went as planned. Next week will be the final week on the first phase of the KISTI project, I will be finalizing the code and documentation.
Work Completed
- Worked on and completed architecture documentation for KNSG Framework, this involved KNSG-87, documentation major classes and extension points. Some class documentation still remains, but the majority of the classes contain information on their use.
- Finished Introduction to RCP Tutorial for Tuesday/Wednesday meeting with KISTI
- Finished howto tutorial guide for creating new applications for the Tuesday/Wednesday meeting with KISTI
- Jong notified me that KISTI would be visiting this week and not next week so we had to complete the tutorials and architecture documents quickly in time for their visit.
Work Planned
- Work on documentation KNSG Framework classes
- Finish architecture document
- Start KNSG tutorial for creating new applications
Comments
I took a personal day on Monday October 11th. This week I had not planned on putting as much time into the tutorials as I did since we thought the KISTI meeting was the week of October 25th. I had only planned on starting the tutorials and putting more time into class documentation and architecture. Since the meeting is the week of October 17th, I had to quickly finish the tutorials.
Work Completed
- KISTI
- Resolved KNSG-96, an issue where the repository was allowing duplicate tags. This is fixed.
- Worked on KNSG-98, an issue involving moving some repeated code for the analysis wizards down to the base class. This is nearly finished (just needs documentation)
- Worked on KNSG-95, updating the user tutorial with the latest changes. I believe this is complete, it just needs to be gone through briefly to make sure all user interfaces match version 0.8's user interfaces.
- Continued documenting the major classes in the framework and updating the architecture tutorial. This work will continue next week.
- Started a tutorial on how a developer can create a new application with our framework. This work will continue next week.
- Met with Jong using google chat to discuss the latest KISTI project status.
Work Planned
- Work on KISTI project Documentation
Comments
This week went as planned. Next week I will continue working on documentation and the tutorial that will be part of the KISTI visit. I took half of a personal day on Wednesday and Thursday.
Work Completed
- KNSG
- Resolved KNSG-91, this issue involved automating the retrieval of results after the workflow finishes. Users can now simply right click on a finished workflow and select to retrieve results. The results are downloaded and added to their scenario. This issue took longer than anticipated due to instability in the scratch space on ABE, links to results would not always resolve properly (a known system issue). We made the decision to allow users to specify where the results should be placed (e.g. user-home) and made this part of the workflow.
- Resolved KNSG-93, this issue involved simplifying the User Interface for launching workflows. The user was asked to specify the gridftp URI to use for transferring the data. This was redundant since the user was already asked to specify the resource to run on and the resource's XML file contained the gridftp URI. This is now fixed.
- Resolved KNSG-94, this issues involved allowing users to delete multiple files from the repository or from their scenario. Initially only single file delete/remove was implemented.
- Built and tested version 0.8 of eAIRS.
- Released version 0.8 of eAIRS on Friday October 1, 2010 and updated the wiki with release details.
- MAEviz
- Sent email to a SYNER-G user with instructions for accessing the MAEviz Subversion Repository. I also updated the MAEviz wiki with the correct instructions (the displayed link was correct, but the actual underlying hyperlink in the wiki was incorrect).
- Filled out and returned to HR my Conflict of Interest form.
Work Planned
- Work on KISTI and release version 0.8 of eAIRS on Friday October 1, 2010.
Comments
This week went as planned. Next week I will continue to work on documentation and tying up loose ends for the KISTI Project as the project wraps up.
Work Completed
- Worked on KNSG-91, this issue involves automating the retrieval of workflow outputs. Al finished up a new attribute in PTPFlow last week that should make this easier. This issue should be resolved this week.
- Worked on KNSG-87, this issue involve documenting all major classes and interfaces for building new applications with our framework. I also added wiki documentation on the architecture of KNSG Version 1.0. This is a work in progress and will continue this week as we begin to wrap up this version of the project.
- Tested eAIRS/PTPFlow with the new feature Al added for recording the working directory as part of the workflow events.
- Met with Jong using google chat to discuss wrapping up the KISTI Project.
- I also answered a few questions about how an analysis in MAEviz works sent by Sheng-Lin, a graduate student of Amr's.
Work Planned
- Work on KISTI project and begin wrapping the project by adding developer documentation and documenting main classes and extension points.
Comments
This week went as planned.
Work Completed
- KISTI Project
- Resolved KNSG-81, the issue involved mocking up a new version of PTPFlow's Workflow Overview view. This view was updated based on the mockup after approval from Jong and Al.
- Resolved KNSG-82. This issue involved allowing the user to specify the global resource to run the workflow on. This will require more work in the future because the RMI Service should specify what global resources are available.
- Resolved KNSG-83, this issue involved providing eAIRS its own implementation of the scenario view, which should be extended from the base class the framework provides, allowing users to control how their data are displayed.
- Resolved KNSG-84, this issue involved updating the sample data that comes with eAIRS RCP. We did an interim release midweek to address this.
- Resolved KNSG-85, this issue involved allowing users to drag and drop multiple datasets from their data repository onto their scenario.
- Resolved KNSG-88, this issue involved updating the text and screenshots of the eAIRS tutorial for the latest release.
- Resolved KNSG-89, this issue involved removing an unnecessary field from the simplified user interface for specifying the RMI Service. The simplified version is intended for users who are connecting to a pre-configured RMI Service where the user only needs to specify a minimum amount of information to connect since everything else is configured for them.
- Resolved KNSG-90, this is a bug where users could not select the same input twice for the parameterized workflow. Users might be using the same mesh with different input parameters so they should be able to select the same mesh. This is fixed.
- Resolved KNSG-92, this issue involved making PTPFlow's help documentation available to eAIRS RCP through the help menu. Users can now access PTPFlow help and I have added placeholders to add eAIRS Help and Tutorial.
- Released 0.6.1 on September 14, 2010, which contained updated sample datasets. The last releases sample data would not run with the eAIRS VTK executable.
- Released version 0.7 on Friday September 17, 2010.
- Requested a feature from Al to have PTPFlow store the job runtime location in the StateInfo made available to the client so we can automate the retrieval of result data.
- PTPFlow Project
- Resolved PTPFLOW-25, I added icons to PTPFlow views included in eAIRS that had no specified icon. This provides a more consistent and professional look to eAIRS and PTPFlow.
Work Planned
- Work on and release version 0.7 of KISTI Project
Comments
This week went as planned. Next week I intend to work on adding the eAIRS Tutorial to the help menu and working on automating the retrieval of result data from the global resource the workflow runs on.
Work Completed
- KISTI Project
- Met with Jong using Google Chat to discuss usability improvements to PTPFlow views that are critical to eAIRS.
- Worked on KNSG-81, designed a new look for the PTPFlow Workflow Overview view. The actual work on the view was tracked as part of the PTPFlow Jira project page.
- Resolved KNSG-58, users now can launch Paraview visualization toolkit with their CFD results
- Resolved KNSG-70, dataset tags are now used to categorize the data in the users scenario view.
- Resolved KNSG-74, added a table of contents to the eAIRS tutorial for easier navigation.
- Resolved KNSG-80, fixed a null pointer exception that was occurring when trying to retrieve result datasets.
- Worked on and nearly completed KNSG-83, I pulled out the common pieces into a super class for the scenario view that subclasses can extend to add their own functionality. As part of the KNSG toolkit, users will build custom applications and extensions need to be identified and documented, this was part of that process. Other parts of KNSG need to be updated with documentation so the release contains all the necessary documentation for building and extending the framework.
- PTPFlow
- Resolved PTPFLOW-24, Updated Workflow Overview view with a new look and feel to improve usability since it is a main view for eAIRS.
Work Planned
- Work on KISTI project and releas version 0.6 on Tuesday, September 7th. We are delaying the release to get another feature into the software that would allow users to launch Paraview with their result files.
Comments
This week went as planned. We will continue to look at views in PTPFlow that are critical to eAIRS-KISTI and update the look and feel to improve usability.
Work Completed
- Worked on KISTI Project
- Resolved KNSG-51, setup dedicated RMI Service. Jong worked with Al Rossi to setup our RMI Service and I helped test it while they worked through various bugs in PTPFlow.
- Resolved KNSG-72, this lets users specify the location of the eAIRS executable on the launch machine and eliminates a hard-coded executable location.
- Resolved KNSG-78, I created a simpler interface for specifying the RMI Service when connecting to a remote RMI, less information needs to be specified.
- Resolved KNSG-79, added a refresh button to the scenarios view allowing users to refresh the contents of the view.
- Resolved KNSG-69, added the options to delete and edit gridftp entries. These entries will be used in the next version to specify which service to use to move local data to the remote location before executing the workflow. Currently, a user must specify the gridftp URI in its entirety.
- Worked on KNSG-58, launch paraview from eAIRS rcp with the result file from the eAIRS workflow. We received a new executable from KISTI that will produce result files that can be viewed in Paraview, an open source scientific visualization application. I am in the middle of testing the new executable now and plan to test and release the software on September 7, 2010.
- Reported PTPFlow-17, convert ServiceInfo object actions to commands framework to provide finer control where menu options are displayed. With the actions framework, they follow the object; however, some of the actions need to be performed in a particular view part or a null pointer is thrown. If these were commands, we could specify which popup menus they are displayed in so they follow the object only where it makes sense for them to be performed. Converting to commands looked simple to do upfront, but was more involved and will require more advanced command framework logic so I will look into this later as time permits.
- Resolved PTPFlow-18, this fixed a bug where ServiceInfo actions that were not executed in the expected view would throw null pointer exceptions. I added a check for null and did an internal search for the part it expected.
Work Planned
- Work on and release version 0.6 of eAIRS (KISTI Project) on Friday September 3, 2010
Comments
This week went mostly as planned. We moved the release date of version 0.6 back to Tuesday September 7th in order to implement an extra feature.
Work Completed
- Worked on KISTI Project
- KNSG-75, Resolved, added MacOS to the builds for eAIRS
- KNSG-77, Resolved, investigated an issue with exporting the eAIRS product with the defined plugins for the features. There were some plugins that were no longer part of the project set that were still in the list of required plugins for the features. This prevented developers who followed the tutorial for setting up their workspace from exporting the product for their machine.
- KNSG-72, in progress, worked on allowing users to specify the location of the eAIRS executable for their workflows, which will remove the restriction of the hard-coded location of eAIRS.
- Worked on Marketmaker Project
- Resolved MM-194, which fixed a bug where census tracts were being selected/unselected, but the table profile with the census tract values was not being updated properly.
- Opened issue MM-235, a bug where not including the history token in the URL causes a null pointer exception
- Worked with Jong to transition from using a python proxy to using a Java servlet for handling XmlHTTPRequests (due to a Javascript security restriction, the domain between requesting machine and providing machine must be the same). Now it is much easier to test code inside of Eclipse since a proxy script is no longer required. This will improve our efficiency when testing new code.
- Finished MM-237, which involved wrapping the Javascript code for a loading indicator in gwt-openlayers. This object can now be used in our code; however, the way they change the visibility of the object does not seem to be compatible with GWT since they alter the DIV element so I am not sure this will be useful for us to indicate that a map is loading.
Work Planned
- Work on KISTI and Marketmaker Project, Jong has an issue that he wants me to look for Marketmaker.
Comments
This week went as planned. It took a little time to get my marketmaker workspace sync'd up with the latest changes since it has been over a month and a half since I checked any code in. Next week I intend to continue working on KISTI towards the next release and any Marketmaker issues that Jong needs me to look into.