Category Archives: Projects

Phil 8.19.2010

8:00 – 11:00 VISIBILITY

  • Added a role for Paul O’s group.
  • Fixed Paul O’s tables. It turns out that the Excel ingestor chokes on ‘-‘. Asked Dong to take a look at informing the user if the ingestor fails.
  • Finished my IA refresher
  • Talked to Anne about Paul’s project and about the upcoming meeting.

11:00 – 1:00 Savage Mill

  • Went over with Brian to explain wall positions and discuss ceiling options. We should have a quote by the end of the day.

1:00 – 4:00 FP

  • Working on provisional patent. Decided not to use the software and just write the thing.
  • Got about 5 pages into the application. It’s coming along well

Dong Shin 08.18.2010

  • Started on Project Lock management
  • packed up PPM for review
    • found commas causing lot of errors everywhere! fixed most of them
    • added total budget amount on each Financial Data title
    • FMP not saved when all the fields are empty

Phil 8.18.2010

7:30 – 11:00, 1:00 – 3:30 VISIBILITY

  • Talked to Ann, it looks like things may be happening in September. Need to work on getting some of John’s data ingested
  • Started IA online course. Should finish tomorrow.
  • QC-ing Dong’s PPM code
  • More talks with Anne. Meeting at 9:00 Sept 3 for now.

11:00 – 1:00 FP

  • Working on refining collision detection and adding more obstacles
  • Contacted Stuart Shapiro ( sbs@usiplaw.com) about filing a provisional patent. Looks like this would be useful: https://www.patentwizard.com/
  • Had to get rid of an infinitely large vector that results from a collision too close to the origins of the two colliding point masses basically a 0/x problem, where x > 0

Phil 8.17.2010

8:00 – 10:00VISIBILITY

  • It’s dark when my alarm goes off. I don’t like this. At all.
  • Nothing new at the site. Need to start coordinating with Jim M. to log into the server and see what we need to set up.
  • Pinged Mike about discussing using reflection/introspection for code completion in Python. We chatted for a while about
  • Helped Keil with some mxml

10:00 -12:00

12:00 – 3:30

  • Tech fair

Phil 8.16.2010

7:30 – 10:30 VISIBILITY

  • Meeting with Trish
    • Values need to be compared to goals, not planned in Financial Reporting
    • Highlight if planning is below the goal, but don’t bring up the 5 questions
    • Add legends to the charts
    • If user saves an empty form, it should behave like cancel, and not save anything
    • Add commas to financials
    • All chart tooltips say “outlay goals”

Mike 8.13.2010

  • Played around with PyDev for a very long time and I was right, there is a standard API in Eclipse for providing language support.  Unfortunately, the PyDev plugin spreads it over 20 java projects.  I’ve found methods for generating completion proposals for local variables, local functions, imported packages (through python introspection!) but not a one stop shop for everything.  The environment variables also make it difficult to pull out snippets of code and run them.  I haven’t given up yet though.
  • I took a short break from reading through poorly commented code to connect the AutoCompleteTextArea to a back end java service that will call the appropriate language assistant.  So far so good with a dummy language assistant that provides crayola colors.

Phil 8.13.2010

7:30 – 12:30 VISIBILITY

  • No Trish for meeting reschedule for Monday. I believe that we will try again Monday.
  • Long meeting with Anne and John D. about using VISIBILITY for their stuff.
  • Did find a few bugs with PPM
    • I attempted to save off a filled out version of the “5 questions” form with some cut and paste data in the fields. The form never saved successfully, and the timer icon never came back
    • The vertical tab that switches between charts and data in the financial entry chart has the full year in it. It should be “Year 1”, “Year 2”, etc.
  • Talked to John W. about documentation some.

FP 12:30 – 3:30

  • Got the EL axis working and integrated
  • Got everything working with the physics and the graphics, although I need to clean up collision detection. THings are ready to demo.
  • Worked on the next version a bit. A full hand interface, without additional motion, but able to sense XYZ on its base.

Phil 11.12.2010

8:00 – 10:00, 3:00 – 4:00 VISIBILITY

  • Clift had forwarded some questions about ingesting data from Paul O. I described how to ingest an excel spreadsheet and pointed to the Ingest Manager tutorials. Paul would also like an Access ingestor. Not sure if it would be as straightforward as setting a jdbc access jar and running the same commands we would otherwise.
  • And FGM’s mail servers are down. Whee.

10:00 – 3:00FP

  • Adding in additional axis. Working out which should be x and which should be Y.
  • When I get the Y-axis working, I’ll need to reset the clamp on the fingers so that they stay centered around the “hand”, Which is going to need some additional graphics to show. We’ll start with that.
  • Graphics are done. Working on getting the collision for the hand to be the average of the finger X-axis collision forces.

Phil 8.11.2010

7:30 – 10:30VISIBILITY

  • Nothing new on site, but two meetings scheduled for Friday

10:30 – 2:00 OVERHEAD

  • 9:30 meeting at Savage Mill
  • Worked on the redesign of the office space to accommodate the sprinkler system

2:00 – 3:30FP

  • Got the fingers working pretty well. There is a lag of about 0.25 sec or so between collision and response, but it’s good enough to see that things will work.
  • Need to add an x-axis in with collision response next. Maybe make a maze of gray balls.

Phil 8.10.2010

8:00 -10:00 VISIBILITY

  • Rescheduled meeting with Anne and others for Friday at 9:30
  • Discussed Sharepoint stuff with Kiel. We’re tentatively thinking about having him write a proof of concept app that has Flex talk to Sarepoint via jdbc

10:00 – 4:00 FP

  • Integrating fingers with graphics. Changed it so I’m passing references to the entire PointMass through the dictionary. Now drawing fingers and the target.
  • Fingers move appropriately in the graphics, and collide with the target. Going to add feedback tomorrow.

Phil 8.9.2010

7:30 – 10:00 VISIBILITY

  • When checking over the query logs this morning, found that Paul O. Had spent most of Friday using VISIBILITY. He made 133 queries in 9 or so hours, spending most of his time on the LOTR and cars examples
  • Went over status of PPM with Dong. We’re going to aim for Wednesday to walk through the new changes.

10:00 – 3:30 FP

  • Working on reading the sensors. They seem to be working well. Working on calibration code – done
  • Making a bi-linear normalizing, since there seems to be a big difference between pushing up on a sensor and pushing down.
  • Update graphics to show new axis

Phil 8.6.2010

7:30 – 11:30 VISIBILITY

  • Meeting with Trish and Chris
    • Tabs in Financial Entry should read “Year 1″, Year 2”, “Year 3” etc., to avoid confusion between appropriation year and budget year
    • Still not starting FY in October
    • Whenever a reported obligation out outlay is less than 99.9% of the (calculated) goal (currently not visible), a dialog with Trish’s questions should come up. There is shared data between obligation and outlay reports for a given month.
    • Dialog should be smart enough to know if an obligation or outlay value is not applicable in a dialog, since one or the other or both may be under the estimate.
  • Responded to Paul O’s questions about how to interact with the report builder.
  • Went over the meeting results with Dong, and gave him Trish’s “5 questions”.
  • Went over the rough layout of the visibility system with Kiel.

11:30 – 3:30 FP

  • Working on fingers. Building the ServoMotorController class. Done.
  • Had to wait for the servoController class to initialize before getting the handle to the motors.
  • Had to initialize and engage all axis in the ServoController. I could do this inside the ServoMotorController.
  • Need to adjust speeds and ramps for the servos

Mike 8.5.10

  • The AutoComplete TextArea is pretty much done.  It could use a few small tweaks but those can wait.
  • After a fruitless search of “python autocomplete library”, “python code completion library”, etc. I’ve decided to go straight to some source code so I’m currently digging in to the PyDev plug-in for eclipse to see how they do code completion.  Perhaps each language implementation in eclipse uses a similar enough of an API that I can make some backend service for analyzing code and providing autocomplete suggestions.  I’ll just have to add one language at a time.

Phil 8.5.2010

7:30 – 12:00 VISIBILITY

  • Sent off CATWALK memo
  • There is some demo activity on VISIBILITY, so we’ll how that develops
  • walked through PPM with Dong
  • Burned a new disk for demo tomorrow

12:00 – 3:30FP

  • Added step-to-angle conversion to StepperController
  • Couldn’t get the servo motors to move. After looking closely, I realized Tom had flipped the plugs on the controller header, so the motors were wired backwards. Fixed. Next is to try talking to the fingers in code.
  • Added “home” command to let the fingers and other motors reset.

Feldman Project Strain Guage Calibration

  • At first I noticed there was a strange thing where the voltage wasn’t constant.  I think this was simply because I was moving the tester because once I found a better positioning for the probes it seemed to stop.
  • Finger 2 needed its strain gauges reversed due to the fact that you want low sensitivity and need to up the offset.  The way it was set up it was balancing to the bottom end and it couldn’t be fixed because the offset was all the way up while gravity was forcing the voltage down.  Anything fighting gravity need to have negative offset up positive offset down unless you want it to be ultra sensitive by turning the gain up.  Even in that case I think this orientation is still better.
  • It is possible when interacting with the elevation to accidentally move the wrist.  I don’t think this was because of loss of holding torque but because the linkage isn’t connected to the motor tightly enough.
  • Overall sensors seem to have very good response range. Fingers are pretty sensitive even with gain all the way down but still in good range of force.  Elevation is the hardest to move but
  • Displacement for each sensor goes as follows
  1. Finger 1: up = negative, down = positive
  2. Finger 2: up = negative, down = positive
  3. Wrist: forward = negative, back = positive
  4. Azimuth: right =negative, left = positive
  5. Elevation: up = negative, down = positive
  • I don’t think we need to shield the wires at the moment but once the motors are moving instead of just holding we might find it necessary.  It is going to be hard to tell though because motion will probably cause fluctuations in general.