7:30 – 3:30 VISIBILITY

• Still got this cold. But now it’s out of my lungs and in my head. Wheee!
• Meeting with (who?) at S33P to talk about email system(?)
  • Trish didn’t have the time or focus to look at the spreadshjeet in detail, but she said that she would look at it tomorrow and reply with comments in an email.
• Interview with Terry

7:30 – 3:30 VISIBILITY

• Playing phone tag with John P. About admin accounts. He also called about some data that they added to the VISIBILITY data stream that may not be showing up. Since there’s no contract there any more, I couldn’t help much.
• Finished up the power and effect size calculations in the Power Wizard
• Lunch for Sophia, who started today. She might also be helping Dong with report-building queries.

7:30 – 4:30 VISIBILITY

• Got a cold. Ick.
• Working on a way to get exact p out of the qf function in R
• Done! Below is my first function in R

function(calcF, dfNum, dfDenom, limit = 0.001, iter=100){

 alpha = 1;
 oldAlpha = 1;
 step = 0.5;

 for(i in 1:iter){
 newCritF = qf(alpha, dfNum, dfDenom, lower.tail = F);    
 #show(c(alpha, newCritF, calcF-newCritF, step));

 newDiff = calcF - newCritF;
 if(abs(newDiff) < limit){
 return (alpha);
 }
 if(newCritF > calcF){
 alpha = oldAlpha;
 step = step * 0.5;
 }

 oldAlpha = alpha;
 alpha = alpha-step;
 }

 return(alpha);
}

7:30 – 7:00 VISIBILITY

• Meeting with Trish and Christie. Trish wants the following:
  • All rows from the Financial Entry tables, including the obligation and outlay goals, and percentages
  • Contract information
  • Trish asked if the online version would be available in 30 days. I said that it was likely that we would be on the test platform in 30, but probably not on the production server
  • Trish has more requests, but they are waiting until these actions are finished and approved
  • We missed this mailing. The next will be the first of May.
• Went over the new requests with Dong, who will now “not be bored”
• Finishing up white paper – done
• Requested more space on fgmdev

7:30 – 6:00 VISIBILITY

• Showed Dong the DB problems with PPM
• More proposal – done with first draft!

7:30 – 5:00

• Writing up laser pointer whitepaper – some progress
• Working on Test Wizard – done
• 2:00 meeting with Bill D. and others? Nope.
• Finished Two-way Anova

7:30 – 2:00 VISIBILITY

• Went over roles with Dong. Here’s the basic idea (I’m adding this to the bug list as well
  • Admin – Everything from NSA and Site, plus
    • Add User
    • Delete User
    • Modify User
    • Modify database elements
  • NSA- Everything from Site, plus
    • Ability to view all projects and financial data
    • Project Creation
    • Project Deletion
    • Assignment of users to projects
    • Request reports
  • Site
    • View only projects that the user is assigned to
    • Modify project data
    • Modify financial data
    • Cannot change project assignments (may possibly be able to change site assignments)
• Meeting with Bill D. The request for admin has been changed to  a change request. Did the regular backups.
• Got started on the Test Wizard

7:30 – 4:30 VISIBILITY

• fgmdev is running really low on RAM. Looking into how to upgrade that.
• Check up on my various accounts
• Got most of the complex contrast parts in. Finish up on Monday
• My request for an admin account has been approved. That doesn’t mean I have the account yet, just the approval to get the account.
• My charting is not working when I’m reading in three values. Need to debug;
• Write up whitepaper on insurgent laser pointer

7:30 – 5:00 VISIBILITY

• They are calling for snow tomorrow. That is the same tomorrow that is the first day of April tomorrow. The one that is 12 days into spring. That tomorrow. Here’s hoping it’s an April fools prank and that it will actually be 72 and sunny!
• Adding an actionscript computed F test class. Done and working. Going to extend that to handle contrasts
• Tomorrow:
  • Add contrast columns
  • Extend Ftest class to ComplexContrast class
  • Add adjustments for number of contrasts and pre-post hoc.

7:30 – 6:30 VISIBILITY

• Installing updates – windows, java, adobe. Let’s see if we can break *everything*!
  • I broke most things. A fix that is mentioned a lot in the forums and helped me was to UNINSTALL all my old Java SDKs/JDKs/JREs and then install the newest one fresh. After that, all was well
• The computer is running faster now, though…
• Doctor’s appt at 9:15
• Going to try to write a straight F test function using R.
• Checking up on the state of my account on the new server
  • Turns out that you need an ADMIN account to request a root account for a server. Put in a ticket. I wonder what’s going to be next?
• Spent some time looking at the programming OOP in R. Looks doable, but different enough from C++/Java/Actionscript that I’m going to get a book. In the meantime, I’m going to calculate the F statistic in Actionscript and hand that off to the qf(.05,dfnum,dferror,lower.tail=F) function in R to see what I can do with it that way. In a way, it’s kind of nice, in that the processing load is kept on the client.

7:30 – 6:30 VISIBILITY

• I think the new Firefox is slowing down my computer…
• Added a lot of post hoc tests into the ANOVA widget. In the process, I discovered something really cool here: http://psych.wisc.edu/moore/Rpdf/610-R3_post-hoc_one-way_betw.pdf. It turns out that you can get the straight critical F values directly out of R using this:
  • qf(.05,dfnum,dfdenom,lower.tail=F)
• Yes, I know that doesn’t seem all that awesome, but it does mean that I can put some functions together to calculate these things from the basic math, and get some of the flexibility that I’ve been looking for. Whee!
• Applied for an account on our newly active server.
• Copied all the webapps folders and sql databases to move over to the new server.

7:30 – 6:30 VISIBILITY

• It’s freaking cold outside.
• Filled out server paperwork
• Found some R code that will run Dunnett. It runs on fgmdev too!
• Loaded up a new version to fgmdev. Minor fixes. Dunnett tomorrow?
• Discussion with Mike about versions of BlazeDS
• Asked Dong to put script output directly into VISIBILITY rather than having a table viewer.

7:30 – 11:30 VISIBILITY

• Swung by the customer site
  • Talked to John W. All development funding is done. There is a little maintenance money. THere are developer dollars to do something with Sharepoint. I reiterated that integrating VISIBILITY with Sharepoint is pretty straightforward, and far easier than re-implementing in Silverlight or HTML5.
  • Wrote up a list of the software and versions that we will need on the new PPM server and sent to Brian
  • Read in on my privac.
• Leaving early today for air conditioning spring maintenance.
• Worked on diagnosing the Runtime.exec() problem with Dong
• Brought Brian up to speed on how things have been going.

8:00 – 5:30 VISIBILITY

• Got my new leather mouse to match my leather mousepad. Need a leather computer and keyboard 🙂
• Doing more upgrades on the R install on fgmdev. Here’s hoping nothing breaks. Pretty much everything is out of date.
• Well, I got 25 errors, but didn’t break anything.
• Talked to Dong about the problems that he’s having with Runtime.exec() on the Mac
• Having Tom install the VISIBILITY suite on one of FGM VA’s servers
• Initial work on adding in MANOVA and MANCOVA
• Walked through table issues with Mike.
• Webinar on cloud computing at 2:00: http://event.on24.com/r.htm?e=292186&s=1&k=69E5FCD268EA18FC4E0A060B96973886

8:00 – 6:00 VISIBILITY

• Write status report for the last 7 hours of CATWALK
• Write reqs for the applet version of the scripting system. Started here: http://www.fgmdev.com/blogs/?page_id=5220
• Making sure that the coefficient order matches the group order
• Commented my WebInterfaceIDE Server and Client code and walked Dong through it
• Phone call with Brian and Bill on server migration
• Finished fixing bugs, trying to port to fgmdev. Getting Java null exception errors back. Going to make a console panel for sending raw queries and getting responses. Done. It turns out that I was using an old version of Rserve (0.4-7) that came with the Debian apt-get system. The version that I was developing against is version 0.6-3. I installed that using install.packages(“Rserve”). Fortunately that compiled and loaded properly (unlike HH). The new version is here: fgmdev:/usr/local/lib/R/site-library/Rserve. That’s the one that needs to be started.
• added rserve_start to the .bashrc file for root.