7:30 – 9:30VISIBILITY

• Meeting with S33P folks Thursday the 23rd at 11:00
• Brought in my other old windows box to Linuxify

9:30 – 12:30 FP

• Finishing paper
• Ordered an audio oscillator and small amplifier kit. I had the thought last night that sound plus vibration could either augment the kinetic interaction, or possibly replace it. If that’s true, then it’s possible that we might be able to get by with simple force sensitive resistors, which would be pretty cool…

12:30 – 5:30 VISIBILITY

• Working on commons-net telnet sandbox application
• Hmmm. MavenAssist isn’t creating eclipse-compatible jars. Looking into that.
• Success!!!  To get the listing below, the program started telnet, logged in to 127.0.0.1, and sent the command ‘ls -la’ to the Cygwin intetd:

• drwxr-xr-x+ 1 phil.feldman ????????       4096 2010-12-15 13:52 .
  drwxrwxrwt+ 1 phil.feldman Administrators    0 2010-12-15 13:26 ..
  -rw-------  1 phil.feldman Administrators  328 2010-12-16 17:02 .bash_history
  -rwxr-xr-x  1 phil.feldman ????????       1150

8:00 – 4:00 VISIBILITY

• Looks like the PPM new features meeting will be here Tuesday or Wednesday of next week at 11:00
• PPM DB still growing and being used. After importing the data for Masha, no activity. Hmmm.
• Let’s see if I can get the Cygwin telnet functioning…
  • Yes, after flailing until 2:00. In the end, I deleted the current install and pulled the version from the gatech ftp mirror. I think that the version at vatech may be corrupted.
• Starting with the java user/telnet manager. I think it will need to do the following:
  • Create/Delete users. Note that this will have to be OS-specific.  In addition, in the case of cygwin on Windows, the user account will have to be added within cygwin as well by running mkpasswd as admin before the user can be logged in through telnet
  • Open a telnet channel on 127.0.0.1. Commands and responses go through this channel. For example “javac foo.java” is the command and error messages from the compiler are the responses.
  • Copy files from the client to the appropriate directory on the server (and with the correct permissions). This may require a pwd command to be run to determine where the user is. Or we do this all graphically (preferred), and keep track on the client
• Spent some time pondering how apps might deploy using dynamically created virtual servers. Looks surprisingly straightforward?
  • http://www.howtoforge.com/creating-virtual-machines-for-xen-kvm-vmware-workstation-6-vmware-server-with-vmbuilder-on-ubuntu-8.10

7:30 – 3:30 VISIBILITY

• For the record, I’d like to state that it is too damn cold.
• Fixed Masha’s data
• Had a conversation with Christine about PPM. It looks like they are going to need a new section to track “real-time” inflows and outflows of cash from projects. Head’s up, Dong! Working on scheduling a meeting to go over things in detail.
• Adding the Cygwin Telnet/Ftp server to my machine. So far, here are the steps (from http://www.wiley.com/legacy/compbooks/nilsson/InetdunderWindows2000.doc):
  1. Run “mkpasswd -l > /etc/passwd” at the Cygwin prompt.  This creates the passwd file which is similar in format and function to the Linux passwd file.  Run mkpasswd with no parameters for a usage summary.  You might need to add a -d parameter to grant domain users access.
  2. Edit /etc/passwd and delete everything but the accounts you want to have access permissions.  (In most cases, you’ll end up with only two lines in this file — your account and Administrator.)
  3. Make sure that your home directory (ie, c:cygwinhomephil.feldman) exists.
  4. Right-click “My Computer” and select Properties.  Click the “Environment Variables” button on the “Advanced” tab.  Add a new system variable…the name is CYGWIN and the value is “ntsec” (no quotes).  This tells Cygwin to use your Windows 2000 security.
  5. Add C:CygwinBin to the PATH system variable.  (Stick it at the end, separated from everything else with a semicolon (“;”)).
  6. Run “/usr/sbin/xinetd.exe –install-as-service”.
  7. Reboot!
  8. Well, that didn’t work. Started up the Windows Telnet server for now.  That’s just a matter of turning on the service.
• Looking at the apache.commons-net package as my telnet interface
  • The main project page: http://commons.apache.org/net/
  • An article about commons-net with an example of how to create and use a TelnetClient: http://www.informit.com/guides/content.aspx?g=java&seqNum=40
  • JTelnet is old and probably not supported, but might be useful for reference: http://sourceforge.net/projects/jtelnet/
  • JavaSSH is less old and has more hits. Could also be useful: http://javassh.org
    • Their javadoc: http://javassh.org/download/source/index.html
  • Tom says that Telnet can be made secure by moving it to another port. Also that the xinetd application from cygwin is supposed to be secure.

8:00 – 4:00 VISIBILITY

• Spent some time with Brian on site this morning working out the details of what our server needs
• Looking at ways to deal with interacting with the shell as a way of dealing with the server. Developers write code as text files that are compiled using whatever tools we integrate. Files are saved as files with some kind of namespacing. The key to all this is to have some kind of workable interface to the OS that is essentially a terminal window. From there we load commands, parse results, use command-line debuggers, etc.
• Found a java telnet server that handles all the issues of sending commands to the OS and getting responses (and errors) back. It’s built around the same Runtime.getRuntime.exec() command that we are using in MavenAssist, just a lot more complete. It’s here: https://sourceforge.net/projects/pwts/. In addition to building a remote object around it, we should refit the version in MavinAssist.
• Talked to Anne about Masha’s data woes. Promised to take a look at it tomorrow

7:30 – 4:30 VISIBILITY

• Deployed Dong’s latest version. It still needs the following:
  • Appropriations Editor:
    • Needs Save/Don’t Save/Cancel for panel and appropriations closing
    • Columns should be wider. We have the room, and some of the text is falling off the edge. Also, add tooltips for cases wherethe text is too wide
  • User Management and Creation
    • Save/Don’tS ave/Cancel when panel is closed
  • Contract Management
    • Save/Don’t Save/Cancel when panel is closed
• Backed up PPM
• Talked to Mike about handling tooltips
• Talked to Dong about the web-based IDE. He decided that he’d rather do the DataManager, so I’m writing up the IDE concept.

7:30 – 1:00 VISIBILITY

• No new data today on the backups
• Talked to John W. He’s still working on getting a meeting set up for this week
• Did my security awareness training
• Discovered a bug in the AirFileImageApp that had to do with bad sequencing of events. Fixed.
• Working on the Dynamic object. Got it so that the object can watch any declared variable. Working on rolling in the dynamic part.
• Call with Anne to discuss CMM meeting and getting some additional data for the data visualizer

8:00 – 4:00 VISIBILITY

• Found the following bugs in PPM
  • In the Project Management Panel, the Portfolio Manager and the Portfolio Admin column contents are switched
  • There are still scroll bar problems with the following screens
    • Project Editor (even with 3FY budgets)
    • Financial activity summary and data entry grids. The charts are OK
• Working on DataManager
  • Seeing how to get ObjectProxy to do what I want it to do
• Lunch at G&M

FP

• Talked to Agnes at Permabond about strain gauge glues. Her number is 201-343-2268

8:00 – 4:30 VISIBILITY

• Added Mike and Christine as users to fgmdev. Passwords on request.
• Burning a copy of PPM to bring in prior to CMM meeting
• CMM Meeting with Anne at 1:30. – not sure how that went. Maybe well?
• Turned out that MavenAssist was pointing at the wrong activator. Worked with Dong for a while to get that fixed.
• Started on DataManager. Built a library and test project.

FP

• Got glue. Tom is testing the mounting of the SGs. He’s working on the wiring harness while the glue sets.

8:00 – 4:30 VISIBILITY

• Backups are working smoothly, and the size of the backup is growing more slowly. Most of the data must be in the DB by now
• Testing Dong’s PPM code
  • Narrower clamp on comboboxes in Project Management
  • Tooltips in project management comboboxes
• Looooooooong discussion with Mike about data architecture for Vis4
• Meeting with Anne? Yep – showed all of the VISIBILITY suite. Not sure what happens next

FP –

• Brought in wire for Tom

7:30 – 3:30 VISIBILITY

• Deployed cpPlusDate and mvPlusDate, which are working just fine
• Made backups
• Talked to John W. He’s going to try to set a demo of PPM with Kathy and the new TD for early next week.
• Chris(tine) has finished Flex in a Week. She sent me a not asking what to look at next. I said to look through AirFilImageApp to get aquainted with a larger project, and start to get familiar with Maven and the flex-mojos. Once she’s done with that, she can start updating MavenAssist to build Air apps correctly.
• Digging into FlexXB
  • General description: http://code.google.com/p/flexxb/wiki/GeneralDescription
  • No luck. FlexXB doesn’t recurse down the tree. Moving on…
• Worked with Dong on getting a new version of PPM to play with the browser window a bit better
• Since I really don’t want to build a generic serializer (again!). I’m thinking about the components that all our applications seem to need. The thought is that maybe we build some higher-level components that have self-serializing functionality that can be put together in simpler ways than a straight mxml experience. My list of visual components so far is:
  • Login
  • DB Table/column browser
  • Charts
    • The regular, data, bar, column, pie
    • Unusual – maps, status, network
  • Data grids
    • for looking at tables of numbers or…
    • Excel-style spreadsheets
  • Data entry forms
  • Formatted output
  • Special widgets, such as speedometer-style displays
• To make this work we need a communication framework to the server that either simply (or automatically) sets up tables to talk to these forms.
• Lastly, there needs to be an authoring/administrative level that creates ids for applications (to aid with data partitioning), and that supports our scripting.

7:00 – 4:30 – VISIBILITY

• Deployed my new copy, which works fine, though it has one diagnostic print statement in it. Need to fix that
• Fixed the bugs, and added move capability. The executable looks at the name it was called when executed and adjusts its behavior. If the file is called cpPlusDate, it will copy. If it’s called mvPlusDate, it will move. Uploaded the files to my ftp site (since we still need CDs!), to deploy on Monday. Also committed to subversion.
• Looking at ways to have a user build a form online and then use it.
  • Read the xml on the server side and have it compiled (http://livedocs.adobe.com/flex/3/html/help.html?content=apache_3.html). This doesn’t allow for graphical creation of the component though.
  • Serialize as xml:
    • http://stackoverflow.com/questions/270046/flex-xml-serialization-and-de-serialization-of-nested-object-structures
    • http://cookbooks.adobe.com/post_Serializing_object_to_XML-11988.html
  • Serialize the top level object using the ByteArray class (http://www.kylehayes.info/2008/01/06/Persisting-an-ActionScript-Data-Model/). This would need the following:
    • A container class that knows how to read and write a serialized object to the server
    • That the app exist entirely inside the above container.
    • The question is – will this store the entire hierarchy under that object in a way that will work later?
  • I like this better because it has the ability to graphically lay out the form. There would be two fundamental types of containers – an Authoring container and a Player container. The authoring container would have the code to lay out the form (and set up the tables on the server?). The player would simply load and show. Aaaaaaaaaaand it doesn’t work. Turns out not all the data for a display object gets stored this way. Seems to only work for value objects.
  • Trying out FlexXB. Changed the compiler options to preserve meta data
  • Set up my path so that the mxmlc compiler can be reached from the command line