9:20 am - 5:20 pm

As you can tell from this week's time logs, I'm really enjoying this concept of flex time!

This morning, I spent some time looking into how the floorplan visualization might be implemented. I have a few ideas, but nothing that I feel is a perfect fit ... yet. I also tried to find out if there are any perl modules that will do 3D graphs, by which I mean not just x-y graphs with depth to them, but graphs of x-y-z data points. While there seems to be a well-established 2D graphing module, I haven't seen a 3D corollary or expansion.

I looked at that auto function, but I just can't get it to work. I don't know what's wrong. My code matches the code on the site in all non-trivial aspects as far as I can tell, but if anything, it's more broken than before. I don't know if the plugins are part of the issue or not (I'm leaning towards "not"). So I'm stuck on that.

I spent some time getting this site up to par, clearing out filler text, looking through relevant sites to put together a good description of ElderCare, and so forth. While the banner graphic could use some work, I think I can say that this site is out of the "construction" phase and into "maintenance."

I looked a little at what Andy was doing, and I tried to help Nick track down a program error. I printed out the code from Nate's "view" folder/Catalyst application, so that I could spread everything out and try to see how it fit together. I installed some more software on my computer and helped Chinonye do the same.

I'm not sure what else I can do Monday, except look for ways to learn Catalyst (the rest of the tutorial involves modules of questionable installation status) and familiarize myself more with Nate's code. Also, there's an interview that was sent out over the listserv that I can go look at. 

9:50 am - 5:50 pm

I listened to the interview on philosophies of aging and care of elders. I decided to go ahead and try to create some mockups of visualizations.

First I did a floorplan mockup, which is where I spent most of my day.  Note to self: The fastest way to make the tigerplace floorplans usable is to manually put black and white where you want it, then in Gimp, delete the white to leave it transparent. There isn't really any way (that I've found yet) to automate the process, due to there being about 3 jillion shades of white and several more between there and black. (Converting it straight to monochrome is NOT pretty. Don't do that.)

Now I'm working on a mockup of a 1-D reading, such as a vibration sensor or something else that has continuous (or many discrete) values. I can't seem to get my code to access the "left" property of a div from javascript, so that's going to make it hard to do the motion I need. I'm looking at online scripts that scroll text to try to find someone who's done it. I found one very good, simple example -- but it works in IE, not Firefox, and I don't know why. I've bookmarked a more promising site to look at tomorrow. 

9:20 am - 5:20 pm

I worked on that second animation. I found my problem. I had been trying to do something like this in Javascript:

    myDiv = document.getElementById("coolJunk");
    myDiv.style.top = myDiv.style.top + 1;  

Apparently this works in HTML 4.01 Transitional, but not in XHTML 1.0 Strict. Here's a place where me throwing in a doctype before doing anything turned out to be a disadvantage. I found a script and stripped out everything nonessential -- see this page and try switching between doctypes. One works, and the other doesn't. I think the DOM must be different between the two. The one is built on HTML, and the other is built on XML. At least, this is the best understanding that I was able to piece together.

Oh, and it seems that if you do the following,

   .coolJunk {
     top: 5px;
   }
   <div id="coolJunk" style="color:red;"></div>

then your javascript call to myDiv.style.top will come up empty, because top wasn't set in the style element (ie, you could call myDiv.style.color, but not .top).

I finally gave up and decided HTML 4.01 Transitional would be just fine. See the fruits of my labor. Note that you can easily edit several parameters all in one place in the source. 

Nate made it in today after being sick for a few days. He showed me how we install Catalyst plugins. I learned that when you open an SSH session, you should log out of it before you close your terminal window. Otherwise, it will leave a process running until the end of time on the server, potentially causing tears and grief for some sysadmin someday.

Nate also showed me a neat Flash graph on Google Finance. I then googled "Google graphs" and found a very cool interface that is probably even more suited to what we want to do: http://tools.google.com/gapminder/ . I've emailed Gapminder to ask about using their software with our data. The FAQ says this is OK, but the license on the first software download says otherwise. And I'm not sure the downloads they have listed are quite what we want. Sidetrack: Gapminder did a presentation that looks like it would be really neat to watch sometime.

Nate and I looked at that auto function that I couldn't get to work. I learned some cool shell commands, but we didn't meet with much success. I did, however, find in the mailing list archives that someone else had had the same problem as I did. The solution given was fairly high-level, so I'm going to have to keep looking to figure out how to do what they said to do (specifically, how to check for a path). It's rather frustrating.

9:15 am - 4:45 pm

Nate showed us how the database tables are structured and taught us some SQL. We went to the weekly meeting, where we had the monthly conference call with the University of Virginia. We asked them about what they can give us to help us develop the Linux app.

Our group of 5 had lunch at Sky Hi, a first for me. Afterwards, Nate showed me several things I can work on while he's out for the next couple of days.

I posted my "auto" question to the Catalyst mailing list. I installed a text editor called Ted, but haven't tried it yet. I need to get some clarification on hours from Dr. Skubic.

Thanks to the Catalyst mailing list, I was finally able to get that auto function working! One of the members also directed me to a more current version of the tutorial. I downloaded the .pod files and learned about turning them into .html (using a command called pod2html -- who'da thunk?).

I found some more Catalyst documentation and explanations, such as a couple of tutorials on concepts such as web framework, MVC (Model-View-Controller), etc. 

I spent the last half or so of the day on Subversion, an open-source version control system designed to be the successor to CVS (Concurrent Versions System). A version control system is a cross between a time machine and a sharing coordinator for your files. It keeps track of changes to the files so that you can compare your file to older versions or even roll it back to an earlier version. It also allows multiple users to check out and edit the same files, then merges (or attempts to merge) the updated files as those users check them back in.

Subversion provides a very good online book that helped me understand what a version control system is, how to get started with Subversion, and so on. Searchable mailing list archives and a FAQ are also available.

I put together a sort of cheat sheet from what I read and did a test run of moving files into Subversion's control. After that, I was going to go ahead and move the files Nate and I have been working with into Subversion, but I decided to hold off until he gives me the OK because I could easily break something pretty badly.