My blog has a new theme! I spent a few hours today working on a design from my friend Dima at TYZ Design, and I’m pretty happy with the results. I plan on making some small changes to the appearance of links, and fix some XHTML validation errors in the next few days.
jQuery and upcoming features
While I have used jQuery (a javascript utility library) in the past, I generally have fallen back to the Prototype JS library. This was mostly because I was already familiar with it from my work on FamilySnap, and because it can be difficult to find the time to fit in learning another tool when it’s not completely necessary to do so.</p>
read more
New Site Design
For those who have seen the previous versions of our site, I hope you will like the new site. My thanks goes out to Dima at TYZ Design for the original design, layout, and graphical elements of our new site!
Dia2POG
I’ve been busy working on several PHP/MySQL web development projects. For the last few of them, I have been using PHP Object Generator to generate all of my Database Abstraction Layer/ORM code. One of the new things I have been trying out is modeling my database using UML. I used to sit down with phpMyAdmin and do my database design directly in SQL, but since I am going through POG to create the database, that wasn’t an option anymore.
So I started using Dia, and started modeling my tables as objects. That still left me with entering the field names from the diagram into the form for my local copy of POG. So I recently decided that I should write a script that transformed the dia XML format into a SOAP request to POG (POG has a built-in SOAP interface to its object generator), and then create the files for each class with the returned code.
I’m releasing the code to my little python script. There isn’t much error checking (or comments), and it may break on certain diagrams. Read the source for more details.
Munkres code v2.
As promised, here is the updated version of my Munkres
algorithm code. The memory leak issue has been fixed, as well as the problems with the code sometimes failing to find a linear assignment for larger matrices.