Over the next few months you might notice this blog transforming into its original purpose: a dumping ground for me to post quick little updates on what I personally do. This is the start and I’ll hopefully catch you all up on some things I’ve worked on over the last year here and there. I’m never actually working on just one project, but always working on several small ideas here and there.
About four years ago I found thinkmap’s visual thesaurus. I always wanted to make something similar, but I’ve always lacked the word smithery to truly understand the English language enough to know if what I had created is actually working or not. To this end, I’ve done the technical stuff like writing a jquery plugin to handle visually displaying data in the same method.
It’s sort of abandoned for now though. I only ever work on it when I have an interest in it. So far I’ve got the front end stuff done, or at least the hard bits. Throw a bit of formatting over it and a quick api to look up a thesaurus database and you’re away. One day, I hope to finish this project off.
The testbed
http://phogue.net/projects/ikefactory/src/examples/testbed.html
Just showing what the script is doing. I recommend hitting “Add Random Branches” to begin with. You can drag nodes around and even pan the view if you hold shift. Zoom was buggy and on the to do list.
Twitter example
http://phogue.net/projects/ikefactory/src/examples/twitter.html
Shows mutual followers, or followers whom you follow. If you double click on a node you will then load their mutual followers and it will interconnect with existing nodes. Chunky as hell when you start getting a few hundred nodes on the screen or just some one that’s really sociable.
Moving some of the workload onto a worker thread would be beneficial here, but would also require a rewrite lest the serialization of the plugin itself takes just as long as calculating the node positions & effects. I wanted the plugin to handle thousands of nodes but only managed to get to handle about 500 smoothly on my computer, but it significantly deteriorates with lower spec computers.
So I had lofty goals to begin with however building the twitter plugin quickly made me realize that formatting data in this way is only helpful for a tens, not hundreds let alone thousands of pieces of data. After a hundred nodes it just starts looking pretty but not being very informative.