Responses - Wordle, Mean Mr. Mustard

A reply to Jonathan Feinberg first, as I personally offended Wordle, much to Klund's amusement. This is presumably what he took umbrage with "Wordle is cool. It's just not accurate." So I'll clarify, although up until that point in the post I'd been pointing out that the cloud generated didn't capture the essence of my blog. Not that Wordle had a problem, only that the resulting cloud wasn't accurate in the sense that it wasn't representative - e.g. as dissociated from the Wordle program. That may sound like criticism of Wordle, but it's not. Wordle did it's job with what it had - that doesn't make the result any more indicative of the content of NodToNothing.

My real issue, and I could have phrased it better, is that when I pointed at a completely different month by putting that month's URL into Wordle, in order to grab the details that weren't related to the last 100 posts in the Atom feed and generate something a bit more representative based on what I hoped was a more representative month squarely in the middle of geocaching and bicycling season, Wordle instead regrabbed the Atom feed for the last month, not the content from the month I selected. That's understandable as well. After all, if it just grabbed the URL and went on it's merry way, it would probably have to parse out all the HTML and CSS-type tags. I sympathize with using a structured Atom or RSS feed and forgoing that mess. I could resolve a bit of that by using the paste-into-a-textbox function, but that was painful as I ended up grabbing the tags (meaning some words were counted twice) and some of the template information, which didn't turn out well. My assertion that the Technorati cloud was more accurate was in no way a reference to a program or algorithm, but only to how Technorati tracked my tags, resulting in a cloud based on what I had personally tagged as important in my blog and that seemed to capture keywords more appropriate to what it's about: bicycling, geocaching, Eryn, beer, and for some reason, the City of Rosemount.

To give Mr. Feinberg a nod (to something, not nothing), and perhaps invite a retaliation from Dave Sifry, Mr. Feinberg's Wordle program a.) creates more aesthetically pleasing clouds and b.) is no doubt more accurate in assessing the actual content cloud of my blog from an appropriate dump of text than Technorati's cloud is, which relies on my own ability to tag my content with appropriate metadata, a skill of questionable ability in anyone who isn't an information architect with a library science background.


I have a much shorter response for Mean Mr. Mustard, who commented on the post Pain. Despite my choice of language in the post, it was still a better call overall than telling an architect at work today, "You can blow me off any time. You're only two minutes away." Fortunately, I've switched jobs enough times that no one realized, or was unwilling to comment upon it, that I was snickering at my own stupidity as I transitioned to a new topic. And, on a more personal note directed specifically at you, I appreciate that I have so many a**holes in my life, so that only one had to have hemorrhoid surgery. I hope your back feels better soon.

Wordle

My coworker Alex (I'll get a link out here for him soon now that his blog is public - I haven't sent the address home yet) sent me a link to Wordle and a word cloud he made from my blog. I've looked at Wordle before and there's a tag cloud somewhere on the blog in the past, but it's always fun to rerun those things to see if it looks different than the last time.

Here's the image Alex generated. I looked at it for a while and decided it only really captured the last month of my blog. It looks like I'm obsessed with Lemony Snicket's Series of Unfortunate Events, instead of obsessed with alcohol, sex, and bicycling, which is really the case.


So I ran it again trying to tag the last 100 posts by using a parameterized blogger url. This seemed a little more accurate. I do talk a lot about Eryn. I do put a lot of photos and pictures on my blog. But Christmas is limited to a single month. And surgery was the topic du jour for December, but otherwise it's not a consistent blog topic. And Sparky and Hannigan were one-post topics. So clearly, this wasn't capturing the essence of my blog either.


Wordle is cool. It's just not accurate. Fortunately, I remembered that Technorati does a word cloud that focuses on your tags, not your words from the last # posts. This seems more accurate, although less aesthetically pleasing. I'm willing to believe that bicycling, geocaching, friends, humor, eryn, memes and "postpourri" occupy quite a bit of my writing. I'm disappointed "Mean Mr. Mustard's Sister" doesn't make the cut.