12 March 2008

Here We Go

Since thoughts worth preserving have been growing about as quickly as glaciers, I haven't really done a lot with this new, shiny blog, but now I have something that's slightly above filler quality, so away we go!

Itihasa posted a fascinating little gender-guessing widget. Check it out; the principle upon which it is supposed to work is interesting enough to toss me along all sorts of speculation that I couldn't begin to substantively answer.

So, what else was there to do but start feeding it material and seeing the verdict?

My writing was first, with appropriate selections taken in a decidedly non-random fashion from my old/other blog. The gender guesser nailed it as male or weakly male every time, which quite frankly pissed me off because I like the idea that I can pretend to be anyone when I'm on the internet. Oh well; I'm a guy, I guess that secret is out there now, and even a stupid bot can figure it out.

But what about the things that I read?

I'm pretty sure that Tamara K is a lady/woman/distaff whatever the preferred term for that is these days, but the gender guesser pegged samples from the blog as male every single time. Likewise Kathy Jackson of Cornered Cat. OK, though, both of those women write about guns and self defense, which is admittedly a male-dominated subject, and the gender guesser did say that it uses word frequency analysis in its inscrutable machinations, and that could be throwing it off.

So then I fed it some paragraphs from Mary Shelley's Frankenstein, only to be reminded that the algorithm only works on American idiomatic English.

Apparently some of the Americans I read don't write like Americans either, since the guesser returned the weak option for several posts form Stephen Bodio's Querencia and Unqualified Reservations.

The gender guesser couldn't nail down a solid answer for Heather Corinna either.

Is that a large enough sample size to draw any conclusions from? I would think not, but it does amuse me to think that everything I read is written by a man, or cleverly disguised to look like the same.

Edit: The widget says that this post is only weakly male. I wonder why?

3 comments:

Tam said...

I know my language has been colored by my career and by mostly having guys for friends.

I knew it had contaminated my vernacular the time my friend Jen started laughing at me for frustratedly cursing a recalcitrant Garand by calling it a "c**ks**ker".

Steve Bodio said...

Are we foreign or gender ambiguous?! Heve never had this problem before.

I think this software needs tweaking.

Neutrino Cannon said...

Since it's based on a simple word-frequency analysis, and doesn't claim especially astonishing powers of perception, a more practical program might be written from the ground up entirely.