How many cows in Texas?
Somewhere in the neighborhood of 13,900,000 as of January 2004, according to the experts on how many cows in Texas, the USDA National Agricultural Statistics Service.
Why am I telling you that? Apparently (judging by my usual casual research == a tiny bit of searching Google), that question has been used as an example of why we need the Semantic Web, or Intelligent Search Agents, or Whatever Someone Wants To Push In Search, for some time now, because of course such crappy search terms don’t pull up the answer. It’s gotten famous enough that the Washington Post uses it as an example of how search doesn’t work, because the top result in Google doesn’t answer the question. (Aside: the question is more often phrased as “how many cattle are there in Texas”, but the WP’s source probably didn’t give that phrasing because all the results are for Semantic Web presentations and W3C mailing lists: bit too Heisenbergian for general consumption.)
So, the Washington Post says that the first result for “how many cows in Texas” is an article on mineral supplementation of beef cows, as though somehow Google’s results are fixed for all time. Gee, I wonder if anything could affect that, other than having a Semantic Web or Intelligent Search Agents? Could, perhaps, someone simply link from the question to the answer?
So we need more intelligent Search-Agents instead of more intelligent searchers?
Well, my thesis (to the extent I had one) was that we need more ordinary people linking to things with ordinary terms. But the first page of the article, where my blood began boiling, brings up the point that one of the things we still really need are intelligent search agents. Like the nice ones who work at that building with the stone lions out front.
And if you think there’s any chance of getting more intelligent search terms, you clearly didn’t work the desk at the library as an undergraduate ;)
If nothing else, it makes for an interesting casual experiment to show how Google’s indexing at the moment. Right now, the top result for ”how many cows in Texas” as a phrase is my previous entry, thanks to the ”next” link, with number 1.5 being my February archive, and number three is DJ’s news reader. I wouldn’t have been surprised to see my entry indexed rather than my link target, that seems to be the way the current scheme works, but I am a bit surprised that they are so cautious about being bombed or spammed that they’ll take a page pointing to a page pointing to, just to avoid it. Given that many of their problems arise from wanting to get too close (empty trackback page must be better than the full entry page), that’s surprising distance. Though it may be the difference between Fresh and Really Indexed.
Not as a phrase (a difficult search, since how many cows in Texas is the same search as * many cows * Texas but not the same as many cows Texas)? Not high enough for me to spot before I got bored with looking. Just trying to help you, Google, cut me some slack.
Right now, the top result for â€how many cattle in Texas†is this sight. So I found the info I was looking for on the first try. (But looking at the rest of the list, I can see it was a fluke and I got lucky, but thanks for posting the answer.)