ResearchBuzz

WatchMouse last week launched a new site called API Status, available at http://api-status.com/. This Web site is very simple; it just monitors the status of APIs.

The site monitors the performance of 26 APIs including Facebook, Twitter, YouTube, Amazon, and PayPal. When you visit the site you’ll get a listing of the APIs with their current status, as well as a table showing the history of the API performance of the last seven days. Three icons denote whether the API is running okay, having issues, or down completely.

If you want to dig a little deeper, you can click on an API name. There you’ll get a bunch of details about the API’s performance, including a graph of its availablity over the last seven days and how it’s performing globally.

This site may not be much use to you in your daily research, but if you’ve ever been in a situation when you’re trying to import XML
into a Google spreadsheet and you can’t get it to work properly and you’re about to START THROWING THINGS — ahem, not that this has ever happened to me of course — it’s good to have a place you can quickly check an API status and better yet see how it’s been doing for the last 24 hours (because maybe you should be working on something else!)

Last Friday, Google announced a couple new features that should make it easy to extract information from the general hodgepodge of search results. One of them is for factual data and one of them is for events.

The first feature is called “Answer Highlighting” and is invoked when you search for factual information. Google has an example in its blog post: ask Google How tall is the Empire State building? and you’ll get the answer highlighted in your search result snippets. Sadly I didn’t have as much luck with this search. Washington Monument height got me relevant results, but the fact of its height was not highlighted. I didn’t have much luck with Statue of Liberty height either.

I tried a search for Abraham Lincoln died and got this at the top of my search results:

I don’t know if this is supposed to be the highlighted search results or not. I do know I didn’t have any luck finding highlighted answers in my search result snippets, using variations of Google’s examples. (And I also wasn’t sure why there one of the examples was in the form of a question and the others had more simple syntax.) Probably for simple fact questions I will continue to use Wolfram|Alpha. Not only can I get the heights of buildings there easily, but I can enter a query like “Empire State Building height minus Washington Monument height” and get an answer!

The second feature is for pulling events out of search results. This is one of Google’s “Rich Snippets.” When you search for a site that supports this snippet, you’ll get an event listing right in your search result, like the example for Irving Plaza:

Note that as a “rich snippet,” this feature requires Web wranglers to insert special markup in their Web pages. Since Google just announced this feature you will not find event listings in your search results for a while. If you yourself are a Web wrangler and have events on your site, you can get the documentation on the new search snippet here.

I didn’t have much luck with the answer highlighting, but I like this new events rich snippet. It’ll be even better if that event data also ends up on the Google Maps listings!

The other day as I was poking around the Internet I kept coming across the following date: December 29, 1967. If you are not a Star Trek person, you might not know that that’s the date that the Trouble With Tribbles episode was first aired (at least according to the many sources I kept coming across — I have not cross-checked and alas, I am not a Star Trek person.)

This amused me and I wondered exactly how much time had passed since that episode aired. Google can do some mathematical calculations so I figured I might as well try it for dates. So I went to Google and tried:

December 29, 1967 – December 29, 2009

That got me results — over 12 million of them as a matter of fact — but nothing about date calculation. But you know, as soon as I saw the results I thought to myself, “I bet Wolfram|Alpha could calculate these dates for me.”

And I was right!

I went to http://www.wolframalpha.com/ and entered December 29, 1967 – December 29, 2009 and WA gave me the following output:

As you can see, WA calculated that I was looking for the space between two dates and gave me the calculations in years, weeks, and days. It also helpfully informed me that December 29, 1967 was on a Friday.

After I played with this for a while (finding out *exactly* how many days you’ve been alive may make you a little tired) I wondered if WA could do calculations based on dates. For example, would the following query work?

Fifteen days after December 29, 1967

Yes, that does work. It gives you the date (Saturday, January 13, 1968) calculates the time difference between that day and this one, and then proceeds to give the information it normally does about a single date (notable events, date formats, moon phase, sunrise and sunset.) It works if you’re looking for dates counting backwards from a specified date as well.

I decided to through a curve ball at WA. Since it lists the moon phrases, I figured it should be able to calculate dates based on the full moon. So I asked it:

Fifth full moon after December 29, 1967

And it gave me an answer! According to WA the fifth full moon after December 29, 1967 was 9:06:35 am EDT Sunday, May 12, 1968.

I tried asking WA about the first solstice after December 29, 1967. It got that one right. I finally tripped it up asking for the first sunrise that was exactly 7:05 AM after December 29, 1967. And even now I’m wondering if my syntax was screwed up.

Wolfram|Alpha is a great compliment to Google; they’re not competitors. Being able to play with dates is just one more proof of that.