How Is Wolfram|Alpha Like a Matryoshka Doll?

Wolfram|Alpha is different from most of the tools out there on the web that you might use to get answers. Rather than inundate you with lists of links to web pages that may or may not be useful, Wolfram|Alpha works to understand your query.

What really sets these different approaches apart is how they deal with complexity in queries. Whether there are many concurrent factors to your question or you have a unique math computation with an answer that simply does not exist on some web page, Wolfram|Alpha is your best bet for a web service that actually understands what you are asking.

One of the ways that complexity can appear in queries is in depth, when there are multiple steps to a question. To understand what we mean by “depth,” think of the beautiful Matryoshka dolls that all fit inside of each other.

To answer a query like “elevation of Steven Spielberg’s birthplace“, the first step is to recognize that “Steven Spielberg” and “birthplace” can be combined to form a location.

(elevation) of ((Steven Spielberg’s) (birthplace))

We can then combine that location with the “elevation” and form a nested statement, and now we’re off to the races:

These kinds of questions can be quite natural:

• “Population of the capital of Italy“
• “Bright stars in Rigel’s constellation“
• “Spin of the antiparticle of muon“

And Wolfram|Alpha can handle multiple layers of this kind of nesting:

This sort of thing can quickly begin to feel contrived. But that’s OK, and if it makes you feel better, you can add a little more structure to help see the relationships in the input, such as in this query: “angle relations (crystal system (symmetry group of a tetrahedron))”

But what happens if the inner combination is not a single thing, but rather multiple things? Wolfram|Alpha still knows what to do:

• “Birthdays of the discoverers of Neptune“
• “GDP of the bordering countries of France“
• “Most common words in the acts of The Merchant of Venice“

Pretty cool, right?

There is still plenty to do in this realm. We’re ensuring that the links between our datasets all conform to the standards required for this kind of nesting; building up the natural-language capacity to handle qualifiers (such as dates for nested population queries); and, of course, continuously extending the relational patterns that Wolfram|Alpha can recognize. And so we remain hard at work building up the computable capacity of the world’s only real knowledge engine.