Thursday, June 10, 2010

Interview with Peter Shor

Here's an interview from Dr. Dobb's with Jack Woehr interviewing Peter Shor- as in Shor's algorithm for factoring. Most discussions on quantum computing focus on the physical implementations, or perhaps algorithms. It was refreshing to see a mention of actually programming quantum computers too:

JW: I've seen quantum computer machine language simulators on the web. What would an implementation of Shor's Algorithm in a high-level quantum computing language look like?

PS: Good question. Certainly I can imagine a Fortran-style high-level language that would make Shor's Algorithm easier, but really, you'd want something more high-level than that.

That being said, I don't think the best approach will be a quantum computing language. Too much computation is classical, even when a quantum computer is used. As an example, a quantum computer is used just for a single step of Shor's algorithm- the rest is classical. Preferable to a quantum computing language I think would be a framework that extends classical languages to allow for quantum computation. There are many benefits, but I see the big two as being:

  1. The framework designer can focus on tackling quantum computation. They don't have to try to include classical techniques that we've spent decades refining.
  2. There is a large user base for the popular classical languages. By creating a framework there is less of a hurdle for those programmers to perform quantum computation. Furthermore it will integrate much better with existing code bases.
Of course, I developed Cove as a framework for quantum computation exactly because of the reasons above.

No comments:

Post a Comment