The Quantum Architecture and Computation group launched Q#, our quantum computing programming language, a year ago on December 11th, 2017.
Q# 0.1 was the result of a lot of hard work from a small, dedicated team of developers, researchers, and program managers. We had made the decision to build a domain-specific language for quantum computing about six months before we launched, so we were on a very tight schedule. We were lucky to have a great team of people who all pitched in and did what needed to be done so that we could meet our extremely aggressive timetable.
Start!
Inside the team, we speculated on what level of interest Q# would attract. We hoped that we might receive a few hundred downloads, but we were blown away when we crossed 1,000 users by about 9 hours after launch. That said, with so many users installing the Quantum Development Kit and trying to write simple programs in it, bugs started popping up. In order to deliver the best experience for our users, we released a patch in January that addressed issues like floating-point literals that were handled incorrectly in certain locales, and allowed the simulator to run on older machines without vector instructions support.
We also addressed portability feature requests in our 0.2 release in February 2018, which saw us move from the .NET Framework to the open-source, cross-platform .NET Core. This allowed us to easily support macOS and Linux as well as Windows for building and running Q# code. We also added support for VS Code on all platforms (the 0.1 release was limited to Visual Studio on Windows). As part of the 0.2 release, we were able to make the majority of our libraries and samples available under an MIT license.
Long Hot Summer
We decided to take advantage of one of our team members’ expertise in organizing coding competitions and run a Q# coding competition to engage non-quantum developers with Q# and quantum computing. After a couple of months of preparation, we ran the Q# Coding Contest in early July. Again, the results exceeded our expectations: 514 participants in the warmup round, and 389 in the actual contest. 100 participants solved all the problems, and a lot of them even asked for more challenging ones!
To help make Q# and quantum computing more accessible to the public, we also launched self-paced programming tutorials: the Quantum Katas. We’re up to 10 katas already, and more are coming!
Spring, Summer, Autumn
We started planning the next major release in the spring of 2018, after shipping our 0.2 release: we wanted to rebuild our compiler to work as a language server, to give Q# developers the same interactive error checking and IntelliSense features they’re used to for languages like C# and F#. We knew this would be a huge amount of work and would require a significant re-architecture of the compiler in order to work incrementally. We didn’t want to wait longer to do this work, though, because we wanted to give our users the kind of modern programming environment they’re used to.
We spent the spring and summer re-architecting and rewriting the Q# compiler and shipped the new Q# compiler as our 0.3 release at the end of October.
The 0.3 release also includes a new, open source quantum chemistry library. This library integrates with NWChem, a powerful and popular open source computational chemistry package. The integration is based on the open source Broombridge schema.
Whatever Next
What’s next for Q#? No spoilers (yet!).
The last blog post of the calendar, scheduled for December 24th, will look at some of the things we’re considering for Q# in the coming year.
Until then, enjoy the holidays!
Congratulations to everyone who can figure out what the section titles have in common…
0 comments