This post was originally published at Q# – a Wish List for the New Year
In previous blog posts you have read about some of the ideas behind Q#, how it came into existence, and its development over the past year. You have read about quantum computing, quantum algorithms and what you can do with Q# today. With the end of the year approaching, there is only one more thing to cover: What is next?
This blog post is about our aspirations for the future and how you can help to accomplish them. It contains some of our visions going forward, and we would love to hear your thoughts in the comment section below.
One of the most exciting things about Q# for us is the growing community around it. Being rooted in the principles of quantum mechanics, quantum computing tends to have this air of unapproachability to the “uninitiated”. However, quantum computing builds on the notion of an idealized quantum system that behaves according to a handful of fairly easy to learn principles. With a little bit of acquired background in linear algebra, some persistence, and patience when wrapping your head around how measurements work it is possible to get knee-deep into quantum algorithms reasonably quickly!
Of course, a couple of good blog posts on some of these principles can help. We strive to actively support you in the adventure of exploring quantum algorithms by providing materials that help you get started, like our growing set of quantum katas. Our arsenal of open source libraries provides a large variety of building blocks to use in your quest of harnessing the power of quantum. One of the main benefits of open source projects is being able to share your work with all the people brave enough to explore the possibilities that quantum has to offer. Share your progress and help others build on your achievements! Whether in kata or library form, we welcome contributions of any size to our repositories. Let us know how we can help to make contributing easier.
Exchange among developers is one of the most important aspects of software development. It is omnipresent and vital to building a sustainable environment around a particular toolchain and topic. Thankfully, modern technology has made that exchange a lot easier than when the first computer programmers started their careers. We intend to make full use of the power of the internet and give a voice and a platform for discussions on topics related to Q# and quantum computing to developers around the world. The Q# dev blog is part of this effort. Contact us or comment below if you have an idea for a blog post or would like to hear more about a specific topic related to Q#. Establishing good feedback channels is always a challenging endeavor and in particular for a small team like ours. We would like this place to become a source of knowledge and exchange, a place where you can find the latest news and voice your take on them.
This brings us back to our plans for Q#. We have built Q# to make quantum development easier and more accessible. Of course, there were also a couple of other considerations that have played into that decision. For instance, we are anticipating the need to automate what is largely done in manual labor today, e.g. qubit layout and gate synthesis that are often still done on a case-by-case basis for each program and targeted hardware. When is the last time you worried about how error correction works on the hardware your code gets executed on? With qubits being an extremely scarce resource, and the long-term ambition to use quantum computing to address the most computationally intensive tasks that cannot be tackled with current hardware, the optimization of large-scale quantum programs needs to be a priority. We chose to develop our own language in order to have full control and flexibility over what information is represented how, and when it is used during compilation in order to be able to support a modular and scalable software architecture for executing quantum programs. But that’s a tale for another time. What is important is that these considerations are key factors in how we design and develop the language going forward.
A programming language is more than just a convenient set of tools for expressing an algorithm. It shapes the way that we think and reason about a problem, how we structure it and break it down into tasks when building a solution. A programming language can have a tremendous impact on our understanding of existing approaches, as well as how to adapt and combine them for our purposes. Particularly so when venturing into new territory.
Our goal is therefore to build a shared understanding of what it is we strive to accomplish, and to evolve Q# into the powerful language needed to drive progress in quantum programming. Our goal is to leverage the expertise of a community of language designers, compiler veterans, quantum physicists, algorithms and hardware experts, and a variety of software developers to shape a new kind of computing architecture. And we want you to be part of it.
Since our 0.3 release at the beginning of November we have been eagerly working on not just the next release, but on defining and preparing the next steps in 2019. While we are in the middle of formulating our plans for the future, I want to give you a brief insight into some of our considerations.
As I am sure you have noticed, the support for data structures in Q# is minimal. While we do provide quite a few high-level language features for abstracting classical and quantum control flow, we intentionally omit some of the more object-oriented mechanisms such as classes. We anticipate remaining heavily focused on transformations that modify the quantum state, expressed as operations in Q#, as well as their characteristics and relations in the future. However, basic bundling of data and manipulations of such is of course an important aspect of many programs and we want to provide suitable mechanisms to express these in a way that allows to make abstractions, is convenient, and is resistant to coding errors. User defined types in the current setting have limited power besides an increased type safety. The “black box approach” to type parameterization currently restricts their usefulness; we do not provide a mechanism for dynamic reflection and it is not possible to apply operators or other type specific functionalities to argument items whose type is resolved for each call individually. In that sense, these items are “black boxes” that can merely be passed around. We want to do as much of the heavy lifting as possible statically in particular since debuggability of quantum devices is a huge challenge. There are several mechanisms one might consider alleviating the consequences of these decisions. On one hand, type constraints are a common mechanism used in several popular languages. In a sense, they can be seen as “specializations based on the properties of a type”. One could also pursue the stricter path of specializing based on the concrete type itself, de-facto adding a form of overloading that we currently explicitly prevent from being used. Either way, by clearly separating user defined types from tuples in the type system we have made a first step towards extending their power.
If you are curious to hear more about possible ideas for Q#, their benefits and caveats, or want to share some thoughts of your own, comment below! Contribute to the discussion and post your speculations to the question: What makes a quantum programming language “quantum”, i.e. what makes it particularly suited for quantum computing?
I hope you join us into a new year of pushing the boundaries of computation by participating in our coding competitions, contributing to our open source repositories, commenting on or writing blog posts and sharing your ideas and experiences!
How about a new year’s resolution of your own? Let us know what you expect to accomplish and how we can help you achieve your new year’s resolution around quantum programming in Q#!
|Bettina Heim, Senior SDE, Quantum Software and Application
Bettina Heim is a quantum physicist and software engineer working in the Quantum Architectures and Computation Group at Microsoft Research. She is responsible for the Q# compiler and part of the Q# language design team. Prior to joining Microsoft she worked on quantum algorithms, adiabatic quantum computing, discrete optimization problems, and the simulation and benchmarking of quantum computing devices.
This post was originally published at Q# – a Wish List for the New Year