The Pyret Code; or A Rationale for the Pyret Programming Language
We need better languages for introductory computing. A good introductory language makes good compromises between expressiveness and performance, and between simplicity and feature-richness. Pyret is our evolving experiment in this space.
Why not just use Java, Python, Racket, OCaml, or Haskell?
One of the enduring lessons from the Racket project is that no full-blown, general-purpose programming language is particularly appropriate for introductory education. By the time a language grows to be useful for building large-scale systems, it tends to have accumulated too many warts, odd corners, and complex features, all of which trip up students. The journal paper for DrScheme (the old name for DrRacket) explains this in some detail.
In that respect, the closest fellow travelers of us Pyreteers are the Racketeers (see how that works?). In fact, the first version of Pyret was merely a
#lang in Racket. Nevertheless, Pyret represents a departure from Racket (for now and for the near future, at least) for several reasons:
We wanted to experiment with syntax. Much as many of us love parentheses, we fear that Racket will always bump into an acceptance threshold due to its syntax. We’re delighted to see the growing adoption in industry of languages like Racket and Clojure, and maybe the days of paren-phobia are over. But academics are far more hidebound than industrial programmers!
Though our pedagogy draws from and is very similar to that of the Racketeers, it also diverges in some ways, reflecting the somewhat different backgrounds and preferences of team. Pyret embodies our educational philosophy. We felt we could more easily experiment if we had a clean-slate design than if we had to keep fitting our work into the constraints of Racket.
#langfacilities, though designed to create new languages—and a great prototyping ground for Pyret—proved to not be quite enough to support a language creation process of the scale of Pyret; and the need for a strong Web run-time system also meant that we did not get enough out of the Racket ecosystem. Nevertheless, that only represents the state of systems today; if someday the two grew closer together, many or all of us would rejoice.
Will Pyret ever be a full-fledged programming language?
Broadly, we view building an awesome teaching language as a useful design discipline that necessarily entails all the features typically expected of a mature language. When building a language, you’re constantly confronted with questions: what to do next, how to do it, and whether to do it at all. Language designers therefore need a mental framework that guides their choices. For some, it might be “Build an awesome language for shared-memory concurrency!” For others it might be “Build an awesome language for writing scientific computations very, very concisely!” For us, it’s “Build an awesome teaching language!” So whenever we’re confronted with a design (or implementation) choice, we first ask how it would play out in this specific context. Then we ask how it would impact our long-term goal. We have yet to find actual contradictions between the two, but it’s certainly re-ordered priorities.
There are lots of kinds of “education”.
That’s right. We are focused on introductory programming education at a high-school and collegiate level — what is often called “CS 1” and “CS 2” (roughly, the first year of college). Pyret is being actively used in everything from high-schools to upper-level collegiate courses, giving us a tight feedback loop.
Of course, even in that setting there are differences of opinion about what needs to be taught. Some believe inheritance is so important it should be taught early in the first semester. We utterly reject this belief (as someone once wisely said, “object-oriented programming does not scale down”: what is the point of teaching classes and inheritance when students have not yet done anything interesting enough to encapsulate or inherit from?). Some have gone so far as to start teaching with Turing Machines. Unsurprisingly, we reject this view as well.
What we do not take a dogmatic stance on is exactly how early state and types should be introduced. Pyret has the usual stateful operations. We discussed this at some length, but eventually decided an introduction to programming must teach state. Pyret also has optional annotations, so different instructors can, depending on their preference, introduce types at different times.
What are some ways the educational philosophy influences the language?
Pyret is driven by the How to Design Programs (HtDP) philosophy of programming education; one could almost view it as a language designed to make teaching from HtDP comfortable. However, we also have our own curricular ideas that extend or run parallel to those of HtDP, which we are steadily writing up in Programming and Programming Languages (PAPL). PAPL both exclusively uses Pyret and is slowly turning into a definitive book for teaching with it.
Some of our central ideas include:
Languages should offer a rich set of values. For instance, images should be primitive values, fully supported by the environment. This improves the teaching of several things, from image-based composition exercises (which are often far more interesting than composition over numbers) to animations and games.
Languages should offer syntactic support for as much as possible of the HtDP design recipe. (Pyret does not offer explicit support for templates, but then again we’re rethinking the pedagogy of the template step.)
Design recipe support includes good data definition primitives, including the expression of invariants. Therefore, Pyret has support for refinements (which are currently checked dynamically; a future version may add static support, but this is not our highest priority). In particular, by allowing algebraic datatypes to also be objects, Pyret offers significant flexibility and expressiveness.
Students should be able to easily write examples as a precursor to writing code. In this regard, Pyret borrows but (we believe) syntactically improves upon Racket’s
check-expectfamily of operations with examples and testing.
Students should easily be able to escalate from simple examples to sophisticated testing, such as writing test oracles.
Equality is subtle and languages should respect its nuances. Students must understand the meaning of equality and its consequences, especially with regard to observations in the presence of state. Therefore, Pyret has carefully developed equality primitives.
Is Pyret also a research project?
Yes and no.
Yes in that, broadly speaking, we view what we are doing as very much a research effort. We have specific, novel goals in mind (some of which we’ve described elsewhere in this document), and finding good solutions to achieve those goals will require innovation. This is the very essence of research. Some of that research is on the curricular side, though, and may not require any language innovation. Indeed, as much as possible we want Pyret to remain a language that is easy to approach and understand.
However, “research” also has a narrow meaning in academia, in the sense of publishing lots of papers. This is an un-goal, and almost an anti-goal. Every paper takes months to write, and those are months we are not spending on the language and curriculum. We view this as a poor time trade-off. We’ll write papers sometime, but our immediate focus is on shipping code and textbooks and other materials of immediate value to students and educators. If this means we get “scooped” on some of our technical innovations, that’s just fine; we consider that a good trade-off.
Ultimately, though, research requires doing new things. We don’t want to do new things just for the sake of it. We’ve long been inspired by this quote from Paul Graham:
The difference between design and research seems to be a question of new versus good. Design doesn’t have to be new, but it has to be good. Research doesn’t have to be good, but it has to be new.
For us, design is essential; research is optional. We actively dogfood Pyret both as teachers (actively teaching with it) and as programmers (actively building things with it), which gives us ongoing feedback about its points of friction. Where necessary we have done and will continue to do research, but publishing academic papers about it — rather than putting it to work educationally — will continue to take a back seat.
We are currently hard at work on three very important features:
Support for static typing. Pyret will have a conventional type system with tagged unions and a type checker, resulting in straightforward type errors without the complications associated with type inference algorithms. We have carefully designed Pyret to always be typeable, but our earlier type systems were not good enough. We’re pretty happy with how this one is going.
Tables are a critical type for storing real-world data. Pyret is adding linguistic and library support for working effectively with tables, which PAPL will use to expose students to “database” thinking from early on.
Our model for interactive computation is based on the “world” model. We are currently revising and updating it in a few ways that will help it better serve our new educational programs.
If you’d like to stay abreast of our developments or get involved in our discussions, please come on board!
On the educational side, Pyret is already used by the Bootstrap project. We are now developing three new curricula for Bootstrap:
A CS1 curriculum, corresponding to a standard introduction to computer science, but with several twists based on our pedagogy and materials.
A CS Principles curriculum, for the new US College Board Advanced Placement exam.
A physics/modeling curriculum, to help teach students physics and modeling through the medium of programming.
If you’d like to talk more about these, join us!
Why is your filename suffix
Because pirates supposedly said “arrr!” (or “yarrr!”) a lot. (Sometimes they sing, too.) Note that you must pronounce it with a guttural sound, so it can’t be mistaken for a file written in the R programming language. You should consider swashing your buckler as you speak.
Does the skull in your logo have a name?
Yes, she’s Bonny!
Where is the proximal end of the second femur?
Look closer. It’s a lambda.