11/16/2018 The PL Part of CS2013 Dan Grossman Department of Computer Science & Engineering University of Washington SPLASH-E October 28, 2013

Before we dive in… History in 1 line: CS2001 generally reviled by the PL community Much work by others to lay groundwork to “fix it” If you don’t like CS2013, it’s my fault  Don’t fret over “what is PL” in “Body of Knowledge” “PL” includes lots of “programming” and all of “compilers and run-time systems”… That’s “good” if selfish about the writing Less “good” if selfish about the hours But the steering committee was thankfully not selfish Experts will never be happy  Imagine a department with 5 CS faculty, none in PL Curriculum must be implementable and [only] a bit forward-looking October 28, 2013 Grossman: PL in CS2013

Acknoweldgements, part 1 Started early with a diverse committee of reasonable size Result much better than I could have done alone Kim Bruce Curt Clifton Ernesto Cuadros-Vargas Kathi Fisler Kathleen Fisher Stephen Freund Matthew Hertz Johan Jeuring Doug Lea Simon Thompson October 28, 2013 Grossman: PL in CS2013

Acknowledgments, part 2 Then many fresh eyes to avoid (?) committee-speak Guilt-trip emails from me and public-comment periods Alex Aiken Ron Cytron Tony Hosking James Noble Jens Palsberg Benjamin Pierce Michael Scott Michelle Strout Stephanie Weirich Steve Zdancewic … October 28, 2013 Grossman: PL in CS2013

High-level goals Language design, use, and semantics side: Less OO-focused Encapsulation not only a sub-topic of OOP Benefits of avoiding side-effects, using closures Modern parlance, understanding, and breadth No “compiled vs. interpreted languages” Rational arguments on static vs. dynamic typing Language implementation side: Program analysis for more than compiler optimization Increasing importance/sophistication of run-time systems Matter of principle + inclusiveness + timeliness: No PL mentioned by name in Body of Knowledge October 28, 2013 Grossman: PL in CS2013

Interactions All of OOP, FP, and Event-Driven is in PL But recursion, decomposition, abstraction is in SDF Because Body of Knowledge won’t “pick a paradigm for CS1” Concurrency and parallelism mostly elsewhere (PD and SF) Committee convinced me to add a PL unit for things like actors Of course, PL touches in everything in computer science  October 28, 2013 Grossman: PL in CS2013

The “core” pieces FP First-class functions + lexically scoped environments Mutation-free programming OOP As an abstraction mechanism Subclassing Event-driven Program Representation: Code is data “Basic Type Systems”: what they do [not] do “Basic Translation Execution”: What a compiler/interpreter/run-time does, but not how to implement one Key details like how recursion is implemented October 28, 2013 Grossman: PL in CS2013

My one axe to grind Matrix of operations and data variants Cf. “the expression problem” Apologies if this came out “how Grossman teaches it” eval toString hasZero … Int Add Negate October 28, 2013 Grossman: PL in CS2013

Elective != unimportant, part 1 Help modernize “the compilers course” While acknowledging lexing/parsing still exists Separate code generation from static analysis New run-time system unit with GC, JITs, VMs Mostly aspirational at this point?? Would love to discuss how to make this happen October 28, 2013 Grossman: PL in CS2013

Elective != unimportant, part 2 “PL side” has units that justify foundational treatment… Formal semantics, type systems, etc. … and units that justify “advanced programming”… Control operators, laziness, macros, mixins, … Reflects a big-tent there-are-many-good-courses view Now let’s fight over “better” with course exemplars and reusable materials October 28, 2013 Grossman: PL in CS2013

In the end… SIGPLAN written SIGPLAN and broader community approved Surely not perfect Don’t ask the cook how the soup tastes  How can we use this to best influence courses/curricula? How can we best prepare the field for an even better curriculum in 2023? Need to successfully connect advances to the classroom! October 28, 2013 Grossman: PL in CS2013