Typelevel Summit Boston

http4s: pure, typeful, functional HTTP in Scala

at 9:05

http4s embraces cats.data.Kleisli for service definitions and fs2.Stream for payload processing. With its foundation on these simple types, we’ll take a whirlwind tour of how http4s can:

• plug into your functional business logic
• snap together with other functional libraries
• relate servers to clients
• test with a simple function call
• run on multiple backends
• support multiple higher level DSLs

This talk will not make you an expert in http4s, but show that it’s a relatively short path to become one.

About Speaker

Ross is a Senior Software Engineer at Takt. He began his open source journey on the Scalatra project in 2009, and has gotten purer, more functional, and more typeful with each passing year. He now contributes to http4s and cats among others, and is a member of Typelevel. He is a co-organizer of IndyScala.

Opaque types: understanding SIP-35

at 9:55

Proposed in SIP-35, opaque types introduce a way to define types which only exist at compile-time. Despite some superficial similarities to value classes, opaque types are significantly more flexible and introduce a number of exciting new possibilities in the Scala design space.

Opaque types are motivated by a number of different concerns:

• desire for a non-class type that exists only at compile-time
• efficiency concerns with value classes
• limitations of existing type aliases
• need to better support phantom types, type tags, etc.

This talk will introduce opaque types, compare them to type aliases and value classes (their two nearest cousins) and then walk through some examples of using opaque types. The focus will be on advantages of using opaque types versus other encodings, including looking at how various types are represented by the JVM at runtime. The talk does not assume in-depth knowledge of the Scala compiler and will motivate the code using plausible real world examples.

Attendees will come away from this talk with a better understanding of what SIP-35 means, why it was proposed, and how it could change how we write Scala code for the better.

About Speaker

Erik Osheim is one of the founders of Typelevel, and maintains several Scala libraries including Cats, Spire, and others. He hacks Scala for a living at Stripe, and is committed to having his cake and eating it too when it comes to functional programming. Besides programming he spends time playing music, drinking tea, and cycling around Providence, Rhode Island.

Big Data at the Intersection of Typed FP and Category Theory

at 10:30

Big data, functional programming, and category theory aren’t just three trendy topics smashed into a talk title as bait! Foundational ideas from typed functional programming and category theory have real and practical applications for working with big data and can also be utilized to write more principled pipelines at scale. Whether it’s aggregating with monoids or writing more typesafe Spark jobs, we’ll try and bridge these topics together in a way that can be immediately useful. Some knowledge of Scala and a big data framework like Apache Hadoop, Spark, or Beam is suggested but not necessary.

About Speaker

Long Cao is a software engineer focusing on Scala, Spark, and data engineering and has been in New York for the last 5 years. He cares deeply about showing newcomers the benefits of Scala and functional programming. On his off time likes to enjoy climbing, Rocket League, music, sports, and coffee.

Tracking with Writer Monad

at 11:20

This talk will tell the story of one team at eBay which used to do data tracking in a healthy side-effecting manner. Until the team realized that it’s not that healthy. The solution was found in a Writer Monad (residing in the cats library) as well as in the fact that the writer monad can stay in shades. Some people, especially when they are new to typed FP, don’t like/feel comfortable to see words like Semigroup, Traversable, Writer and such in their domain code. The talk will show how those “scary” parts can be “hidden” by domain specific extension methods.

About Speaker

I’m a Scala Dev at eBay Inc, working commercially with Scala as my main programming language since early 2011, Java Dev in my past life. Occasional contributor to variety of Scala related OS projects. Scala evangelist who successfully converted individuals and teams to Scala from Java.

Duality and How to Delete Half (minus ε) of Your Code

at 11:40

In functional programming, we often refer to category theory to explain various concepts. We’ll go over where these concepts do and don’t map well to Scala, as well as what duality is, how we can take advantage of it in Scala, and how to distinguish other concepts that are often confused with it.

About Speaker

Greg Pfeil is a compiler writer and programming language designer. At this point they have written four recursion scheme libraries in three languages, with the goal of getting to half a recursion scheme library for all languages. Greg works at Formation, writing Haskell, using only the purest artisanal FP.

Keynote: Planning for Rainfall

at 14:00

Soloway's Rainfall problem, a classic benchmark in computing education research, has proven difficult for many CS1 students. Rainfall tests students' abilities at plan composition, the task of integrating code fragments that implement subparts of a problem into a single program. Nearly all early studies of Rainfall involved students who were learning imperative programming with arrays. Over the last few years, we've conducted studies with students who were learning functional programming instead. These students have produced atypical profiles of compositions and errors on Rainfall (and similar problems). What do these results suggest about the role of programming languages in novice programming education? This talk raises various questions about the relationships between programming languages, program design, curricula, and how students perceive code structure. The talk assumes no experience with having been rained upon.

About Speaker

Kathi Fisler is a Research Professor in Computer Science at Brown University, and co-director of Bootstrap, a national-scale K-12 project that integrates introductory CS into existing middle- and high-school classes. She spent many years doing software and security verification research before deciding that people were harder (and more interesting) to model than systems. She is currently on a mission (with partner-in-crime Shriram Krishnamurthi) to explore how classical CSEd studies might have turned out differently had they considered functional programming. She's teaching with Scala for the first time this semester.

Why Monads?

at 15:20

Monads remain a somewhat mysterious concept in Functional Programming, even though the number of tutorials and blog posts trying to “monadsplain” is at an all-time high. Rather than answering the classical question “What is a Monad?”, we are going to dig more into “Why Monads?”. Building intuition on why monads are useful will help better understand what they are as well. We’ll start with a simple function in a monadless world and we’ll see how annoying it would be to use it in different contexts (List, Maybe, Either). As soon as we are sufficiently frustrated we’ll invoke our friendly Monad and see how much easier our life becomes.

About Speaker

Luca Belli is a Senior Software Engineer at Twitter Cortex, the centralized deep learning hub within the company. Previously he was a Senior Scientist at Conversant Media where he helped bootstrapping image classification using deep learning. His first job was at Wolfram Alpha in Boston. He got his Ph.D. in Mathematics from Tor Vergata University in Rome.

Pants and Monorepos

at 15:55

Large or quickly growing projects that consist of many interdependent sub-projects with complex dependencies on third-party libraries can be difficult to handle with standard language build tools. Add on to that code generators and the use of multiple languages and suddenly a lot of your coding life is spent figuring out the right commands to run for the right language, and waiting for all of your code to build. This is where Pants can help! Pants is an open source build tool developed and used by Twitter, Square, Foursquare, Medium, and others. This talk will begin with a brief overview of what Pants is and how it can help, and then discuss new features we have been adding to make the tool faster. In particular, I will discuss the work we have done to restrict what is going on the JVM compile classpaths to make building Scala and Java projects faster, and the work we are doing to implement a remotely executing build system.

About Speaker

I’m a member of the Build team at Twitter, meaning it is my full-time job to contribute to Pants. I’ve been working at Twitter for 1.5 years, the first 11 months I spent on a team building integration test frameworks, and then switched to the Build team after realizing how interesting it was. My favorite part is learning how things work under the surface! I broke into computer science in my early twenties after earning a degree in psychology. Chocolate makes the world a better place.

Declarative Control Flow with fs2 Stream

at 16:30

fs2 is a purely functional streaming library, with support for concurrent and nondeterministic merging of arbitrary streams. Concurrency support means that we can use Stream not only to process data in constant memory, but also as a very general abstraction for control flow: whilst IO gives us an excellent model for a single effectful action, assembling behaviour with it often has a very imperative flavour (pure, but still imperative). This talk will introduce fs2 combinators by example, and will hopefully show how we can model control flow in a declarative, high level, composable fashion. In particular, we will focus on concurrent combinators.

About Speaker

I'm a Principal Software Engineer at Ovo Energy in London, specialised in distributed systems and purely functional programming. I'm also an Open Source author and speaker as SystemFw: I'm one of the maintainers of fs2, cats-effect and http4s, and a contributor to cats, shapeless and several other libraries in the Scala FP ecosystem. Passionate about learning and teaching.

Scalafix @ Twitter scale

at 17:05

Scalafix is a fairly popular OSS tool that is useful for performing syntactic and semantic rewrites of Scala code. At Twitter we use it for migration to new library interfaces and maintenance of code health by removal of deprecated code. In this talk we walk through examples of simple and complex Scalafix custom rule specifications for rewrites. We describe the core infrastructure we have set up to support rewrites across our entire monorepo, several orders of magnitude faster than if we were to apply them manually. A simple demo will be included to provide a glimpse of our developer workflow and the user experience with our code base. We envision leveraging this tool for more purposes such as improving performance, upgrading compiler revisions, and assisting developers to automatically recognize and prevent commits of disallowed code patterns.

About Speaker

Uma is a Staff Software Engineer in the Advanced Scala Tools team at Twitter. She brings with her multiple decades of experience and expertise in the area of compilers, code generation and related hardware/software co-design. Prior to joining Twitter she worked at Intel and Hewlett Packard. She has a Bachelor’s degree in Electrical Engineering and a Master’s in CS. She holds several patents and technical publications in her field of expertise.