Download as PDF, PPTX
Uploaded byIndicThreads
Using Scala for building DSLs
The document discusses the concept of Domain Specific Languages (DSLs) and their application in various domains, including cloud computing, using Scala as a foundation. It explains the differences between internal and external DSLs, Scala's features such as higher-order functions and pattern matching, and the advantages of using Scala for creating DSLs that are concise and expressive. The content culminates in demonstrating how to build cloud computing DSLs using Scala's syntax and constructs.
More Related Content
Similar to Using Scala for building DSLs
Using Scala for building DSLs
- 1. Using Scala forbuilding DSL’s Abhijit Sharma Innovation Lab, BMC Software 1
- 2. What is aDSL? • Domain Specific Language • Appropriate abstraction level for domain - uses precise concepts and semantics of domain • Concise and expressive for a specific domain – not general purpose • Domain experts can communicate, critique better with programmers and amongst themselves • Math – Mathematica, UI – HTML, CSS, Database - SQL 2
- 3. DSL’s at large •Build tool Ant is an XML based DSL • Task to build a jar with a dependency on task compile • Web App framework Ruby on Rails • ActiveRecord to model domain objects and persist them – • Domain constraint implementations do not clutter API – uniqueness, cardinality, null check 3
- 4. Cloud Computing DSL •Audience – Non tech savvy cloud end users • DSL - English language sentence for requesting a machine with technical and price specifications • Domain Concept – Machine • Technical specifications - cpu, os • Pricing specifications - spot Machine price, pricing strategy – •cpu:Cpu •os:String default or define inline •spotPrice:Int Cpu •arch: String •cpus: Int 4
- 5. Cloud Computing DSLin Java • Builder pattern - Method Chaining Fluent Interface • Issues • Syntax restrictions, verbosity – parenthesis, dot, semi-colons • Non Domain complexity – Builder • No Inline strategy – no higher order functions 5
- 6. DSL Classification • InternalDSL • Embedded in a host language like Ruby, Scala, Groovy – use their features • Bound by host language syntax and semantics • External DSL – standalone developed ground up • Define syntax and semantics as a grammar • Use tools like lexical analyzers, parsers, interpretation, code generators 6
- 7. Internal DSL Classification •Internal DSL • Generative - Ruby, Groovy Techniques like runtime metaprogramming • Meta Objects – inject new behaviour at runtime • Embedded in host language • Smart API – Method chaining – Java etc. • Syntax tree manipulation – Groovy, Ruby libraries • Type Embedding – Scala – statically typed, type constraints, no invalid operations on types 7
- 8. Scala Language • Scalais a “scalable” language • JVM based – leverage libs, JVM perf, tools, install base etc • Mixed Paradigm – Object Oriented + Functional Programming • Object Oriented Programming - Improves on Java OOP (Traits, no statics, advanced types) 8
- 9. Scala Language • FunctionalProgramming - Functions • No side effects and immutable variables • “First-class” citizens - Can be assigned, passed around, returned • Higher order functions promote composition using other more primitive functions • Lots of powerful features – discussed later • Statically Typed – Type Safe DSL 9
- 10. Scala – Readablestyle • Type inference minimizes the need for explicit type information – still Type safe • Elegant, succinct syntax unlike verbose Java • Optional dots, semi-colons and parentheses • Operators like methods • Syntactic sugar method takes one/zero argument, drop period and parentheses 10
- 11. Scala - Implicits •Implicits • Conversions – return a wrapped original type e.g. Integer • Implicit argument to function – don’t need to pass – Concise syntax • Generative expression - Implicit conversion converts the 1, an Int, o a RichInt which defines a ‘to’ method • Lexically scoped – Unlike other languages like Groovy where such modifications are global in scope 11
- 12. Scala - Higherorder functions • Functions which take other functions as parameters or return them as results are called higher-order functions • Flexible mechanism for composition • Currying 12
- 13. Scala – FunctionalCombinators • Calculate the total price of all Linux machines – uses several combinators – filter, map, foldLeft – all take other functions as predicates 13
- 14. Scala – CloudComputing DSL - Implicits • Consider excerpt - 8 cpus “64bit” – Using Implicit conversion we get the object representing the CPU - Cpu(8, 64bit) 14
- 15. Scala – CloudComputing DSL – E2E DSL - new Machine having (8 cpus "64bit") with_os “Linux“ • Implicit Conversion • Method Chaining – Builder pattern – without the cruft • Syntactic sugar no parenthesis, dot, brackets 15
- 16. Scala – CloudComputing DSL – Functions • Using Higher Order Functions – Flexible pricing • Spot Price Threshold - Inline strategy 16
- 17. Scala - PatternMatching • Pattern Matching – Switch Case on Steroids • Cases can include value, types, wild-cards, sequences, tuples, deep inspection of objects 17
- 18. Scala - PatternMatching &Case Classes • Case Classes – simplified construction and can be used in pattern matching • Pattern matching on Case Classes • Deep pattern matching on object contents • Make good succinct powerful DSL 18
- 19. Scala - PatternMatching – Visitor Pattern • Pattern match and case classes – extensible visitor • Different operations on tree • Expression Evaluation • Prefix Notation • Very expressive, flexible and concise code 19
- 20. Scala – ForComprehensions • Loop through Iterable sequences and comprehend/compute something • E.g. Filter 32, 64 bit architectures 20
- 21. Scala – ForComprehensions + Option • Wrap vars & function returns as Option – Null Checks, resilient programming • Option sub classes : None and Some • Options with for comprehensions, automatic removal of None elements from comprehensions 21
- 22. Scala – ForComprehensions + Option • Validate and audit machines • Using Options with for comprehensions eliminate the need for most “null/empty” checks. • Succinct, safe DSL with uncluttered API 22
- 23. Scala - Traits • Traits are collections of fields and behaviors that you can extend or mixin to your classes. • Modularize these concerns, yet enable the fine-grained “mixing” of their behaviors with other concerns at build or run time – Callbacks & Ordered • Traits can be mixed-in at class level or at instance creation • AOP Pervasive concerns - Logging, Ordering, Callback Handling 23
- 24. External DSL inScala DSL - having (8 cpus "64bit") with_os "Linux" at_spot_price 30 • Parser Combinator library – available as a library on host language – Scala • External Parser Generators like Javacc – use tools to generate code for tokenizing, parsing • Parser Combinator Specification is like a BNF grammar 24
- 25. External DSL inScala • Each function is a parser - works on a portion of the input, parses it and may optionally pass on the remaining part to the next parser in the chain via the combinator • Several combinators provided by the library like ‘~’ the sequencing combinator composes two parsers sequentially. • Optional function application combinator (^^) can work, applying the function on the result of the sequencing combinator. 25
- 26. Thanks abhijit.sharma@gmail.com Twitter : sharmaabhijit Blog : abhijitsharma.blogspot.com 26