Can F# make us better as .NET developers?

Can F# make us better as .NET developers?

• 1.
  CAN F# MAKEUS BETTER AS .NET DEVELOPERS? KEVIN AVIGNON
• 2.
  COMPANIES USING F#IN THEIR TOOLSET
• 3.
  KEY POINTS OFTHE LANGUAGE Conciseness • Less coding noise: No curly braces, semicolons and so on Convenience • Many tasks are simpler in F# such as list processing, comparison & equality and state machines Correctness • Powerful type system backing up the language which prevent common errors such as nulls. Completeness • Although F# is functional, it supports other styles making it easy to communicate with websites, databases and so on. Concurrency • Asynchronous programming is very easy as is parallelism because of the nature of functional programming
• 4.
  TESTIMONIALS
• 5.
  RACHEL REESE –SENIOR SOFTWARE ENGINEER @ JET We’ve compared C# and F# solutions to decide with which to move forward. (F# won) Easier to work it. Can solve issues in 3-4 lines and make the code cleaner. Moved from 30k to 2.5 M users in three months Barely stressed our systems
• 6.
  DOM SYME –PRINCIPAL ENGINEER @ MICROSOFT F#’S POWERFUL TYPE INFERENCE MEANS LESS TYPING, MORE THINKING TYPE-INFERRED CODE IS EASILY REFACTORED IMMEDIATE SCALING TO MASSIVE DATA SETS
• 7.
  PRODUCTIVITY GAINS FROM RUDIMENTARY FEATURE SET
• 8.
  OUT-OF-THE-BOX BEHAVIOR FORTYPES • F# types have built-in immutability • To have mutable values, you need to use the mutable keyword • Most F# types have built-in pretty printing • No need to override ToString() -> Pretty printing is given for free • Most F# types have built-in structural equality • No need for IEquatable to validate equality between types • Most F# types are automatically comparable • No need for IComparable to sort types
• 9.
  LOW OVER-HEAD FORTYPE DEFINITION • Every time you need a structure, you create a type • Not reusing strings and list for instance. • Systems are more type-safe, more self-documenting • Changing those types would cause compile-time errors • Small visual footprint to create a type
• 10.
  POWERFUL PATTERN MATCHING •More flexible than a simple switch/case statements • Match expressions, values, conditions • Assign and extract values at the same time • Identify design issues to the developer • Missing case – New choice because of a change of requirement for instance • An impossible case – Existing choice has been removed • Not handling a case is a compile-time error, not a run-time error.
• 11.
  ADVANCED REWARDS FROM F#
• 12.
  UNITS OF MEASURE •It’s a type • Operations can happen only with types with the same unit (float<m/s>) • Values can be added/subtracted/multiplied/divided • They’re not a part of .NET type system • C# & VB.NET don’t have access to units of measure
• 13.
  TYPE PROVIDERS • Componentthat provides types, properties, and methods • Match a specific interface • Loaded as .NET type at design time • The provider can adapt to change to its data source • Input can come in different form • JSON file, URL point to a service, connection string to a data source • Uniform mechanism for data access • Ex: SQL, CSV, XML, JSON, R & Powershell
• 14.
  WHY SELECTING F#IS A GOOD IDEA Variables should not be allowed to change their type. 1 Objects containing the same values should be equal by default. 2 Comparing objects of different types is a compile-time error. 3 Objects must always be initialized to a valid state. Not doing so is a compile-time error. 4 Once created, objects and collections must be immutable. 5No nulls allowed.6 Missing data or errors must be made explicit in the function signature.
• 15.
  THANK YOU!