Author: Andreas Rumpf ## A new approach to metaprogramming ## Talk structure 1. What is Nimrod? 2. Implementation aspects 3."Hello World" 4. Meta programming features 5. Optimizing "Hello rewriting macros 7. Summary of Nimrod's meta programming features ## What is Nimrod? ## What is Nimrod? ## • A statically typed var input: TaintedString ## What is Nimrod? ## • A statically typed var input: TaintedString • systems programming language var a = cast[int](gch.stackBottom) ## What is Nimrod? ## • A statically typed var input: TaintedString • systems programming language var a = cast[int](gch

Thesis 3 1.2 Hypothesis 3 2 Ethics 3 3 The Game Nim 4 3.1 Rules and How to Play It 4 3.2 Nimrod 5 3.3 Impartial Games 5 3.4 The Sprague-Grundy Theorem 6 4 Machine Learning 6 4.1 Reinforcement n objects for a pile of size n. ### 3.2 Nimrod During the “Festival of Britain” in 1951 in the UK, in the science exhibition, there was a computer called Nimrod with which visitors could play a game of perspective. Second. CRC Press/Taylor & Francis Group, 2015. The Ferranti NIMROD Digital Computer. URL: https://www.goodeveca.net/nimrod. ## List of Figures 1 A model of a McCulloch Pitts neuron. . . . .

textual input not on its own line but directly after the prompt, such as 'What is your name: Nimrod'. As the echo() proc always writes a newline character after the last argument has been written necessary. No, should not be necessary for recent Nim versions, maybe that is a legacy from the old Nimrod days. And is the threadvar pragma in var clients {.threadvar.}: seq[AsyncSocket] really needed? Our string data type in the C language, which is basically a pointer to a character (*char). In early Nimrod, we often used the Nim cstring data type as an alias for a string in the C language. In modern Nim

textual input not on its own line but directly after the prompt, such as 'What is your name: Nimrod'. As the echo() proc always writes a newline character after the last argument has been written necessary. No, should not be necessary for recent Nim versions, maybe that is a legacy from the old Nimrod days. And is the threadvar pragma in var clients {.threadvar.}: seq[AsyncSocket] really needed? Our string data type in the C language, which is basically a pointer to a character (*char). In early Nimrod, we often used the Nim cstring data type as an alias for a string in the C language. In modern Nim

textual input not on its own line but directly after the prompt, such as 'What is your name: Nimrod'. As the echo() proc always writes a newline character after the last argument has been written necessary. No, should not be necessary for recent Nim versions, maybe that is a legacy from the old Nimrod days. And is the threadvar pragma in var clients {.threadvar.}: seq[AsyncSocket] really needed? Our string data type in the C language, which is basically a pointer to a character (*char). In early Nimrod, we often used the Nim cstring data type as an alias for a string in the C language. In modern Nim

textual input not on its own line but directly after the prompt, such as 'What is your name: Nimrod'. As the echo() proc always writes a newline character after the last argument has been written necessary. No, should not be necessary for recent Nim versions, maybe that is a legacy from the old Nimrod days. And is the threadvar pragma in var clients {.threadvar.}: seq[AsyncSocket] really needed? Our string data type in the C language, which is basically a pointer to a character (*char). In early Nimrod, we often used the Nim cstring data type as an alias for a string in the C language. In modern Nim

textual input not on its own line but directly after the prompt, such as 'What is your name: Nimrod'. As the echo() proc always writes a newline character after the last argument has been written necessary. No, should not be necessary for recent Nim versions, maybe that is a legacy from the old Nimrod days. And is the threadvar pragma in var clients {.threadvar.}: seq[AsyncSocket] really needed? Our string data type in the C language, which is basically a pointer to a character (*char). In early Nimrod, we often used the Nim cstring data type as an alias for a string in the C language. In modern Nim

textual input not on its own line but directly after the prompt, such as 'What is your name: Nimrod'. As the echo() proc always writes a newline character after the last argument has been written necessary. No, should not be necessary for recent Nim versions, maybe that is a legacy from the old Nimrod days. And is the threadvar pragma in var clients {.threadvar.}: seq[AsyncSocket] really needed? Our string data type in the C language, which is basically a pointer to a character (*char). In early Nimrod, we often used the Nim cstring data type as an alias for a string in the C language. In modern Nim

textual input not on its own line but directly after the prompt, such as 'What is your name: Nimrod'. As the echo() proc always writes a newline character after the last argument has been written necessary. No, should not be necessary for recent Nim versions, maybe that is a legacy from the old Nimrod days. And is the threadvar pragma in var clients {.threadvar.}: seq[AsyncSocket] really needed? Our string data type in the C language, which is basically a pointer to a character (*char). In early Nimrod, we often used the Nim cstring data type as an alias for a string in the C language. In modern Nim

textual input not on its own line but directly after the prompt, such as 'What is your name: Nimrod'. As the echo() proc always writes a newline character after the last argument has been written necessary. No, should not be necessary for recent Nim versions, maybe that is a legacy from the old Nimrod days. And is the threadvar pragma in var clients {.threadvar.}: seq[AsyncSocket] really needed? Our string data type in the C language, which is basically a pointer to a character (*char). In early Nimrod, we often used the Nim cstring data type as an alias for a string in the C language. In modern Nim