CSE341: Programming Languages CSE341: Programming Languages CSE 401 – Compilers Lecture 1 CourseLecture Mechanics 1 CourseML Variable Mechanics Bindings Lecture 2: Languages, Automata, Regular ML Variable Bindings Expressions & Scanners Dan Grossman Michael Ringenburg Fall 2011 Dan Grossman Winter 2013 Fall 2011 Winter 2013 UW CSE 401 (Michael Ringenburg) ADministrave Notes CSE341: ProgrammingCSE341: Programming Languages Languages • Reading – Cooper & TorczonLecture: Chapter 1, anD SecEons 2.1-2.4 1 Lecture 1 – Try to finish by the enD of the week – it’ll be helpful Course MechanicsCourse Mechanics for the first homework. ML Variable BindingsML Variable Bindings • First homework Dan Grossman Dan Grossman Fall 2011 – ShoulD be out on FriDay (I’ll post on course website Fall 2011 anD senD an email). – Will be Due a week from FriDay (January 18). – Note: You have 4 late Days for the en)re quarter. Use them wisely (see syllabus for Details). Winter 2013 UW CSE 401 (Michael Ringenburg) B-2 ReminDers CSE341: ProgrammingCSE341: Programming Languages Languages • Please vote for office hours by enD-of-Day ThursDay (see link on course home page). Lecture 1 Lecture 1 Course Mechanics – Select whichever slots you think you coulD Course Mechanics ML Variable Bindings reasonably aenD. ML Variable Bindings Dan Grossman – We will use this to help DeciDe office hours for the Dan Grossman Fall 2011 TAs anD the instructor. Fall 2011 • Please pick your project partner, anD senD mail to cse401-staff[at]cs. – First piece of the project will be releaseD (early) next week, so you shoulD pick partners this week. Winter 2013 UW CSE 401 (Michael Ringenburg) B-3 Snow CSE341: ProgrammingCSE341: Programming Languages Languages • It’s the Eme of year where the “S”-worD starts to show up occasionally in weather forecasts. Lecture 1 Lecture 1 Course MechanicsCourse Mechanics • The scheDule for the quarter is Eght, so if we ML Variable Bindings Do have a snow Day at some point, we may ML Variable Bindings Dan Grossman have to rush through some of the material. Dan Grossman Fall 2011 – If this happens, take advantage of the extra Eme Fall 2011 on the snow Day to pay extra aenEon to the readings – with less Eme to cover the material in class, the readings become corresponDingly more important. Winter 2013 UW CSE 401 (Michael Ringenburg) B-4 AgenDa CSE341: ProgrammingCSE341: Programming Languages Languages • Finish course intro (history) • IntroDuce Scanning (part 1 of your project) Lecture 1 Lecture 1 Course MechanicsCourse Mechanics – Quick review of basic concepts of formal ML Variable BindingsML Variable Bindings grammars Dan Grossman – Regular expressions Dan Grossman Fall 2011 – Lexical specificaon of programming languages Fall 2011 – Using finite automata to recognize regular expressions – Scanners anD Tokens Winter 2013 UW CSE 401 (Michael Ringenburg) B-5 Some History • Early computers – hanD coDeD assembly language CSE341: ProgrammingCSE341: Programming Languages Languages (punchcarDs!) – HarD to write anything complex – but earliest computers Lecture 1 Lecture 1 coulDn’t execute any thing that Course MechanicsCoursecomlex Mechanics. • 1952: Grace Hopper writes first compiler (for A-0), anD ML Variable Bindings coins the term “compiler”. ML Variable Bindings – EssenEally a collecEon of mathemacal subrouEnes that coulD Dan Grossman be calleD. The compiler woulD take a series of calls anD convert Dan Grossman them into an executable. Fall 2011 – Successors: A-1, A-2 (first “open source” sojware), anD later … Fall 2011 B-0! • 1957: IBM writes first real “high-level” language compiler, for FORTRAN. (FORTRAN is high level compareD to assembly.) – CompeEEve with hanD-opEmizeD coDe. – RequireD 18 person-years (hopefully your projects won’t take this long!) Winter 2013 UW CSE 401 (Michael Ringenburg) B-6 Some History • 1962: First CSE341:bootstrapped ProgrammingCSE341: compiler (for LISP) Programming Languages Languages – A compiler that was compileD by itself, rather than wri\en in assembly (or another language). Lecture 1 Lecture 1 – Requires iniEally creang a very simple compiler in assembly or Course Mechanics another language, anD then using that to compile the iniEal Course Mechanics bootstrappeD compiler. ML Variable BindingsML Variable Bindings • IniEal compiler may contain just a subset of the language. As this compiler is refineD to compile more of the language, the compiler itself Dan Grossman can begin to use more of the language. Dan Grossman – Much more efficient that wriEng in assembly (like the first Fall 2011 compilers). Fall 2011 – Great way to test a compiler. • Rest of 1960’s, into 1970’s – Work on formalizing scanning anD parsing (theory anD pracEce). – Automac parser anD scanner generators • Lex (lexical analyzer) anD Yacc (Yet Another Compiler Compiler) • JFlex anD Cup are Direct DescenDants of these C-baseD tools. Winter 2013 UW CSE 401 (Michael Ringenburg) B-7 Some History CSE341: ProgrammingCSE341: Programming Languages Languages • Late 1970’s, 1980’s – New languages (funcEonal; object-orienteD) Lecture 1 Lecture 1 – New architectures (RISC, parallel machines, caches, …) Course MechanicsCourse Mechanics – Back-enD improvements: OpEmizaon, Register ML Variable BindingsML Variable Bindings Allocaon, Automac parallelizaon Dan Grossman • 1990s Dan Grossman Fall 2011 – ImproveD techniques for compiling object orienteD coDe Fall 2011 • Efficiency in the presence of Dynamic Dispatch anD small methoDs – Just-in-Eme compilers (JITs) – Compiler technology to effecEvely use new harDware (RISC, parallel machines, complex memory hierarchies) Winter 2013 UW CSE 401 (Michael Ringenburg) B-8 Some History CSE341: ProgrammingCSE341: Programming Languages Languages • Last Decade – Compilaon techniques in many new places Lecture 1 Lecture 1 Course Mechanics • E.g., parsing, semanEc analysis, source-to-source Course Mechanics translaon useD for sojware analysis, verificaon, ML Variable BindingsML Variable Bindings security Dan Grossman Dan Grossman – PhaseD compilaon – blurring the lines between Fall 2011 “compile Eme” anD “runEme” Fall 2011 • Programs can generate anD compile specializeD versions of rouEnes “on the fly”. • Can use machine learning to control opEmizaons – MulEcore: parallelism everywhere! Winter 2013 UW CSE 401 (Michael Ringenburg) B-9 Any quesEons? CSE341: ProgrammingCSE341: Programming Languages Languages • Don’t hesitate to ask – I’m teaching this course because I enjoy talking about Lecture 1 Lecture 1 Course Mechanics compilers. Course Mechanics ML Variable BindingsML Variable Bindings • If you have a quesEon, it’s likely other people Dan Grossman Dan Grossman Do as well, but they are too shy to ask. So Fall 2011 Fall 2011 you’ll be Doing them a favor too. Winter 2013 UW CSE 401 (Michael Ringenburg) B-10 AgenDa CSE341: ProgrammingCSE341: Programming Languages Languages • Finish course intro • Introduce Scanning (part 1 of your project) Lecture 1 Lecture 1 Course MechanicsCourse Mechanics – Quick review of basic concepts of formal ML Variable BindingsML Variable Bindings grammars Dan Grossman – Regular expressions Dan Grossman Fall 2011 – Lexical specificaon of programming languages Fall 2011 – Using finite automata to recognize regular expressions – Scanners anD Tokens Winter 2013 UW CSE 401 (Michael Ringenburg) B-11 Programming Language Specificaons CSE341: ProgrammingCSE341: Programming Languages Languages • Since the 1960s, the syntax of every significant programming language has been specifieD by a Lecture 1 Lecture 1 formal grammar Course MechanicsCourse Mechanics – If you ever have the “pleasure” of reading a language ML Variable BindingsML Variable Bindings specificaon Document, you’ll see that each secEon Dan Grossman typically consists of a formal grammar for some piece Dan Grossman Fall 2011 of the syntax, followeD by notes Describing the Fall 2011 semanEcs. – First Done in 1959 with BNF (Backus-Naur Form) grammar useD to specify ALGOL 60 syntax – BorroweD from the linguisEcs community (Chomsky) Winter 2013 UW CSE 401 (Michael Ringenburg) B-12 Review of Formal Languages anD Automata Theory CSE341: ProgrammingCSE341: Programming Languages Languages Oink!!! • Starring Mr. Pig Lecture 1 Lecture 1 • Alphabet: a finite set of symbols anD characters Course MechanicsCourse Mechanics – E.g., {‘iML’, ‘k’, ‘n’, ‘o’, ‘!’, ‘ ’} Variable BindingsML Variable Bindings • String: a finite, possibly empty sequence of Dan Grossman Dan Grossman symbols from an alphabet Fall 2011 Fall 2011 – E.g., “oink” • Language: a set of strings (possibly empty or infinite) – E.g., {“oink!”, “oink oink!”, “oink oink oink!”, …} Winter 2013 UW CSE 401 (Michael Ringenburg) B-13 Finite Specificaons of Possibly Infinite Languages • Automaton – a recognizer; a machine that accepts all CSE341: ProgrammingCSE341: Programming Languages Languages strings in a language (anD rejects all other strings) Lecture 1 Lecture 1 – E.g., a pig Detector: accepts all sequences of oinks, rejects Course Mechanics “moo”s or “Coursebaa”s Mechanics ML Variable Bindings • Grammar – a generator that proDuceD all strings in the ML Variable Bindings language (anD nothing else) Dan Grossman Dan Grossman – Unfortunately, we can’t use a pig as our grammar – no pig Fall 2011 (that I’ve met) can generate infinite “oink” sequences. Fall 2011 – Instead we use formal (aka mathemacal) grammars. • A parEcular language may be specifieD by many Different grammars anD automata – But, a grammar or automaton specifies only one language Winter 2013 UW CSE 401 (Michael Ringenburg) B-14 Language (Chomsky) hierarchy: quick reminder •