Writing for Computer Science Writing up Overview n

Writing for Computer Science: Writing up

Overview n Getting Started q q n Publishing q q q n Writing Tips Paper Scope External Sorting Paper Organization First Draft Submission Thesis Parting Words 2

Getting Started n Research is exciting! q q q Solve real problems Enrich world’s knowledge Slow, steady progress n q Enjoy the successes Even discovering something that does not work is progress. 3

Getting Started n n n Read background articles Brainstorm with advisor Scribble down ideas: q q n Spiral notebook Messy & creative Research ≈ Mouse in a Maze q q Dead-ends Cheese at the end 4

Writing Tips n n Start writing early to organize your thoughts Know your audience q q n What is common knowledge? What needs more details? Research = Innovation Evaluation 5

Paper Scope n What is the problem? q q n Why is it interesting? How would it help others? Organization q q Background results first OR New results first Audience determines amount of background material 6

External Sorting(p. 139) Research scenario: Audience Question coming! n Database external sorting q n While (more data) { Load data from HDD into memory Sort data in memory Save sorted result to temporary HDD file } Merge temporary files Audience: Is compressing the data a good idea? 7

External Sorting(p. 139) n Compression Disadvantage: q n Compression Advantage: q q n Fewer reads/writes from the HDD Less network traffic If all data fits in memory q n More processing time Avoid compression Otherwise q Use compression save HDD reads/writes 8

Paper Organization n Educate readers n Logic Sequences: q Chain n q State the Problem Background Improvement Specificity n n Outline Fill in details (e. g. , external sorting) – HDD/network bottleneck – Previous sorting approaches… – Compressed merge sort (e. g. , compiler) – Front-end & Back-end – Parser, Syntax-Analyzer, … 9

Paper Organization n Logic Sequences: q By Example n n q Start with familiar setting – Merge sort Apply to new problem – External merge sort By Complexity n n (e. g. , external sorting) Simple cases Complex cases (e. g. , Recursive proof) – Base cases – Recursive cases 10

Paper Organization n Common Sections: q q Title & Author Abstract n n q 50 to 200 words Summarize paper Introduction n Extended abstract Motivation & results Avoid jargon and notation 11

Paper Organization n Common Sections: q Body n n q Literature Review n n q Background & terminology Detailed proofs & results Standing on the shoulders of giants Most work extends other works -- inheritance Conclusions n n Review results Future work idea 12

Paper Organization n Common Sections: q Bibliography n Bib. Te. X Editor 13

Ly. X Word Processor n Ly. X word processor q q q La. Te. X editor Windows or Linux Pre-compiles formulas & citations n WYSIWYM Uncompiled Formula: Ly. X Source Code: $S(x)=frac{S(x-1)+O(N^2)}{2}$ 14

The First Draft n First Draft q Written while you do research n n q Innovation Evaluation Organizes your thoughts Finds mistakes Focus on concepts over grammar n n n Problem definition Mathematical content Be messy & creative 15

The First Draft n First Draft q q Use simple sentences to avoid writer’s block Brainstorm ideas with advisor Outline ideas When you get stuck: n Try different angles n Talk with advisor n Go <do your favorite activity> (I go jogging ) 16

From Draft to Submission n n Anticipate objections Make conservative claims Be thorough with evidence Keep a journal q Ideas, proofs, & hypotheses 17

From Draft to Submission n Co-Authoring Strategies: q Each author writes different sections n q Disadvantages: q Disjoint styles q Inconsistent tables & figures q Repetition & omission Taking Turns + Exclusive locks n n Person a starts writing Person b revises & extends Person a revises & extends Person b revises & extends 18

Prepublication n Technical Reports q q Timestamp “Stake claim” to your ideas 19

Thesis n Thesis Goals: q q Demonstrate original & meaningful research Critique your own results: n Why is this algorithm better? q q Less network congestion? Fewer CPU cycles? Show importance of the topic Show that results are useful to others 20

Writing Checklist n n Identify goals & scope Logically organize the paper Keep a journal of brainstormed ideas Write while researching evaluate ideas 21

Parting words n n Work diligently. Make a conscious decision to be happy in your work. Research is a journey, not a destination. 22

Reference: n Zobel, Justin. Writing for Computer Science, 2 nd ed. , 2004. 23