剑桥大学 writing

nullÀ la recherche des réseaux perdus. 9 décembre 2007. Le tutoriel sera donné par Jon Crowcroft,À la recherche des réseaux perdus. 9 décembre 2007. Le tutoriel sera donné par Jon Crowcroft,Jon Crowcroft, University of Cambridge Currently CNRS/LIP6/UPMC And Thomson Labs, Paris. Jon.Crowcroft@cl.cam.ac.uk http://www.cl.cam.ac.uk/~jac22This talk is in EnglishThis talk is in EnglishUnfortunately (desole) my French is insufficient1.4 Billion Indians and Chinese have chosen English (luckily we don’t have to learn mandarin or Urdu!) Even British have to learn American English for science writing We all use Greek for maths anyhow :-)There are 6 parts to this talkThere are 6 parts to this talkCold Topics in Networks Reading a paper Writing a paper Giving a talk Writing a proposal Hot Topics in Networks Acknowledgements to Keshav Simon Peyton Jones Brad Karp and HT Kung for materials used with permission. 1. Cold Topics in Networks1. Cold Topics in NetworksJon Crowcroft, CambridgeHot v. ColdHot v. ColdResearch goes in cycles - possibly Carnot CyclesTopics become hot Initially, even, controversial (active nets, social nets, etc) Lots of people flock to the topic Paradoxically, higher density of researchers reduces temperature Topic goes cold. Hot TopicsHot TopicsCan be detected by brainstorming Socialising and off-the-wall thinking is good Invert a traditional approach Stretch one dimension to an extreme Bisociation/lateral/interdisciplinary Tennenhouse at DARPA (Active Nets) and as head of Intel Research deliberately indulged disruptive ideasSome measures of cold topicsSome measures of cold topicsNumber of low cited papers in low impact conferences Fractional Performance delta in systems papers Massive uptake of automatic tools for research (NS2, Planetlab, etc)Some examples of cold topicsSome examples of cold topics[DHT and Structured P2P] Even bad guys like the Storm Botnet use them [Internet Coordinate Systems] now secured too! [Faster packet classification] If you aren’t working with cisco, juniper or huawei? [BGP] The Border Gateway Protocol We even have a meta-replacement. Now is the time to deploy. More cold topicsMore cold topicsDDoS Define the problem - DoS on a best effort doesn’t mean much - see Newarch. Spam Is not largely a technical problem (see social nets and closed user groups) see ddos Overlays Were made up as a tool for research, not a research goal!Even more cold topicsEven more cold topicsTINA The Intelligent Network Architecture = Knowledge Plane = network management TCP+AQM Mostly wrong Multicast 20 years without deployment? Newarch Not even wrong (see String Theory)Yet more cold topicsYet more cold topicsSelf similarity Surely there is a horizon effect MANETs 5000 protocols cannto be good Self Organising WSNs Unexpected behaviour may not be a plus Small World Networks Ask epidemiologistsNow you know what I don’t like…Now you know what I don’t like…For now :-)Your PhD topic will have been hot in Year 1. By year 3,4,5 this is unlikely to be still true - consider journals rather than conferences or workshops for later work:-).2. How to Read a Paper2. How to Read a PaperJon Crowcroft, Cambridge Based on CCR Article by Keshav (Waterloo)Stand on the Shoulders of GiantsStand on the Shoulders of GiantsAnd do not stand on their toesYou read other papers so that You are learning what papers are like You are current in the field You may be writing survey (literature review) You want to find what to compare with We propose a 3 pass reading approach Pass 1Pass 1Structural overview of paper Read abstract/title/intro Read section headings, ignore bodies Read conclusions Scan references noting ones you knowPass 1 outputPass 1 outputYou can now say Is this a system, theory or simulation paper (category defines methodology) Check system measurement methodology Check expressiveness/fit for purpose of formalism Check simulation assumptions What other papers/projects relate to this? Are the assumptions valid? What are the key novel contributions Is the paper clear? Takes about 5 minutes 95% of reviewers will stop at pass 1 :-( See Section 3 of this (on writing papers)Pass 2Pass 2Check integrity of paper Look at figures/diagrams/exes/definitions Note unfamiliar references Do not check proofs yet Takes around 1 hour You should be able to summarise the paper to someone else now If it is unclear, you may need to pasuse overnight Pass 3Pass 3Virtually re-implement the paper Challenge all assumptions Think adversarially about experiments, proofs, simulation scenarios Takes 4-5 hours You should be able to reconstruct paper completely nowReading batches of papersReading batches of papersE.g. for literature survey excercise pick topic (hot or cold), and search on google scholar or citeseer for 10 top papers Find shared citations and repeated author names - key papers (look at citation count/impact too) Go to venues for these papers and look at other papersSee alsoSee alsoTimothy Roscoe’s Writing reviews for Systems Conferences Writing Technical Articles Henning Schulzrinne’s Now you can review papers…Now you can review papers…For me:)You will read 100 papers to every one you write. 90 of them will be much worse, some will be better. A few the same.3. How to write a great research paper3. How to write a great research paperSimon Peyton Jones Microsoft Research, Cambridge Writing papers is a skillWriting papers is a skillMany papers are badly written Good writing is a skill you can learn It’s a skill that is worth learning: You will get more brownie points (more papers accepted etc) Your ideas will have more impact You will have better ideasIncreasing importanceWriting papers: model 1Writing papers: model 1IdeaDo researchWrite paperWriting papers: model 2Writing papers: model 2Forces us to be clear, focused Crystallises what we don’t understand Opens the way to dialogue with others: reality check, critique, and collaborationIdeaDo researchWrite paperIdeaWrite paperDo researchDo not be intimidatedDo not be intimidatedWrite a paper, and give a talk, about any idea, no matter how weedy and insignificant it may seem to youFallacy You need to have a fantastic idea before you can write a paper. (Everyone else seems to.)Do not be intimidatedDo not be intimidatedWrite a paper, and give a talk, about any idea, no matter how insignificant it may seem to youWriting the paper is how you develop the idea in the first place It usually turns out to be more interesting and challenging that it seemed at first The purpose of your paperThe purpose of your paperWhy bother?Why bother?Good papers and talks are a fundamental part of research excellenceFallacy we write papers and give talks mainly to impress others, gain recognition, and get promotedPapers communicate ideasPapers communicate ideasYour goal: to infect the mind of your reader with your idea, like a virus Papers are far more durable than programs (think Mozart)The greatest ideas are (literally) worthless if you keep them to yourselfThe IdeaThe IdeaFigure out what your idea is Make certain that the reader is in no doubt what the idea is. Be 100% explicit: “The main idea of this paper is....” “In this section we present the main contributions of the paper.” Many papers contain good ideas, but do not distil what they are.Idea A re-usable insight, useful to the readerOne pingOne pingYour paper should have just one “ping”: one clear, sharp idea Read your paper again: can you hear the “ping”? You may not know exactly what the ping is when you start writing; but you must know when you finish If you have lots of ideas, write lots of papersThanks to Joe Touch for “one ping”The purpose of your paper is not...The purpose of your paper is not...To describe the WizWoz systemYour reader does not have a WizWoz She is primarily interested in re-usable brain-stuff, not executable artefactsExamples of WizWozExamples of WizWozCrash Proof OS for Mobile Phones (singularity in F# on an iPhone) Go Faster VM (Xen) NimrodYour narrative flowYour narrative flowHere is a problem It’s an interesting problem It’s an unsolved problem Here is my idea My idea works (details, data) Here’s how my idea compares to other people’s approachesI wish I knew how to solve that!I see how that works. Ingenious!Structure (conference paper)Structure (conference paper)Title (1000 readers) Abstract (4 sentences, 100 readers) Introduction (1 page, 100 readers) The problem (1 page, 10 readers) My idea (2 pages, 10 readers) The details (5 pages, 3 readers) Related work (1-2 pages, 10 readers) Conclusions and further work (0.5 pages) See section 2 (on reading!)The abstractThe abstractI usually write the abstract last Used by program committee members to decide which papers to read Four sentences [Kent Beck] State the problem Say why it’s an interesting problem Say what your solution achieves Say what follows from your solutionExampleExampleMany papers are badly written and hard to understand This is a pity, because their good ideas may go unappreciated Following simple guidelines can dramatically improve the quality of your papers Your work will be used more, and the feedback you get from others will in turn improve your researchStructureStructureAbstract (4 sentences) Introduction (1 page) The problem (1 page) My idea (2 pages) The details (5 pages) Related work (1-2 pages) Conclusions and further work (0.5 pages)The introduction (1 page)The introduction (1 page)Describe the problem State your contributions ...and that is all ONE PAGE!Describe the problemDescribe the problemUse an example to introduce the probleme.g. of systems probleme.g. of systems problemMobile Phones crash a lot Wireless media is vulnerable Bad software on mobile phone can hurt user (cost money, time, pain) Bad software on radio can hurt all users We have a lot better tools to write safer software and have done so on desktops and servers Can they work on small devices with limited resources, and if so, how well?State your contributionsState your contributionsWrite the list of contributions first The list of contributions drives the entire paper: the paper substantiates the claims you have made Reader thinks “gosh, if they can really deliver this, that’s be exciting; I’d better read on”State your contributionsState your contributionsBulleted list of contributionsDo not leave the reader to guess what your contributions are!E.g. of systems contributionsE.g. of systems contributionsWe encapsulate all the modules of software on a cell phone in F# behavioural description wrappers, and Run a model checker on them (e.g. isobel) And then try various well known attacks that fail on desk top but succeed on windows mobile and symbian phones We then show our software is also smaller and faster….Contributions should be refutableContributions should be refutableNo “rest of this paper is...”No “rest of this paper is...”Not: Instead, use forward references from the narrative in the introduction. The introduction (including the contributions) should survey the whole paper, and therefore forward reference every important part.“The rest of this paper is structured as follows. Section 2 introduces the problem. Section 3 ... Finally, Section 8 concludes”.StructureStructureAbstract (4 sentences) Introduction (1 page) Related work The problem (1 page) My idea (2 pages) The details (5 pages) Related work (1-2 pages) Conclusions and further work (0.5 pages)No related work yet!No related work yet!Related workYour readerYour ideaWe adopt the notion of transaction from Brown [1], as modified for distributed systems by White [2], using the four-phase interpolation algorithm of Green [3]. Our work differs from White in our advanced revocation protocol, which deals with the case of priority inversion as described by Yellow [4].No related work yetNo related work yetProblem 1: the reader knows nothing about the problem yet; so your (carefully trimmed) description of various technical tradeoffs is absolutely incomprehensible Problem 2: describing alternative approaches gets between the reader and your ideaI feel tiredI feel stupidRelated work and surveyRelated work and surveyObviously, if your paper is a survey… …then it is all related work Gap analysis is sometimes useful in an introduction, but it is not quite the same as related work Taxonomies…are quite handy in that caseStructureStructureAbstract (4 sentences) Introduction (1 page) The problem (1 page) My idea (2 pages) The details (5 pages) Related work (1-2 pages) Conclusions and further work (0.5 pages)Presenting the ideaPresenting the idea3. The idea Consider a bifircuated semi-lattice D, over a hyper-modulated signature S. Suppose pi is an element of D. Then we know for every such pi there is an epi-modulus j, such that pj < pi.Sounds impressive...but Sends readers to sleep In a paper you MUST provide the details, but FIRST convey the ideaPresenting the ideaPresenting the ideaExplain it as if you were speaking to someone using a whiteboard Conveying the intuition is primary, not secondary Once your reader has the intuition, she can follow the details (but not vice versa) Even if she skips the details, she still takes away something valuablePutting the reader firstPutting the reader firstDo not recapitulate your personal journey of discovery. This route may be soaked with your blood, but that is not interesting to the reader. Instead, choose the most direct route to the idea.The payload of your paperThe payload of your paperIntroduce the problem, and your idea, using EXAMPLES and only then present the general caseUsing examplesUsing examplesExample right awayThe Simon PJ question: is there any typewriter font?The details: evidence The details: evidence Your introduction makes claims The body of the paper provides evidence to support each claim Check each claim in the introduction, identify the evidence, and forward-reference it from the claim Evidence can be: analysis and comparison, theorems, measurements, case studiesIn my safeOS for handset e.g.In my safeOS for handset e.g.One would give code fragments of unsafe code And examples of threats… …and examples of safe code And some performance results… Before launching in to the description of the small fast efficient compile and runtime checks possible in new system…StructureStructureAbstract (4 sentences) Introduction (1 page) The problem (1 page) My idea (2 pages) The details (5 pages) Related work (1-2 pages) Conclusions and further work (0.5 pages)Related workRelated workFallacy To make my work look good, I have to make other people’s work look badThe truth: credit is not like moneyThe truth: credit is not like moneyGiving credit to others does not diminish the credit you get from your paperWarmly acknowledge people who have helped you Be generous to the competition. “In his inspiring paper [Foo98] Foogle shows.... We develop his foundation in the following ways...” Acknowledge weaknesses in your approachCredit is not like moneyCredit is not like moneyFailing to give credit to others can kill your paperIf you imply that an idea is yours, and the referee knows it is not, then either You don’t know that it’s an old idea (bad) You do know, but are pretending it’s yours (very bad)StructureStructureAbstract (4 sentences) Introduction (1 page) The problem (1 page) My idea (2 pages) The details (5 pages) Related work (1-2 pages) Conclusions and further work (0.5 pages)Conclusions and further workConclusions and further workBe brief.The process of writingThe process of writingThe processThe processStart early. Very early. Hastily-written papers get rejected. Papers are like wine: they need time to mature Collaborate Use CVS to support collaborationGetting helpGetting helpExperts are good Non-experts are also very good Each reader can only read your paper for the first time once! So use them carefully Explain carefully what you want (“I got lost here” is much more important than “Jarva is mis-spelt”.)Get your paper read by as many friendly guinea pigs as possibleGetting expert helpGetting expert helpA good plan: when you think you are done, send the draft to the competition saying “could you help me ensure that I describe your work fairly?”. Often they will respond with helpful critique (they are interested in the area) They are likely to be your referees anyway, so getting their comments or criticism up front is Jolly Good.Listening to your reviewersListening to your reviewersThis is really, really, really hard But it’s really, really, really, really, really, really, really, really, really, really importantTreat every review like gold dust Be (truly) grateful for criticism as well as praiseListening to your reviewersListening to your reviewersRead every criticism as a positive suggestion for something you could explain more clearly DO NOT respond “you stupid person, I meant X”. Fix the paper so that X is apparent even to the stupidest reader. Thank them warmly. They have given up their time for you.Language and styleLanguage and styleBasic stuffBasic stuffSubmit by the deadline Keep to the length restrictions Do not narrow the margins Do not use 6pt font On occasion, supply supporting evidence (e.g. experimental data, or a written-out proof) in an appendix Always use a spell checkerVisual structureVisual structureGive strong visual structure to your paper using sections and sub-sections bullets italics laid-out code Find out how to draw pictures, and use themVisual structureVisual structureUse the active voiceUse the active voiceThe passive voice is “respectable” but it DEADENS your paper. Avoid it at all costs.“We” = you and the reader“We” = the authors“You” = the readerUse simple, direct languageUse simple, direct languageSummarySummaryIf you remember nothing else: Identify your key idea Make your contributions explicit Use examples A good starting point: “Advice on Research and Writing”http://www-2.cs.cmu.edu/afs/cs.cmu.edu/user/ mleone/web/how-to.html4. How to give a good research talk4. How to give a good research talkSimon Peyton Jones Microsoft Research, Cambridge 1993 paper joint with John Hughes (Chalmers), John Launchbury (Oregon Graduate Institute)Research is communicationResearch is communicationThe greatest ideas are worthless if you keep them to yourselfYour papers and talks Crystalise your ideas Communicate them to others Get feedback Build relationships (And garner research brownie points)Do it! Do it! Do it!Do it! Do it! Do it!Good papers and talks are a fundamental part of research excellenceInvest time Learn skills PracticeWrite a paper, and give a talk, about any idea, no matter how weedy and insignificant it may seem to youGiving a good talkGiving a good talkThis presentation is about how to give a good research talk What your talk is for What to put in it (and what not to) How to present itWhat your talk is forWhat your talk is forYour paper = The beef Your talk = The beef advertismentDo not confuse the two, even if you are vegetarianThe purpose of your talk…The purpose of your talk…..is not: To impress your audience with your brainpower To tell them all you know about your topic To present all the technical detailsThe purpose of your talk…The purpose of your talk…..but is: To give your audience an intuitive feel for your idea To make them foam at the mouth with eagerness to read your paper To engage, excite, provoke themYour audience…Your audience…The audience you would like Have read all your earlier papers Thoroughly understand all the relevant theory of cartesian closed endomorphic bifunctors Are all agog to hear about the latest developments in your work Are fresh, alert, and ready for actionYour actual audience…Your actual audience…The audience you get Have never heard of you Have heard of bifunctors, but wish they hadn’t Have just had lunch or been skiing for 5 hours and are ready for a dozeYour mission is to WAKE THEM UP And make them glad they didWhat to put inWhat to put inWhat to put inWhat