Research Project Written Presentation Evaluation Form Final Paper: Fill & Download for Free

I did my M.Sc. in Electrical (Power) Engineering 2018–2020 batch. I will give a detailed experience broadly classified as study experience, living experience, money expenses and suggestions to incoming students.In the beginning, I thought why the person specifically highlighted the experience for an ‘Indian’ student and not just any student or any generic international student. But then, perhaps since the education system, lifestyle and cuisine varies from country to country, the answer is expected such that it covers the transition of someone with an Indian background to that of an international one.Study Experience:My background: I did my Bachelors in Electrical and Electronics Engineering from India. In TU Delft, I was in the program - Electrical (Power) Engineering where I chose compulsory courses from both Smart Grids track as well Power Electronics and Machines track. But my final year project was in Power Electronics and Machines track where the thesis was about design, modelling and optimization of machines.Work load: Unlike India, where the semester system (6 months) is followed, a quarterly system is followed (3 months). Considering time interval for exams, results and some holidays, 1 quarter lasts 2.5 months during which you have to study 3–4 courses. Usually lab courses are part of the main course. So overall, I felt that there were too many courses. For instance, a semester system might have 5 main courses. A quarter system for the same duration now has 6 to 8 courses. Some of the courses I felt were too irrelevant for the program. For instance, a course on measurements was mostly about calculating and eliminating noises in micro electronic circuits which is completely irrelevant to power engineering students. Overall, I felt that 15 credits worth of courses (which contributes to 3 months of study time) were totally not useful for my program.Quality of courses: Many of the courses are of high quality. You will end up with remarkably higher skills and knowledge after you finish these coursesStyle of exam: Exams are of the following types (for the courses I undertook or know from another person). In all exams, the faculty member who taught you evaluates the exam. Which means it is not going for ‘external evaluation’ like state public universities in India. Sometimes, teaching assistants also evaluate the exam. But in case of dispute, the final authority rests with your faculty member.Closed book descriptive exam: This is same as the exams held by most colleges in India.Open book descriptive exam: You can take the book with you. But that is for exams with huge design procedure and formulae.Assignment plus written (paper / digital) exam - Some courses have assignments throughout every week or every two weeks that have 20% to 60% weightage with the remaining weightage given to end exams. In digital exam, you have to type instead of writing. Assignments can be group or individual. I personally felt this was the best type of testing. In courses with this of type of assessment, I was able to study systemically and do well.Assignment plus oral exam - Some courses have assignments throughout every week or every two weeks that have 60% to 90% weightage with the remaining weightage given to oral exams. In oral exams, your faculty will ask you questions directly. Sometimes they give you paper to answer in case of equations. This is exactly same as viva exams for labs in India. Usually, I find oral exams harder because unlike written exams, you have no time to think. Any delay that occurs because you were thinking will adversely affect your performance if the faculty thinks so. In one oral exam I attended, the faculty gave time to think and answer and I was able to do well. In another oral exam, it was like rapid fire where I was expected to immediately write and derive expressions and I ended up creating a bad impression. So it totally depends on who is asking and how fast you are expected to respond.MCQ Exams - This is same as entrance exam type in India.Structure of courses: Some courses are organized extremely well. In a given number of hours, you can study systematically, gain enough understanding and pass with good grades. Also some courses are extremely disorganized where you spend hours without any direction. It depends upon who your course instructor is. Your faculty has the authority to organize course, choose the type of assessment and your final grades. Usually I have noticed that when it comes to organization of courses, courses belonging to a track are organized in a similar manner. I mean suppose there is track - ‘X’, if one course is organized well in X, then most other courses in X are also organized well. Similarly, if one course in another track - ‘Y’ is not well organized, then other courses under Y are also not well organized. So perhaps, there is some sort of a head for each track who is responsible for regulating courses, assessment and grading methodologies.The Master Thesis - In my program, the master thesis consists of a long research project worth 45 credits (out of 120 total credits). Note that 37.5% of your entire CGPA depends upon this.Out of the total 120 credits, first year has 60 credits and second year has the remaining 60 credits. If you get less than 30 credits in first year, you may be deported back to India since you are no longer meeting even half the study requirements. In order to start the master thesis, you must not only finish 60 credits but also these 60 credits should include credits from the mandatory courses of the track you selected.The master thesis comprises of a research project you are expected to work ‘independently’. Basically among a broader topic, you are expected to form some research objectives, find ways to achieve that and arrive at the final outcome. You will get 30 minutes of weekly assistance from one supervisor who is a faculty in your department.You can also do your master thesis at a company. Usually I observed that among my classmates, those who took a company project instead of the university research project ended up getting paid for their work (from the company), completed it faster and also received job offers.Often company projects are at different locations and you might have to relocate to that place. Maybe the money for breaking present rental contract and shifting would break even with the internship like salary. But time and job offers appear as pros in case of company thesis. When it comes to quality of research, the university research projects may have a slightly better quality owing to a higher starting level of difficulty. Once again, all this is from a small sample space of 10–15 students I know.2. Living experienceGetting a place to stay - University arranges accommodation for students depending on a first come first serve basis. If you are late to pay fees / not interesting in university arranging your accommodation, you can search for a place yourself. Accommodation are of the following types.a. Studio - You live alone in a tiny apartment -> kitchen and bathroom included. Depending upon your age and rental contract, you may be eligible for rental allowance. This is expensive if you don’t get a rental allowance.b. Shared - You share bathroom or kitchen or both with others. If you are searching alone (without university assistance); when it comes to gender, ‘all male’ and ‘mixed accommodations’ are available a lot. But ‘all female’ apartments are rare to find. I also felt this to be socially unfair. When there are a lot of ‘all male’ apartments, shouldn’t there be ‘all female’ apartments too? Or maybe there aren’t enough girls seeking accommodation.c. Couple - There are options where you can live with a husband or a registered partner.There are many rental scams. So please use only reputed sites when you are looking for an accommodation. Finding accommodation on your own is difficult - especially if you are a single (unmarried) woman searching for an ‘all female’ accommodation. Please make sure that you have someone (relative or friend) to temporarily stay with for a few days or weeks in case you run out of accommodation - either your rental contract is over or you have to shift to another location because of some internship.(I understand that the Netherlands is very egalitarian with high safety standards for women. But also, I presume that most Indian female students are unwilling or not modern enough for a mixed accommodation.)Food - Unless you know to cook, food can be very difficult. When it comes to Indian cuisine - all raw materials from different types to rice to spices can be obtained in Indian/ Pakistani/ Bangladeshi shops or even shops like Albert Heijn here. I personally didn’t find food to be a problem. I cooked Indian food regularly. If you have to eat out urgently and if you are very specific that the food should be vegetarian, most shops have one vegan option. But eating out is expensive. Even fast food is relatively expensive.Lifestyle and comfort- Weather was a problem for me. It was very windy and I found cycling very difficult when wind speeds exceeded 30 kmph. Winters were harsh and cold and I had dry skin problems.The quality of life (except weather) is high. Food is hygienic. Tap water is very good. Roads and public transport are excellent. Winter clothes are affordable here. Please don’t buy them all the way from India.I found most people to be polite. I didn’t face any direct racism in the university or from the general public.Socializing- There may be events in university where you can eat together, meet representative from companies and have a chat with your faculty. Also there are cultural events in college. When it comes to events specific to Indians, Diwali is celebrated as Diya with a venue that uses university facilities. Also, state wise you can form groups and find common rooms in accommodations to celebrate. Example, Malayalis celebrate Onam in the form of Potluck. Socializing is going to be difficult if you are a teetotaler as you will find most Indians who don’t drink alcohol in India drinking it here. But events like Diya don’t serve alcohol.3. Money and expensesHouse rentFor renting a studio the cost varies from €550 to €850 per month. If you have rental allowance, depending upon the breakup of rent, cost can come down by €100 per month or even more. For example, if your studio is €600 per month, then eventually you may end up paying €480 per month. For shared accommodation, the overall rent may range from €450 to €600 per month depending upon the space and quality of accommodation. Couple accommodation costs between €700 and €800 which is €350 to €400 per person.Living expenses: Health insurance and travel charges may totally comes to €60 per month. This is given that you stay close to university and don’t have to travel much. Note that dental insurance is not covered under general health insurance.A single meal at a restaurant may amount to €8 to €10. Fast food costs €3.5 to €6 per meal. Cooking reduces cost per meal to €2 to €3.I cooked most of my meals after buying raw ingredients suitable for Indian cuisine. Weekly, I ate 2 meals outside (like Subway). My total food expense came to €150 per month.Overall, I estimate that the total cost of living in NL for an Indian student (including rent, food, insurance and travel) will amount to €500 (too stringent) to €800 (somewhat lavish). €650 per month was ‘comfortable’ for me. If you live in a medium size independent house, eat out and have parties every weekend, you may end up spending more than €1000 per monthTuition expenses: Tuition fee in Netherlands is relatively highly expensive compared to most other EU countries. For every month of study delay, you have to pay additional fees for the whole month. After your defense, don’t forget to un-enroll. Even if you unenroll of Nov 1st, you will be charged for the whole month of November. Also if you un-enroll on a vacation period, you have to pay for the whole of that period. If you un-enroll on the day summer holidays, you have to unnecessarily pay for three more months! Be extremely careful with this. Green light -> defense -> Unenroll -> Graduate. For graduation, you need not be enrolled.Overall suggestions for future students:Before you join, please make sure that you know some people who took the exact same course and exact same track. It is very important for you to get a guidance on what courses to follow and what track to follow. Your overall CGPA may improve by 1.5 to 2 points just with proper guidance. Sometimes it may even effect your passing chance depending upon the course. For instance, in one of the (bad) courses I took, the final exam was completely MCQ type. The examiner had asked 5 (out of 20) questions (so 25% of total marks) directly from a past exam which he hadn’t uploaded in past year exam section. Those students who knew good contacts (such as immediate seniors) had access to that exam questions along with solutions. They ended up getting above 8 while many just passed or narrowly failed. Now of course, all courses are not like this where past year exam matters. But you need to know which course is good and for that a trustworthy network is needed. Overall please avoid courses with MCQ exams as it just a matter of chance. Written descriptive exams and assignments are more fair.Also please don’t choose a track just because you are interested in it. Suppose you like track X, but it happens to be disorganized then your whole study is ruined.Ideally if you start in September 2021, you are supposed to finish by Aug 2023. Every month after that will cost you more tuition fee. And also, you have the additional task of finding another accommodation. Even though your course gets overs in Aug 2023, your rental contract will end by July 2023. So even if you finish on time, you still may end up searching for another whole month’s period of accommodation. Also short stay accommodations are hard to find. Maybe you have to take a one year contract if you need accommodation from some credible place. Instead, you may go and stay in some friends place if it is for a few weeks or a month. So ideally try to graduate by July. If you want to finish by July, you need your report ready by June. Which means you have to finish your thesis by May. Which implies that the core of your thesis must be completed by April! It takes one month of post processing, one month of writing and more time for your faculty to evaluate.In my batch, I do not know of anyone who graduated by July. Only a few students I know graduated by Aug. Most of my classmates graduated during September to December period. Which means that most of them had a study delay ranging from at least one month to more than three months.When you come here don’t forget to bring a pressure cooker, two spare specs if you wear glasses, calculators of all models (Casio 82, 991-es and 991-ms) and any specific product (say Himalaya shampoo) that you personally need. Different exams specify calculator models and nothing else is allowed. Calculators are very expensive here.Try to work independently without much assistance from faculty or TAs. Before you start any project, please ask a rubric of the evaluation criteria. Sometimes you may not get that. Ask any senior students if they have one. It is extremely important that you have a rubric that states how you will be assessed right from the beginning of the master thesis/project, so that you will know what to do.Finally, remember that every small decision you take is irreversible and will have a huge impact on your wellbeing since you are an international student who is paying a massive amount of fees. Get appropriate feedback and advice from trusted seniors and friends before you make any move.

When such a question is asked - the answer should always include what is the alternate.Lets not go into Caste based reservations or any political issues and look at the Education System as a whole.PROBLEM 1 - DIFFERENT SYLLABUSES BUT SAME EXAMINATIONSWe have Three Different Syllabuses in India - One is the State Board Syllabus which varies from State to State, the Other is the CBSE Syllabus which is throughout the Country and the Last is the ICSE Syllabus which is an O and A Level Cambridge Certification which is becoming the norm in an increasing number of SchoolsBased on what people say - ICSE is the most comprehensive and State Boards are the least comprehensive. This is true for at least Tamil Nadu.Yet all these students take the same Competitive Examination - be it the NEET or the IIT JEE Mains or AIEEE etc.This means that some students are unfairly disadvantaged while some students are fairly advantaged. Some Students of a certain Board are spoon fed their course - to get 200 marks out 200 at the Board level but this spoon feeding almost certainly blows their chances to shine in a Competitive examination.It means some students are unfairly disadvantaged in overall aggregates in entrance examinations which take between 40–60% of your 12th Standard marks and it is far tougher to secure 100 marks in an ICSE exam than 200 Marks in a State Board Exam. So a Kid who has scored 95 Marks in ICSE Exam will have a lower aggregate than a Kid who has scored 199 Marks in the State Board Exam for Mathematics - regardless of how different their papers were.Due to quality of teachers being different, quality of curriculum being different - the Students have different approaches to each subject which is a Problem.A Single Curriculum Syllabus - Nationally could change this entirely.One single Course - One single curriculum - One Single National ExaminationPROBLEM 2 - TOO MUCH EMPHASIS ON LANGUAGEThis is a big big deal. There is simply too much emphasis on Language. Language is regarded as a subject. While i do agree that English should be regarded as a subject - other languages should not be. A Student cannot be penalized overall for doing badly in a language.An Average Student has to learn 2 more Languages while in School - and these Languages often have zero intrinsic value. For instance - take Hindi. Conversationally Hindi or Tamil or Telugu may have value but they have little intrinsic value unless you want to read Tamil newspapers or Telugu Newspapers. Even worse is Sanskrit. The Hype of Sanskrit with modern technology apart - Sanskrit is a near dead language used mainly for Shlokas. Yet Children until Class IX are expected to learn Sanskrit in CBSE Schools.This problem persists until Class X in CBSE Schools.There are meanwhile a variety of subjects that can be taught instead which can help a student a lot such as - Electrical Works, Automotive Works, Home Economics, Carpentry, Engineering Drawing, Basic Design, Video and Photography - all of which are Taught in US SchoolsA 12 Year old is taught Carpentry and can build a Birds Feeding Table using Wood and Nails and a Drill purchased from a Superstore. A 15 year old can check out his fathers car.Even in China - Emphasis is on Practical Things rather than on Obsolete Languages.Any Language other than English should not be a subject and its marks should not be taken into the final aggregate. Instead the Student should get a Certificate after writing only 1 Exam in Class VIII - that he can read and write in that language. Only One Language should be compulsory - the other languages should be entirely voluntary.Instead of a Kid getting 76 marks in Hindi - the Kid should get a PASS Certificate in Hindi after writing his exam in Class VIII. Only One Exam - Oral and Written. Even if he gets 76 marks or 52 marks - Pass is Pass.P.S:- Our Fondness for Languages came not from UK but from USSR which has the three language formula.PROBLEM 3 - TOO MUCH CRAMMING TO UNDERSTAND FUNDAMENTALSOur Syllabus has too much content crammed into one subject to really understand concepts and fundamentals.Students are forced to learn mundanely and learn through problem solving rather than learning through fundamentals. This means a Twisted Question will result in most students getting stumped.Lets take the 12th Standard Syllabus - Differentiation, Integration. Definite Integrals, Differential Equations, Functions - Limits - Continuity - Come in Calculus, Probability, Matrices & Determinants, Vectors & Dimensional Mathematics - There is simply TOO MUCH CRAMMEDIn Foreign Countries - Calculus is a separate Subject in itself. It is different from Math.Likewise in Class VI you have too much crammed into mathematics - Algebra, Linear Equations, Algebraic Expressions in addition to a Lot of Geometry, Constructions, Congruency Rules in addition to Integers, Operations on Integers, Rational Numbers, Number Lines etc etc etc.In Foreign Countries - Algebra is a separate subject.In the Indian way of teaching - very few can understand the fundamentals or basics. Without understanding the basics - it is very tough to fully digest these topics.There should be a fairer distribution of topics in mathematics or Physics or chemistry or biology or even English. Fewer Topics but thorough exposure to all concepts and basics.For instance Class XI and XII in India offer 47 Topics in Physics combined - compared to only 23 Topics in US Schools and Syllabuses and 31 Topics in UK and Australian Schools and Syllabuses. Yet the teaching of these topics is far more thorough.PROBLEM 4 - LITTLE OR NO EMPHASIS ON PROJECT WORK OR ESSAYSSomehow our Schools regard Practical Work or Project work as beneath dignity.There is very little emphasis on Essays or Project work. Instead Emphasis is on answering questions.Example in History - we focus on- “What were the causes of World War II?” - while in US they focus on “Write an Essay on the World between 1919 and 1939”US Students have to be creative and this teaches them how to research, how to write good Essays and how to arrange your text concisely.Indian Students have a standard set of points to write. It is more of memory work than creativity.Likewise Project work is a joke in most schools. Previous years Projects are pushed to the present year for work to be done and copied. New Projects are frowned upon. Thus students rarely gain practical experience unless they are in College.Even Labs are utterly 20 years behind. Even today Salt Analysis is the basic practical in Class 12th - which was present during my Grandpas day. In Physics - Outdated devices like Beam Balance or Vernier Calipers or Galvanometers are used which offer zero intrinsic value.So in my opinion here are some things that could help in Education:-Only English be Mandatory as Language. Students can take a Second Language but the Language will have only 1 exam in Class VIII and Student will get a certificate if he clears the Exam. Everything else will be Voluntary.Only one Board of Education.Trained Teachers and More emphasis on Powerpoint Lectures rather than Blackboard Lectures with Chalk Pieces. This way Powerpoint Slides can be used for references by students.Reducing the number of topics in all classes but increasing the depth of each topic to teach the student Fundamentals. Also Trigonometry, Calculus should be divided into Basic and Advanced - with only the best students offered Advanced Trigonometry or Advanced Calculus.More Emphasis on Essays and Project work. They should form at least 40% of the Final Grade or Mark.One Single National Exam instead of hundreds of Competitive Exams. Children can choose which subjects to take based on what they want to do. This is what happens in China and it has proven efficient and effective.At least 70% Budget Increase for Education to ensure Good Teachers are attracted to teaching.Lesser Examinations and more Pop Quiz Culture to be followed to evaluate a students understanding of the subject in classes rather than 4 hours of cramming.School Hours be reduced to 6 1/2 Hours a day and 40 Hours a week maximum instead of the present 8 Hours a day and 48 Hours a week.

I think there are two intertwined questions to address here:Why do we need a special type to do IO?What do monads have to do with this?A Type for IOThe best way to start is to forget about monads altogether. Instead, look at the first part of your question: why do we need an IO type? Isn't functional programming about not using effects at all?You're probably imagining imperative programming as "generic programming" and functional programming as a subset of this, strictly less capable. This is not a good perspective to take. Functional programming is not "imperative programming but less". This is one of the most common ways people—even some experienced functional programmers—seem to see the paradigm, but it's flawed and doesn't capture the full breadth and depth of functional programming.Functional programming is a different abstraction from the foundation up. It's a distinct perspective on the programming universe. The core abstraction is not doing one thing after another; it's simplifying expressions by applying functions. It's programming by calculation and, crucially, exactly how the calculations are performed is an implementation detail, below our level of abstraction. This is what we mean, by the way, when we say Haskell is a "declarative language": it lets us specify what something means without specifying how it's achieved.So where do observable effects like input and output fit into all this? They don't! Crucially, we can't have observable effects that depend on when an expression is evaluated because that would break our abstraction¹, just like having code depend on when an object is collected breaks the abstraction level in a language with managed memory².Of course we still need to perform effects. Real world programs have to interact with the user, read files and talk with the outside world; additionally, performance considerations on modern architectures sometimes dictate the use of mutable state, especially with arrays. How can we do this when we can't rely on the mechanism that evaluates our code?The answer is that we introduce a new abstraction that handles "doing IO things" and ensures effectful things are executed correctly without leaking details about how things are evaluated. In essence, "pure functional programming" means we separate execution from evaluation. In Haskell this type happens to be called IO. I like to think of it as comprised of "actions" which can be executed to get some result; for example, getLine :: IO String is an action that reads from STDIN and gives us a String.If you think about it, IO in anything short of kernel programming has some magic. You probably don't know the special registers that need to be polled to get inputs from your mouse position. Unless you've paid a lot of attention to strace over the years, you probably don't know what system calls go into even basic actions like reading a file. Your language and runtime library take care of everything for you.Haskell takes this a step further, but also makes all the magic explicit by putting it into IO. The runtime takes care of all the low-level details and takes care of executing IO actions in the right order³, independent of how things were evaluated.Practical BenefitsSo far, I've given you a way to think about Haskell's approach to IO given that you're doing functional programming. Ideally you should see how it emerges naturally once you realize that functional programming is a parallel world to imperative programming, not just "imperative programming without side effects". But even if you see this, you're left with the (perfectly reasonable) question of: why is this interesting as opposed to normal imperative programming?This question really boils down to "Why would I use functional programming instead of imperative programming?" which has been written about a lot. Since it's so well-covered elsewhere, and my answer is already getting pretty long⁴, I'll just give you a quick summary of what I think the main advantages are:expressiveness: it's hard to deny that Haskell is incredibly expressive, and part of this (often relying on laziness) comes from separating effects and evaluation.safety: it's easier to avoid bugs and write test code when observable effects are explicitly tracked in the type system. Part of this is that the system keeps you honest and makes writing better-organized code more natural which is an underrated effect.refactoring: Haskell code is incredibly easy to refactor, to the point where it feels qualitatively different. When I dive into some old code I've written in Haskell I consistently leave it in a better state than I found it just because the mental effort needed to do that is so low.I've had projects in Haskell where I'd add features and the code would stay the same size—or get shorter—because I refactored it so aggressively.parallelism and concurrency: Haskell offers an incredible suite of tools for writing correct parallel or concurrent code. There are libraries that can parallelize pure functions deterministically (the parallel version always gives the same answer as the normal one) and some exciting concurrency features like software transactional memory (STM) are significantly easier to use and implement in Haskell for the same reason.high-level optimizations: the compiler can perform all sorts of high-level transformations on your code because it knows they are safe. Haskell can have optimizations like loop fusion as a library (ie stream fusion for Data.Vector) which is pretty incredible.There are other advantages (and disadvantages) to be sure; this is just a rough overview. The point is that people aren't just using this for kicks or because it's elegant.MonadsSo, I just spent what is probably a couple screenfuls of text talking about IO in Haskell and the IO type without ever mentioning monads. That should tell you just how orthogonal the whole monad abstraction is to IO: the IO type forms a monad in a useful way, but you don't have to understand anything about monads to understand IO in Haskell.Again, this is a topic that people have written about extensively, although most monad tutorials out there are not great. I'm certainly not going to embed a monad tutorial in this answer! What happens is that monad is an abstraction that's really useful and captures the behavior of many different types, not just IO. In particular it's a good way to embed little domain-specific languages into Haskell, especially working in tandem with do-notation⁵.That's the important takeaway here: the IO type (comprised of "actions" that can be executed to do effects) is a like a domain-specific language for talking about effects embedded into Haskell. It lets us write little imperative programs that talk with the outside world and whose behavior is completely independent of how Haskell itself is evaluated. Other DSLs achieve this by having a normal Haskell function that acts as an interpreter; IO is actually built into the language because it requires all sorts of primitive operations from the compiler. (Remember how I talked about how IO is ultimately magical in any language?)Ultimately though, the idea is the same. Here's a trivial example using IO that looks just like a "normal" imperative program in a funky syntax:main = do putStr "Please enter your name: "  name <- getLine  putStrLn ("Hello, " ++ name) Not all that different from, say, Python!print("Please enter your name:") name = raw_input() print("Hello, " + name) That's really how "monads" come into play when doing IO in Haskell: they let us use a very general abstraction to express IO actions as a little imperative DSL inside our normal Haskell programs. The runtime system then knows how to execute this DSL to actually run the program.Is IO Imperative?Presented this way it seems that IO in Haskell—perhaps IO in general—is inherently imperative.This is not really the case.Unfortunately, this is where our terminology gets in the way of a good explanation. The problem is that we have IO the concept (our program talking to the external world) and IO the type (the specific abstraction Haskell provides to do IO the concept). IO in general is not inherently imperative but IO specifically is. Try reading that aloud without the benefit of typesetting!The IO type is imperative. This is a practical consideration in designing the Haskell language: for better or worse, imperative programming is the best-understood way of working with effects. Since basically all the other languages use some variant of imperative programming to do side-effects there is a lot of existing knowledge on how to implement and optimize programs like that. Moreover, perhaps more importantly, an imperative approach to IO makes it much easier to interact with existing libraries and operating system utilities.While the world itself might not be inherently imperative in nature, most of the code in the world is, which is a pretty big deal!However, IO the type is not the only way to work with effects in Haskell. We can provide other abstractions that are declarative in nature built on top of IO. This would let us write programs that talk to the outside world in a more declarative sort of way that fits better with functional programming.My favorite example of this is functional reactive programming, which is a fascinating research topic unto itself. Instead of providing an imperative-style DSL for doing effects, it provides two abstractions for dealing with time-varying values, making time (rather than state) the core of the programming model:events are values that only exist at instantaneous points in time. A keypress is a good example since we generally only care that a key was pressed, not how long it was held. (You'd use a different API if you cared about that.)A keypress would be represented as an Event Keycode or Event Char.behaviors are values that always exist but can be varying continuously. Your mouse position is a great example: the mouse always has a position but when it's moving the position is constantly changing. How often the actual number is updated (ie the polling interval) should just be an implementation detail.The mouse position would be represented as a Behavior (Int, Int) (with whatever coordinate scheme you're using).Then instead of working with these values imperatively, we have a bunch of high-level combinators⁶ that combine them to create more complex events and behaviors which can finally be rendered to the screen or plugged into some static UI element layout (like HTML). A totally different model to IO and one that I often prefer, although it is still under active research.As it happens, both Event and Behavior also form monads (although with very different semantics from IO). However, libraries often don't even provide access to all the monadic operations out of performance considerations and simply because they're not necessary! This shows two important things about monads:it's an incredibly general abstraction that applies not only to different models for doing IO but also to all sorts of other types and DSLsit's not necessary for doing IO whether in Haskell or elsewhere; a purely functional language could provide FRP for effects without needing "monads" at all. (In fact, this is sort of what Elm does.)If you want a high-level takeaway about monads and IO, there you have it: the two topics are a lot less related than they seem. Haskell provides IO using a type of "imperative IO actions" that forms a monad in a useful way, but you don't really have to understand the general monad abstraction to understand IO and another purely functional language could provide some totally other abstraction for IO that didn't rely on being a monad at all.footnotes¹ Which is why unsafePerformIO is unsafe: it's a fundamentally foreign operation in Haskell's world and should only be used when you purposefully want to expose the language's internal behavior (like for debugging), or if you understand that behavior well enough to ensure you're safe (like for binding pure C functions through the ffi).² This is why Java finalizers are just as unsafe (more so, even!) than Haskell's unsafePerformIO. If you've just taken a couple of Java courses you probably haven't even heard of them because they're reserved for some really niche uses; here's an excerpt from an article by Joshua Bloch that explains this:Finalizers are unpredictable, often dangerous, and generally unnecessary. Their use can cause erratic behavior, poor performance, and portability problems. Finalizers have a few valid uses, which we’ll cover later in this item, but as a rule of thumb, you should avoid finalizers.That's basically exactly the same advice you hear about unsafePerformIO!³ Even if actions were just executed one after the other sequentially, Haskell would have to do something special for them since this is completely separate from evaluation. However, as it happens, Haskell has a sophisticated system of N:M threads for concurrency with its own scheduler so "executing IO actions" is actually pretty intricate, with a lot of research and development effort put into providing good concurrency primitives and performance. (See "MIO: A High-Performance Multicore IO Manager for GHC" for a paper about its most recent developments.)It might seem unintuitive at first, but Haskell is one of the best languages for doing IO!⁴ Especially thanks to these footnotes. A responsible editor would probably take revoke my footnote license for this sort of thing, but it's so satisfying! But also, some of my answers lately have been getting longer and longer, and I really should cut down a bit because answers past a certain length don't fit as well into the Quora experience.⁵ I actually wrote about this in a recent answer. Look to the final section of Tikhon Jelvis's answer to What is an embedded domain-specific language?⁶ If you want to learn more about FRP, I wrote about it in more detail in another Quora answer: Tikhon Jelvis's answer to What is Functional Reactive Programming?