 All right, so those 1, 2, 3, 4, 5, 6, 7, 8 people who are present, can you just share with me how far you have progressed in the literature survey part that you had undertaken? First of all, are there people who are doing seminar this year? Can you raise your hands? 1, 2, 3, 4, 5, okay. So how far have you progressed? What about you also doing seminar? Okay, any other person who is doing seminar this year? So how much have you progressed in your literature survey? No papers yet. What about you? So when you say we, you mean there are a group of students. If I notice, most people, majority of them are actually looking at specified papers. Generally 2, 3, 4, 5, 6 or 7 kind of thing. But very few seem to have followed only one so far. They actually perused a whole lot of literature and collected about 25 papers, right? Now that was the approach that we had suggested as part of the communication skills course. So that is a mandatory assignment by the way. That has not been done. Those of you who are already busy looking at very specific papers in details might feel that this is a waste of time. Because you have already identified, either jointly with your guide or individually, a few papers which appear to be definitely relevant to the topic that you are working. Having done that, it does not seem to be very useful to go back in time and look at or identify or search for some 40, 50 papers in that case, right? That is correct as far as your seminar is concerned. Because that is the approach you are following, guided by either your friends yourself or your guide or whatever. And I have no problems with that. But as far as this course is concerned, we had outlined a methodology. Even though you have not followed that sequence, it is important that you do that task. You will recall that I had uploaded a sample of, I think, summaries of 3 or 4 papers where I had said what would you write about a paper. I had even suggested that if you have a soft copy, you can just cut and paste and include it, provided you read that abstract. If not, you would have to type in a few lines describing that paper. But there were two important components of that exercise. One was you methodically learn and in fact entrench it in your head as to how to correctly write the complete reference. The names of the author, the title, the publication place, etc., etc. So that is something which is an important exercise and you should be able to do it for a large number of days. The second thing is this exercise would help you increase your breadth in whatever field you are in. Now we have a dilemma that you have progressed already in your seminar literature survey by doing these exercises. But you will still have to go back in time. I have requested Firuza to create an upload link where all of you will have to submit a list of, I would say, a large number of papers pertaining to your field of study in your seminar. We already agreed that those who are not doing a formal seminar can take up any one topic which I will describe very shortly, any one of the end topics that will list. And since it is not a formal seminar for which you are working, you would do relatively less work as compared to other colleagues who are required to do a seminar formally. But still, substantiate this approach that for the chosen field you will select a search for a few papers which appear prima facie relevant as I said flipping through them just looking at the abstract and noting it down and then reading maybe at least one paper more thoroughly and providing a literature survey only comprising of that paper. So this is for those who are not formally registered for a seminar. All those who are registered for a seminar, the submission has to be more elaborate. You have to have a list of 30, 40, 50 papers. Number is not important like somebody already has 20 papers identified. You might want to search a few more just to ensure that you have adequate breadth. But that is an activity that has to be done. It will take at least one day of work. You can do it over three or four days or whatever. I am stipulating next week, the next Tuesday, the next Monday evening will be the submission date for the first submission which is a repeat again which is a summary of a large number of papers. I will leave large undefined. It need not be 50 or 40 or 30. It could even be 20. But there has to be a representation of breadth on whatever topic you choose. This submission will have to be made before next Tuesday. And next Tuesday is what? 20 seconds. The next Tuesday will be that submission. Exactly one week from then which is April 29th March. Exactly one week from then, 29th March, you will have to submit your literature survey portion for the chosen topic of seminar. And the first part can be edited using any editor. You must submit two files, an editable file. So if it is a latex thing then a latex file. It is a word processor that you are using, that word processor file. And in addition a PDF file. So these two files will have to submit by 20 seconds. The second submission which will follow one week later should be done using latex. Because that is the way I expect you to submit your regular literature survey or the seminar report. We will follow this up by individual presentations on your research topic which will be done in the first and second week of April. We will announce the schedule. These will be longer presentations and you should treat them as actually a practice session. When are your formal seminar presentations? They will be before the ensign or after the ensign? They are after the ensign. Good. Then you will have a mock dress rehearsal before the ensign so that you can actually do it. You are most welcome to invite your guides also for that mock seminar if you so wish. But we will record it and we will keep that recording available. Since I have been a bit tardy in ensuring enough enthusiasm for physical attendance in the class sessions and you can see the result, I will swing my pendulum to the other extreme and insist that unless these two submissions are made, people will get a failed grade in this course. And there is no compromise with that. That announcement with a mail will go to everyone. So one week hence, a summary document listing briefs of as many papers as you wish. As I said, 20, 30, 40, 50, whatever numbers you can go through. For those who have a regular seminar topic, for others who will select one of the topics that I will mention because this is not a part of their registered seminar and therefore all those people who are doing an extra work I would suggest that they would try to get the bread in that particular chosen topic by just searching for say 5, 7, 10 papers. So limit your time, not the amount of work that you do. And that is the time you would spend. Those who are not registered for a regular seminar. Such people in their second submission are expected to study only one or two papers of their choice in depth and submit a literature survey of those two papers. Others should be guided by whatever is the regular seminar guidance you are getting from your guide and whatever papers like you have already reviewed 7 papers or something. So if such is an activity, then I would like all of that to be reflected in the literature survey for your formal seminar. Is that a fair game? Fine. The papers are collected. Yes. They are. Yes. And then they can be abstracted. No, no, they can be. They ought to be. There should not be any detail. A segment of the paper at all. Oh, there seems to be some confusion. There are two documents that you are supposed to submit. For the first document, I have already uploaded long time ago a sample description of what this collection of small abstracts is. That's right. So I think it contains some three or four papers, arbitrary selected from teaching computer programming. And I had collected the references where the references were properly written. That is the more important part. And the abstract was simply cut and pasted. The only requirement was that you must read that abstract line by line. It is not cut and pasted. If you read it, that is your first perusal. Just that abstract. If in addition you can flip through the paper as Sahana Murthy had suggested, that would be useful. But the whole exercise is to enhance your breadth of knowledge in that particular. That's the first submission. The second submission is the proper literature survey part of your seminar report that you would be submitting for your query. And that literature survey should actually contain a more detailed understanding of a few papers. And every paper that you have cited in that literature survey should be put at the end of the reference list. So in short, if my seminar right up is going to be chapter one introduction, chapter two literature survey, chapter three something something. I want only the chapter two literature survey portion to be submitted as part of the assignment for this. But the list of references that you will code, because you might be preparing that list for your entire seminar report, it might be slightly longer than the actual references that you have cited only in the literature. Please do feel free for example, if you so wish, to submit the draft of entire seminar report if you are ready by then. But the dates are sacrosan, next Tuesday the first submission and following Tuesday the second submission. I do hope that this will help in speeding up your activity for your seminar talk. So what I found out, like we are not aware of the proper format of the seminar file. Sorry. Like you said that first chapter will be introduction, second will be literature. It depends on how you decide to organize your seminar report based on your discussion with your guide. Because there is no fixed format per se. So your entire seminar report may be just one section without a name or number. Or it could follow the footsteps of a regular EMTEC project or a PAD thesis which has chapter sections and so on. But that is the matter between you and your guide. Okay. There is no fixed format for that, like you were telling. No, as far as this course is concerned, I want you to minimally submit the literature survey portion of your seminar. Whether it is a separate chapter or a section or something is up to you. But what it must accompany is the list of references which you will cite in the literature. So what is important is to ensure that your citation is appropriately made. So if you cite reference number 2 somewhere or reference number, let's say HK74, then in the list of references that HK74 should be appropriately identifiable with the list correctly mentioning all the details that are required as per either IEEE or Department of Standards. Authors, title, yeah that's it. Any doubt about that? So through that today we will upload the link and we'll send a mail. It may come late tonight because or it may come tomorrow by the way. The link will be uploaded tomorrow. So those who are already ready might have to wait at least for one day. So let me come back to the discussion that we had on the issue of building and nurturing collaborative communities. Although as I said, this is a generic topic meant to be handled by those students who are not registered for any regular seminar. This topic is of general interest and definitely of specific interest to computer scientists all over the world who are actually trying to build systems which will permit this. The other day we identified players, contributors, reviewers, teaching assistants, teachers, all kinds of names were given and I only pointed out how synthesis is difficult and how just writing down unique taxonomy could tax us so much that we may not be able to uniquely identify or even agree to different names for something similar. Today I propose to look at some of the actions that these players will take. The actions which result in something tangible that needs to be handled by the system that we build. Can you see that? The first thing we talk about is contributions. Like the reports that you are going to write are your contributions. The contributions could take many forms. For example, there could be an article. The article could be a research paper, essay, note and I have put dot dot dot saying you can actually expand on as many things as you can think of. Technically, since I will be dealing with text and images and audio video, I have generically written photographs to represent images and written videoclip. Why these three types? They could embrace any field of activity by the way. But generically, when we speak of content, what comes to our mind is a written text or a photograph or an image or an audio videoclip. When I say videoclip, it subsumes audio. So I have not written it separately but there could be an audio clip itself as a content which could be given up. Once a contribution is made by anyone, the contribution needs to be assessed. So how do you carry out assessment? One thing that occurs to us as students and teachers is grading. Do we do grading ordinarily for every contribution that is made? When we talk about manual or automatic grading in the context of education or teaching and learning, what kind of contributions we are trying to grade? Answer books in a class test or a quiz or submissions of the project reports for a course project or submission of your literature survey, for example, if it has to be graded. Generally, large-scale submissions have to be graded either manually or automatically. But this brings us to another question. Did we write as part of the contributions here? Did we write a test paper or an exam paper or a quiz? Would that not be a contribution? So we will have to add that to the contribution list. This is a very simple example of how our process of thinking and elaborating has to be iterative. So that is why when we write contributions, it might be prudent to write this in a separate page and keep this as a growing list. I will show that as contribution plus on a separate slide. But in practice, this is what you should do. That whenever you think about, let's say, assessment types, grading, now you suddenly realize, oh, grading is done for answer books. And where are the answer books as a contribution? Not listed. If answer book is a contribution, question paper is also a contribution. Where is question paper listed here? Then I suddenly realize there have to be quizzes, there have to be project problems, there have to be tutorials. So all of these become additional contribution types which I must incorporate under contribution. This is what I meant by iteration. When I start looking at something else, I remember that I have done a bit of a sloppy thinking on some previous slide and I have enlarged it. Assessment by voting. Are you familiar with this? All the many contests on the TV shows. So how do you decide who is the top singer? There is, of course, an examination panel and so on. But have you not come across voting vote for me? And the person who gets maximum votes is declared a winner. What do you think is the sanctity of such a voting process? Depends on the, depends on the crowd. In what way can you elaborate? If the crowd is knowledgeable, okay? Depends on sentiment. So knowledgeable depends on sentiment. This is a third dimension I might want to have. This is a perception. I might not be knowledgeable. I might have no sentiment involved in one of the three contestants. But I have a perception. Take music, for example. I may not be very knowledgeable to understand what Raja is being sung or how well prepared that singer is. So I am not knowledgeable. As she says, I have no sentiment because all the three people are unknown to me, I don't know. But my perception depends upon how I feel when I listen to the singing of that person. Because music relates to your soul in a way. Now if there are three contestants on whom I am voting I might genuinely vote based on my perception or how I like a particular song, not the person. Or how I like the particular singer. So this could be a third parameter. Is it possible at all to analyze a given vote to find out whether that vote was cast based on knowledgeableity whether that vote was cast based on the sentiment which is often could be a political process. There are a lot of, my friend comes and tells me, Arre Fadak, vote for that person. So that is also sentiment, which is being artificial. Or whether the voting was based on my perception. Now you would agree that you would like to count the votes which are given A, based on knowledge and B, based on perception. But not necessarily on sentiment of any other kind including influences. But is it possible to distinguish? No. General voting, it says either rank votes number one, number two, number three or this person and no one else. Or if it is voting in terms of what you like best is that thousand submissions, twenty get the best of this thing, majority wins kind of. Surprisingly what has been found is that if the crowd is large enough and if you can actually do some kind of partitioning based on identity amongst the crowd of people who have previously voted and they have actually assigned the best grade to artifacts which were termed as based by a majority of people including a subsequent review process etc. which I will come to. So there is a lot of science behind this kind of voting based on the crowd and based on an appropriate partitioning which we will look at later. There are a lot of peer review. All of you should be familiar with peer review. How many of you have published a paper or sent a paper for publication? One, two, three, four. So you are aware of the process that when you submit a paper it is reviewed by a panel and generally you will get a feedback. Now comes an interesting part. First of all, the review is often blind. That means you don't know who reviewed it. To enhance the effectiveness of the review. If I know that my name will not be declared I am likely to be more truthful and rather harsh in reviewing and giving the feedback. But what does that feedback result in? So I give you a feedback. Let us say you submitted a paper and there is a panel which collectively adjusts the submission at level 6 but says that if the following shortcomings of the submitted paper can be taken care of and paper be revised then it might be upgraded to 7 or 8 which probably is the minimal requirement for acceptance. You all agree that this is the normal peer review process? But then this will result in other contribution which is the revised draft. Have we listed that under our contributions up there? We have not. So that means we will have to add the revision and the revision process also as part of contribution. So at this stage I have put this slide as contribution plus. That means I have a separate page on which I started writing contribution types. I had forgotten to add, you remember, quiz, test, examination, tutorial, project description, set of problems, etc. These are all submissions and these are relevant to the field of education in fact. These may not be relevant in some other context but they are relevant here. It might be useful to actually write a context. So you are also developing a taxonomy as you write the different types of content which you need to handle. The second point which just arose when we discussed the peer review mechanism that there will have to be revision. Now when you talk about very large scale content contribution review, etc. You need to introduce a solid mechanism for version control. So how do you control your versions? I remember 40 years ago when I started teaching programming you did not have files on a desk because there was no desk. There used to be a bunch of cards that you punched your program in. That was from 74 onwards. Before that it was a paper type. So there is no question of a file name for a program. You actually own that bunch of cards and that is your program. So you say what is the name of this program? You say dv fatal. That's my card day. When I submit it at a place under my roll number and name and they would assign a number to that like an inward number. 3 days later when I go to the same window I will get that take back and a print out of the compilation. Usually for the first attempt the print out used to be one page only which had only half page which would say that the control cards themselves were not properly read. So the whole deck is written. I modify that, submit it again. So it used to be one single compilation it was not uncommon to have three visits to that window which means 8 to 9 days. Can you imagine that situation? To compile a program correctly it would take 9 days. Not 9 minutes not 9 seconds not 1 9th of a second. So those were hard days therefore revision control became essential when we had to spin off a program which was a working program but we had to modify something. When you modified it what would happen is that the modified program would stop working. Now you are required to reconstruct your original program because that was working and the changes that you have made if you are not clear of what changes you have made where then you are sunk. When the first unique systems developed you will notice that the notion of delta that came in it came in precisely because of such requirements but I digress. In case of content the revision and version control has to be implicit and we are talking about building a very large system then we must provide this version control to the contributors. You cannot expect individual contributors who remember we are talking about building and nurturing large collaborative communities a single community could be 10,000, 15,000 people strong and you cannot expect every one of those 15,000 people to have recourse to appropriate computer tools so many might depend upon what tools are provided by the system and therefore revision control has to be inbuilt here. In any case whenever a formal submission is made if that submission is a revision of a previous submission independent of how the original author is maintaining revisions at his or her end I must maintain these revisions in the system that is a requirement that is why it is written. So again to come back I will always have a growing list what I have said contributions contributions could keep on growing I must always maintain a growing list as and when I come across something new something additional I must write it. After the contents are revised or something they have to finally pass muster. So for example when your paper is accepted for publication it goes to a publication publishing house elsewhere or whatever some company Tata McGraw-Hill whatever what does that company do now does it accept the final manuscript as cleared by the peer review process saying the paper is accepted and just publish whatever format you have submitted your paper in they do that so have you heard of editorial boards what do editorial boards do what do editors do editors will ensure a variety of things but the main crux is that whatever material is finally published is absolutely perfect in terms of both its appearance and content in the context of the publication magazine that it is supposed to be contained. For example certain scientific journals insist that no political views be expressed I might have inadvertently in one of the paragraphs mentioned something about some place which my peers who are evaluating only the technical content might have overlooked but the editorial board must say that no this is not all some simple grammatical mistakes might be overlooked by peers concentrating on the content an editorial board cannot neglect that the language that is used has to be perfect, has to be proper has to be correct a format would be prescribed for publication two column format this format, this font you know all the regulations for example IEEE publishes them and you are required to follow them generally to ease that final process many people insist that your initial contribution themselves must be in that form but it is not necessary in any case you would have finally an editorial board and therefore editing of content would be an important let us look at the names of people and tasks that I have mentioned editorial team how is editorial team selected for let us say IEEE communications or IEEE journal and computer any idea how is the editorial team selected if you have seen a journal you would see the editorial team there often the editorial team is augmented by nominations so people are nominated the very senior well known professors who have actually published quality papers are invited to join the editorial board and you will suddenly find their name appearing so editorial boards of such large technical literature is often done by nomination when you have a crowd how do you nominate voting so what kind of voting so let us say there are 10,000 people in a community and you want to create an editorial team in that community let us say that community looks at a specific sub-sub topic data structures and algorithms so there is a community who actually works together on data structures and algorithms there are 10,000 people in the community they all become members through your earlier mechanisms we have not yet discussed them but I assume that they have become so how do you go about voting do you ask all 10,000 people that please select some members say 5 from the remaining 9999 and on what basis do you want 10,000 people to submit their bio-data and each one to read all 9999 submissions half the population of your community will disappear if we ask such things to be done so direct voting for that editorial board kind of thing will not work is not a meaningful lecture any other suggestions yeah, reputation so remember when we are talking about a few people or a small community the reputation is known to a large number of people but when we are talking about 10,000 people who arbitrarily joined that community because you have opened it up saying this is a community I want how do you manage that reputation how do you understand or capture that reputation ok, so here is one possible way one possible way is that to kickstart the process you ask all 10,000 people to make some contributions then you ask all those people in partition groups to access those contributions then you quantify the access contents and the assessors find out which contents have been assessed as the best quality you say scale 1 to 10 and you just look at only those content which have been assessed as 10 or 9 now you find out who are the people who assess those 10 or 9 content and are there any commonality you may do this process automatically but iterate him so you may get the same content assessed by multiple people particularly those who filter up or which bubble up so you can set up a process an automated process which will actually throw out the reviewers who generally do good things correctly that means they are able to decipher between a good quality and a bad quality but you want to arrive at this judgment automatically based on the work which itself is part and parcel of the building of community this is one possibility we do not know whether it will work or not but this is one possibility similarly we have to think of other possibilities to create such teams automatically it is not just the team selection but leadership please remember that in a crowd leadership is voluntary so the task allocation which is to be done both to the leaders and by the leaders for the entire community auto monitoring of the activities of the people if 10,000 people who form that community if you find that about 1,000 of them have not participated even in one activity for more than 2 months by prior announcement you can say that such people will be removed from the community others can join so there is a dynamic equilibrium of some kind clearance by the editorial team for some contents which have been reviewed and this clearance will require adherence to all the formatting specifications that it should and finally it will result in a publication now these are the tasks which are ordinarily done in small community we are talking about automating these tasks and automating these activities for crowds and not for one kind of crowd but for different crowds or different subsets of crowd so we are really really talking about building a system which will permit building and nurturing collaborative communities in large numbers so I can think of about 2,000 communities coming up each one with a minimum membership of 10,000 people and even that number is a miniscule proportion of the population of just this country not the global population that is the sort of dream in that context I have listed here some topics which might be relevant for developing such large system here we are obviously in a communication skills course not talking about developing that software but we are talking about doing literature survey on one part a partial literature survey or one part one of the end parts which together might actually contribute to building so here are here is a sample list for example content management will require content management currently we are looking at Drupal as the content management how many of you are familiar with Drupal one oh very few oh wow very thriving global community of Drupal we had actually a global event here in IIT Bombay recently they are all computer science people not even heard of Drupal heard of it so all of you know the spelling useful anyway but that is something that we would like to explore in particular we know for example that the latest version of Drupal has a set of powerful APIs in fact an API mechanism which will permit systems to be constructed where Drupal is the binding glue and through APIs you can connect to anything and everything behind it so that is something that we are building the national virtual library also using that glue incidentally the local expert in IIT Bombay is not a computer science professor but a professor in chemical engineering Professor Sundar so those of you are interested can look at the second thing is crowd sourcing now this term you are all familiar with have you read any papers on crowd sourcing no you have one to find I have just given one sample name here just like I gave one sample Drupal Professor Mausam who is professor at IIT Delhi has written some papers so those of you are going to do some literature survey on crowd sourcing might just limit themselves to look at Professor Mausam's papers as far as this course is concerned the third item I have written is cloud computing cloud many people should be familiar with not many what is happening your CS department students right one two three four cloud computing you are not familiar with generally familiar with how many of you know about open stack cloud one two very few so what do you know about cloud virtualization have you studied virtualization I now have to talk seriously to the head of the CS department because I find minimal required thing in the 21st century that must be known to every CS graduate somehow are not exactly vociferously available they are sort of available here not good I will tell you cloud computing has to be minimally understood by everyone so it is not part of the communication skills course but I would advise there is a part of the computer science skills course should acquire adequate knowledge about cloud computing in any case in this particular instance we want to look at the open stack cloud implement so what is cloud what are the different virtualization mechanisms that are available etc etc I don't know but if I mean has nobody worked on things like process migration in a virtualized environment and such things not read any papers or any literature or any textbook on it this is a serious lacuna as an independent innovation from a teacher in computer science I would say that you should acquire that knowledge however agile development how many of you have heard of this many good agile development is if you have studied software engine conventional classical waterfall model iterative model etc etc but today most people are using agile development methodology so this is related to software and the last is project management project management is not something that is unfortunately taught very carefully to engineering students in general and to computer science students it is considered part of the management how many of you have done a course in project management or studied a topic in project management as part of some other course nobody those of you who are going to join the profession after passing out from here and are going to work in industries better learn as much as you can about project management it is a vital thing because you will all be working in teams and teams cannot deliver unless proper project management is put in place how many of you have heard of JIRA one two three four very few so JIRA is another important open source software tool often used in the past for software bug tracking so software developers used to use JIRA very very effectively for managing their software development projects a very large number of open source projects all over the world particularly the large ones use JIRA JIRA also has a management component a project management component in this particular instance I am talking about project management I will upload these slides as a pdf file today but this is the list one two three four five those people who are not registered for a regular seminar as a part of their activity this semester are requested to pick up any one of these unless they have already chosen a topic because someone said you are reading a book for example so that is perfectly fine so those of you who have not fixed the topic and are not doing a formal seminar are encouraged to pick any one of these and make your submissions next week and subsequent week based on this is that okay? fine the submissions are mandatory attendance is still not mandatory but submission is mandated and those of you who actually wish to attend like many of you came I would suggest that it will be useful if a 8.30 session starts at 8.30 and not at 8.45 is a humble request from me and about six of your colleagues who were there from 8.25 onwards so for their sake at least if all of you can show up in time it will be useful let us see on Thursday morning if the session can start with your kind cooperation at 9.30 sharp thank you