 Okay, so let us wrap up this workshop with the last session. So in this session, I want to have essentially a discussion, but slightly guided discussion. So I am going to put up a few slides with some topics which I would like to discuss. But then we can, you know, tailor it to whatever people here want to discuss in more detail. So I will spend a few minutes on each slide and then open it up for discussion. But before that, let me mention a few things which we have been setting up. There have been a lot of requests for running the workshop on Windows. So we will put up, you know, how you can handle these things on Windows. So we have it somewhere here, if I can find it, here, lab setup instructions on Windows. There may not be much here yet, but that is enough to get you started. So far, Hadoop on Windows, we have not got it working. Some people are trying it out. So if somebody succeeds, let us know how to do it and we will put up the instructions here. The other things are relatively straightforward. Then the lab setup instructions for Linux, currently empty, but the session that you had in the morning, those things will be put up over here. This is slightly different from the Windows lab setup. The Windows, I am going to change the title. The Windows lab setup is for this course. This is actually system administration. The other stuff is already there for Linux. And all the other stuff you have used already. And finally, I am putting up a few things down at the bottom. The last topic was discussion on research and teaching. A few people have been asking me about resources, various kinds of resources. I am working on this. This is a work in progress. But let me just show you what I have done so far, which is not much. But I have put up links to our courses here, courses which I teach. And also some of my colleagues teach some of these courses. So the first course is the undergrad database and information systems course. So there is a Moodle page for it. You can click on that link and then click on guest login. You should be able to view the things over there. Hopefully, you will be able to see most of the stuff that is in there. If there is a problem, let me know. Then the next thing is the implementation techniques course, which is basically a database internals course. It is a second course and it goes into full depth of database internals. And there is a page which shows the how we organize the course. And also some extra material on PostgreSQL internals. This course is primarily based on, well, it covers the theory of internals. But there is an extensive project for this course which uses PostgreSQL. And the goal is to get the PostgreSQL source and modify it in some way. So there are instructions, laboratory, sorry, project suggestions and how to go about doing these projects in PostgreSQL. So if you are teaching an internals course, some of the material here may be useful. Or if you have students who want to do BE project or something in internals of databases, there are some good suggestions here. There are many, many more things which can be done. So there is a link to PostgreSQL Wiki, which is basically PostgreSQL developers saying what all would be nice to put into PostgreSQL. So that is another nice resource for figuring out what to do. But these are not trivial in the sense that hacking PostgreSQL internals takes a fair amount of effort. It is not a course project for a regular undergrad course. It may be for an internals course, especially PostgreSQL. The third one is a research paper-based course, which surveys recent or old standard papers on various topics in data management. This I teach that course, it tends to be pretty biased towards my research areas, which are query processing. And this year, there has been a very significant bias towards big data also. So if you are interested in these areas, there is a list of papers there which can it is like a guided reading list, which can help you with it. And there are also talks on every one of these. So these are talks, which either I gave or students in the course gave. So every single paper has a talk associated with it. So you can actually browse the talk and get an idea of what is going on in this area. This is by no means an exhaustive collection of papers. It is a selection which is, like I said, biased towards my interests. Other people may have other interests. So what I will do is I will try to add more links to similar courses in other places, wherever they are available. And if you find any links, let me know. If you think it should go into this, let me know I will add it, if I think it is a good link. And finally, there is some stuff on database conferences. A lot of people want to know how to do research. The best way to figure out how to do research, the first step actually, is to find out what is going on out there. And how do you find out what is going on? One way is to see the websites of the leading conferences and also journals. But in computer science, what happens is the people generally submit the most recent work to conferences first, and later on they submit more polished versions to journals. So if you want to know what is the most recent stuff, conferences are usually a little more recent. And at one side, you do not want to focus only on the most recent, on the most hot topic of the day. You do not want to overdo it. On the other hand, if a topic has been kind of cold for 10 years, the chance that you will be able to do something new in that area diminishes with age. So the newer the topic, the more chances for doing something. But that does not mean you need to change every year with the flavor of the year. So I will add more resources here for both of these. Yeah, question? Yes, let me talk about that. In fact, let me also plug our COMAD conference. So tier one and tier two, tier one are very hard to get papers into. They also tend to have maybe slightly better quality papers than tier two conferences. And we have a conference here in India which rotates around different places in India called COMAD, Conference on Management of Data. So that is run by the CSI special interest group on data which currently I am chat for that group. And this year in 2013 December, this will be in Ahmedabad. So please do come for this conference. This conference has a mix of keynotes, tutorials and research papers. So you will get a flavor of many things in there, tutorials and industry sessions too. Let us not forget. So you will get a very nice sample of what is going on out there. If you want to get into a research area, there are usually tutorials on several three, four different research areas which lead you fairly deep into the research area. There is a good starting point for learning more about the area. If you want to know what is happening in the industry, there are industry speakers who we have invited to come and give talks. And then there are keynotes which we have basically we recognize some of the eminent people across the world and we get them to come and give talks. And then in the research sessions, we have two kinds of things. One is contributed research papers. So if you have done some interesting work, consider submitting it to COMAD. The deadline is usually sometime in August, July or August. So there is a site COMAD.in which I do not think it has details of this year's COMAD yet but it will come up, keep a watch on it and do consider submitting a paper. In fact, we have multiple tracks that we have research papers. We usually have an applications track where people who have built significant applications, it need not be new research in the sense of something which the research conferences would accept. But if you have built an interesting application and would like the world to know about it, at least people from India, not the whole world, this is an international one. We do get papers submitted from across the world. So you can get visibility for interesting applications which you have developed. So do consider sending papers in those categories also. So I have put a very short list here. Now these are subjective but there is a Australian core ranking of conferences in computer science across all branches of computer science and they list conferences as A, B, C and so on. So tier one here is essentially their A level, tier two is their B level, tier three is C level. Up to C is what they consider good conferences and then there are others which they do not rank which are, they do not consider as of now as being good enough to be in the ranking. This is on model or the core ranking, yeah I can add that, just send me an email to remind me if I forget. So you can search for it, Australian core C-O-R-E ranking, okay. Any other questions on this, good. Are there are many more sites for research? So these conferences are run by, the CIGMOD conference is run by the ACM special interest group and they have in addition to the conference proceedings they have other information also there, blogs and other stuff, okay. So now let us get to the last part before we have a final wrapping up which is to start off the discussion on research and teaching. I am going to start with research, so let me put up the first slide here. So here are a bunch of topics which I would like to discuss, first of all there is all these meta question, what is research, what constitutes a publication, when do you know you have done enough work to have submitted as a publication. And then there is something which I think is a very serious issue, so Sofartak mentioned this earlier, plagiarism for assignments, unfortunately we have a very sad situation where people are plagiarizing papers and there are journals which will happily publish papers regardless of the fact that they have been plagiarized. So we have seen at least two or three cases involving my own papers, I do not know about other people's papers. So people who publish it and there are these journals which basically make money from authors and they will publish anything you submit to them, you can put it in your resume, they are generally called International Journal of XYZ. So I had a professor from Hyderabad who said IJ Star Conferences, I mean IJ Star Journal, International Journal of something or the other. So there are a lot of disreputable journals, that does not mean all the papers they have there are disreputable, but they do not care basically, they will publish anything as long as you pay them money. And so that is abused, so that is something very serious because it is spoiling India's name. And the problem is you know China also, Chinese people also do this, there is a lot of plagiarism there. But the only good thing going for them is it is all in Chinese, we do not know what is going on there, we know what is happening but there is no clear evidence. But these journals are open access journals, so they say we are open access, so guess what Google crawls them and if you search Google for something those papers do show up and they do get exposed as plagiarism, so it is a very serious matter in many ways. And a lot of this unfortunately is driven both in China and in India by the fact that our governments have in their wisdom or I would say wrong wisdom decided that a PhD is a must for promotion and that is very bad, it is devaluating the whole thing because we have excellent teachers who are you know who want to teach and a lot of institutions are about teaching. Then there are research institutions, we should not be mixing that up and putting pressure on people to do research and publish something or the other without actually having achieved anything. It becomes paper generation, the point of research is not generating papers, the point of research is generating knowledge but we have a system which is pressurizing people into generating papers rather than knowledge, so this is something I think we should push the government to tackle I feel but give me your suggestions on it. And I have a few bullets about how to do research, we will come back to it, this is by no means exhausted, this is a few points thoughts that I had, you can read it and if you have comments we can discuss. So I will stop there on this slide and open up the floor. So what software is used to detect the plagiarism in IIT sir? What software do we use for detecting plagiarism? So if you are talking of plagiarism of research papers, you have to compare it against everything on the net. So for this there is a very popular site called turn it in. Identicate it. Turn it in. That is authenticate, it is a commercial software, authenticate and commercial software is available and the Viper AI open source is available. So I am not sure about the open source version of it, the problem is it needs a web crawl to detect plagiarism for these kinds of things. So it is difficult to download software and run it because it has to have access to a lot of journals out there, you cannot just do keyword search with Google and get results. So you pretty much have to go with one of the commercial offerings for this kind. There is a different kind of plagiarism which is student assignment. If you give a student assignment and you want to check if people have copied, the first step is of course you should make your assignment different enough from what is out there such that students cannot just take something out there and submit it as their own. If you have done that step then there are other tools which will compare different student submissions for plagiarism. Those there are tools available, both you know free and commercial both are available. So name some tools here. Name some? Tools. So a lot of people here have been using something called MOS that is MOSS. MOSS. I will put up links on it with the proper link. So there are a few tools which you have to register for but otherwise free. Then there are some which you can download but you have to do some work and then there are some which are commercial. Turn it in and do all of these. So many colleges are using it and many journals, the reputable journals are using it to check submissions to see that they have not been plagiarized. Sir, as a student once I complete my paper how can myself I can check? Is it? How do you check by yourself? That's a good question. So if you did not copy from a paper directly if you thought about it and you wrote it you know that you didn't plagiarize. A few of the things we use so that we acknowledge them like we are from this view I have taken from this author. That is all. From that I have improved my technology. So I acknowledge the author. See if you have taken a definition or an algorithm from somewhere and you have cited it saying that here we give this algorithm from some other paper and then you proceed to say how you can modify it. That is not plagiarism. That is you are giving something and citing it. You say somebody else did it, this is not my contribution, this is the background. That is fine. Now if you run a plagiarism detector it will find these also but as long as it is limited and it has been cited properly that is not considered plagiarism. But if you reproduce a significant part of the paper with or without citation then there is a problem. And the worst case is just reproduce the whole paper. I saw a very nicely written paper in one of these IJ star journals and immediately you become suspicious. I found it because it cited a paper of mine and I have a Google registered for it to notify me of any papers that cite my papers. So it notified me and I said hey that looks interesting. Let me take a look at this paper and it is a IJ star journal and it is a very nicely written paper. I say wait, I sure this is and I Google the first sentence in that paper and sure enough I find a paper published somewhere else with my friend of mine which had done all of this and they have just copied it straight. Sir can you write on the paper that name, software name you told? I will put it up on this side but yeah since you asked this is off I will put it here. So the commercial one is turn it in but you need an account. Now I think MHRD has somebody has purchased it IIT Bombay has a license and some other places have licenses. So you could request somebody in one of these places to run your paper through it if you want to check. But the point is if you wrote the paper you would know but supposing a student wrote it you don't know maybe the student copied it in which case you can run it through this. Thank you. I don't know if they give you trials of this but you can always request somebody somewhere to run it for you. Thank you sir. Excuse me sir this is regarding that fake journals so I want to give some input to that. In our NIH university website they have been listed a number of journals with impact factors. So if the people those who are interested in good publishing good journals they can refer that particular website. There are the list of journals along with the impact factors so through that we can avoid that fake journals like that. There are more things regarding that flagration checking tool. So we are using white smoke type of tools which is checking not only flagration and although it's checking the grammatical errors also how to write the papers and latex also might be the good tool to write the papers in ITB format some other format. Yeah thanks so if you send me a link to that tool I will add it here. I have not used that the first tool which you mentioned for grammar checking. Of course you use Microsoft Word it automatically highlights grammar. OpenOffice doesn't do such a good job but I mean but it does check. Spelling of course and since you mentioned latex that's a very nice typesetting package for research papers it's a lot nicer to use latex than Word. I have tried both and Word really sucks for writing research papers. In fact I was visiting Microsoft for a year and I found that people in Microsoft use latex for technical paper many of them not awful. Hello sir I have a question other than research actually I want to know what is the significance of the XML in database especially XML especially in temporal statics collection a temporal type of the database like network traffic and statics generated by the simulator. So these are two separate issues in some sense they are orthogonal XML is very good for data interchange it's a way for two different applications to talk to each other they cannot agree on a relational schema but it's a lot easier to represent common data in XML schema there are standards for this and even if there isn't a standard many tools will export data and some XML format which they define but they have given you what this XML schema is they have explained what the structure means to you. So now you can write an application which takes that XML and ingests the data into your system. So XML has been very very successful in this field it has also been very successful in document representation now all the document formats which are used currently whether it is open office or Microsoft's current format they are all based on XML so it's very successful that in that sense. XML was considered also as an option for storing data and databases and every major vendor built a significant part of their database today is handles XML per say I suppose to relational but in the market this is kind of flopped not all that many people are storing their regular data which they process through the relational system that as of now very few people are storing it in XML format there are many technical and non-technical reasons for this but at one time XML was considered a very hot area for research in databases today it's kind of pitted out because the market interest in running you know storing data in XML and running database queries has come down that doesn't mean it's dead it may come back at some point that's it now with respect to temporal data that is totally of argument so temporal data has a very long history and it has been there was a lot of research on it at one point but surprisingly it had very little impact unfortunately on practice the theory was there but very few databases actually supported temporal data type beyond timestamp timestamp is very weak form of it what you need are time intervals or time ranges for it to be more useful and then the SQL language extensions were proposed to deal with time but very few people implemented it that is now changing a little bit many more vendors are now paying attention to temporal data the previous really the most recent stable release of first-class 9.2 has a data type called range which is a range of time now they don't do anything at the SQL level with it yet but it is a built-in data type and they have some SQL syntax to deal with it so you can actually do some interesting stuff with it but it's not yet fully integrated I think that's part of their strategy to eventually support temporal data basically it is used in like BSD systems RRD database that generate the collect the database and generate as a histogram output as a histogram especially using the cricket match and IPL for going the graphics so it has also good or just around Robin database RRD RRD tool I don't know about it we will look into it up as tools for spatial databases okay that's a good question other tools for spatial data so actually spatial data you know everybody uses today you know if you have a smartphone you have a map application so it has been very successful in many levels in terms of tools which you can use for it there are quite a few there are geographical information systems there are many tools you can look it up on the web and then there are geographical extensions to databases so post-crisquial and oracle both have add-ons which can deal with spatial data there's a excellent post GIS which deals with spatial data oracle has its own thing the last time I used post GIS I found the performance was not very good it was good for learning but it was not quite you know production ready at that point might have changed since then but at least for some kinds of research you can use it sir to do something new in some particular area it requires complete domain knowledge it takes a lot of time to read the things yeah like for doing my MS all those things if we able come with some incremental things which is new it's not so significant but it is a bit new with idea yeah it is considered a research or not okay that's a very good question what is research right is it something so in an ideal world the only research will be ground breaking amazingly new research but that's not how research lot of time yeah so hand it mean never happen in one's lifetime also so typical publication has some interesting ideas a few interesting idea sometimes even just one interesting idea and then once you have that core idea you build around it you you know figure out how to make it work properly with other existing ideas how to integrate it with existing systems there are lot of details in putting it to make it actually work properly and then you have to compare it with alternatives anything you do usually there is some other way of doing it so you have to show that this works well compared to the others so there's a performance study so when you put all of this together what at a core is a small idea becomes a medium-sized paper say 10 12-page paper conference paper so most conference papers are like that they have at their core a few interesting ideas and so if you have a few interesting ideas which are new you know once you work out all the details and show that it actually has some benefits you are ready to publish a paper now if that idea is really deep you can target a tier one conference if it is a reasonable idea some novelty you can target say tier 3 kind of conference or tier 2 depending on the incremental work is also good then yeah all work is you know most papers advance the field in some way so there is an existing technique you figure out how to do it better in some cases so that is incremental in some sense but as long as there is some novelty to it it is publishable but you of course how much is that delta really really trivial idea is not going to get published it has to be it's hard to define what is big enough to publish that's a judgment call so there are reviewers for conferences who decide whether your idea is interesting enough to be worth publishing and sometimes it's rejected from one you send it somewhere else they may think that the first person might say no I don't think this is new enough the other person may say yes this is what publishing if you're not interested with a degree and I want to do something new in the contrary to that particular area yeah okay then that case what I should do so the first thing is to learn about the area you can just you know so most area somebody has already started working it's very hard to pick an area where nobody has worked there are a lot of people who are working full-time on research so the first step is to see what research is there in that area today that's a lot easier because we have all these search engines and everything is online so the first step is to find out about it and the problem in most areas is there's so much research so you pick up an area and then you have to narrow narrow narrow so like they say a bachelor student knows nothing about everything a PhD knows everything about nothing so ultimately your area is very narrow but you know everything about that area so that's you kind of drill down you look at the area read papers then drill down to a sub area read papers till the point where you realize okay here is an interesting thing which people have not really thought about or the solution they have proposed are not good enough you can improve on it and that's where you get a research publication out of it thank you sir hello sir according to you which one is best tool and oracle SQL or Postgres which we should have a beauty contest of the tools just a minute no I do have a serious answer to that so it depends on your application the open source tools have been steadily improving postgres now has a lot of features which it didn't have earlier if you ask me this question I think about six years ago I would say for any high availability application don't use postgres don't use my SQL use go with oracle okay so our financial applications in IIT Bombay we chose oracle this was about 12 years ago we didn't choose postgres SQL however for our academic data at the same time we chose postgres SQL because you know that data can be regenerated if it fails today and I've lost a few hours of data it's okay I can regenerate it I have the paperwork necessary to regenerate worst case has not happened so far worst case I can do that but today even postgres SQL has high availability feature what I mean by high availability features that means if you you can couple two postgres SQL systems you run a transaction on the first one all the logs that it generates are shipped automatically to the second one if this machine dies now the other one is up to date as up to the last log record that it received and it can take over and there are tools now there's a tool called heartbeat and other tools which can automate the process of switching over so application which was using this one can now switch over to use the other one with almost no noticeable downtime so this year gate you know gate is a huge undertaking the gate application for gate you know gate had about a million candidates this year all applying within a one month period and tail end of that period they had something like I don't know they were looking at 60,000 per day or up to some 6,000 or more per hour or 10,000 per hour at peak time it's a fairly high intensity site and they use postgres SQL with high availability for that so that was done by CDAC under the guidance of one of my colleagues who was he's a computer scientist he was also chairman of gate so they used it work beautifully it handled the load no problem at all is there any data data repository for relational database any data data repository for relational database so we can download and run test of our approaches I'm talking PhD purpose okay that's a good question there is a flat files is available but we are not finding the relational database that's a good question what kind of test databases do test your things on right so there are several benchmarks called the TPC let me write it here PC or transaction processing console I don't know if you can read that there's a set of benchmarks and each of those benchmarks has an associated schema and a program to generate data so a lot of papers and databases use one of these data bases to test their ideas so most of my papers for example use one of the TPC schemas and the data generator to generate large data that you can generate it to whatever size you want one gigabyte 10 gigabytes 100 gigabytes whatever you want and then you can test your things on that but these are for a few schema each benchmark has its own schema size of the file means in terms of the records yeah so they define it in terms of total data size but you can control it you can each of these benchmarks you can generate the data where the total data size is one gigabyte ten gigabytes hundred gigabytes one terabyte I mean the generator program can generate it to whatever size you want they give you the program you can download it and learn it free free they're free sir sir one query sir for every research we do we do we have to develop a mathematical model to say that that's a real problem I have solid do we have must have a mathematical model to back it up is it a must when we do our research not exactly in computer science there are many things where you know mathematical model per se but you are you have to abstract away what you are doing so you may be motivated by one specific application but you have to abstract away what you are doing so that you can handle a variety of applications and in computer science usually there's an algorithmic component but it's not like the traditional mathematical things necessarily that but in data mining it is different there some of them have a lot of probabilistic models associated with it so that's a different catalog so it depends on the sub-area if you're working in networks there are systems papers but there are also some more theoretical papers which build queuing models so it depends on your area there's no one solution so can we get any Indian data sets where we can do research on data mining in UCI machine learning repository most of the data sets will be on other country yeah we are not able to get any free data sets pertaining to India yeah that's a good question I don't work in data mining so I would have to forward this query to my colleagues what I know is national informatics in the NIC generates a lot of data sets but I don't know if they will share it with you that's a problem with data sets okay nobody wants to share their data no company wants to share its data because they'd gain nothing and they stand to lose so that was a very famous incident this is company America online which is kind of half dead now but at one time it was one of the biggest internet companies in the world so at some point they decided they wanted to encourage people to work on a search and to help them they would publish query logs people have used to query on AOL so they released query logs and somebody got these query logs and found out that they could identify exactly who had run many of those queries and some embarrassing queries are in there and they could identify an individual the person who released the data was fired from AOL okay so he that person is actually an Indian he had good intentions at heart the good of the community and he paid for it with his job so you know that's an example of why nobody wants to release data anymore and that one or two attempts ended up in problems so it's very hard to get data sets so that's not a good answer for it so the other thing you can do is many companies are willing to share data with you as long as you don't reveal it to somebody else that you can work with the company to do something so if you have contacts in companies which have data you can work with them to analyze their data and they will allow you to publish the findings without publishing the data that's the agreement you'll have to so I know some people have done those kinds of things thank you so there are many companies in India which have a lot of data so if you have contacts in those companies you can try them hello sir actually you have listed the different courses which you would which you used to conduct yeah but for outside participants what is the procedure to get enrolled for your course as well as the different courses for their research topic which has been conducted by IIT professors these courses are only for our students there was an attempt made many years ago when distance education I mean professor Fartak when he started the distance education program around 1999 I think 2009 was this teacher-teacher but he started distance education back then and at that time we actually shared this course with outside participants but it didn't work out very well so after that that was kind of abandoned so at this point there is no infrastructure to do the course formally you can see what we did in the course you can I mean you're all teachers you can run similar courses but you want to take a course there are these massive online courses from many universities IIT so far has not offered such a thing I think professor Fartak has started thinking about it so at some point it probably will happen but right now online courses not yet come out but it will happen and the government is serious about it in fact I think one of the ministers is going to be meeting some of these companies that run these online courses and professor Fartak is going with him and professor Fartak has already started discussion so it will happen sooner or later so there are these companies Coursera is one of them which came out of Stanford so they offer a lot of courses you can take the course for free but you have to pay for certificates so that model is now available in fact what is interesting is many of our own students have been taking those courses okay we offer an equivalent course here they may do that course and do the other course or maybe they didn't do the course here because they you know they can only do a limited number of courses register a limited number of courses so they did some other course here then they go do an equivalent course online so that's a nice avenue now for learning about a new area a lot of courses out there business intelligence related to big data analytics yeah if yes how do they both fit in how do bi and big data fit in together so there's a all the traditional bi with OLAP and other tools those work with moderate amounts of data now the big data frameworks have tended to work with much much larger volumes of data and till now there's not been much interaction between the two fields but obviously people want to be able to do bi on big data so one of the first steps that has happened is that now there are SQL interfaces to big data we covered Hadoop in the lab there is a Hive project which I mentioned earlier which is SQL and then people are building stuff on it to couple business intelligence with Hive so that is happening and it will be an important area excuse me sir whether the implementation is necessary for publishing the journals yes pretty much most of the journals without implementation they are published so like I said if you publish in IJ star they'll publish anything but if you want to publish in a venue which is actually selective these days it's very hard to just tell a given idea and say publish if you're a theoretician you come up with this brilliant algorithm or something like that you come up with an amazing proof of something like Prasad Manindra Garwal you might have heard of his IIT Kanpur he and his students came up with some amazing results which hit the you know New York Times headlines and so on that kind of stuff will obviously be published without any implementation but if you're doing systems work like databases you pretty much have to implement your system to show that it works just an idea is not considered good maybe a good idea may not be but people will say prove your idea works before they publish it how do you answer that fifth point sir should we push dropping PhD as a which one should we push for dropping PhD as a requirement for promotion yeah so what do we do about it this is my personal feeling that you know this should not be coupled so tightly and there are a few others have also been talking about it now what do we do about it I don't know so I wanted to get feedback from me what do you think you think it should be kept or dropped dropped dropped these are few people would be interested in teaching few people were interested in research if you combine both both will not do anything yeah you might end up with the worst of both yeah so so I don't know what is the solution to it but people have been talking about it so you eventually these things you have to keep talking till it reaches a critical mass enough people feel that something should be done and the first step is of course to say you know why is it a bad seems like a good idea you know they do it in the US why shouldn't we do it here and arguments against that have to be built up part of that is to show what is happening with all these fake journals so you told that it is there in US US to join an academy in store you would have to complete your page and join correct in our case it's like if you have a m-tech you can join currently ought to work you ought to do research neither I can do justification what my teaching nor our research yeah it's hard now that doesn't mean it's impossible many people have successfully done it in fact I think professor Fartek and professor Sardar this was long back in the 70s they were working here as a faculty with an m-tech and doing their PhD in parallel so that doesn't mean it's not possible but it's all of you who are doing it know that it is hard it's a lot easier if you're working full-time on a PhD so that model makes more sense and we don't generate enough PhDs to man so many colleges so you know my personal feeling is this should be dealing so there may be research institutes where PhD is a requirement for joining and for regular colleges PhD should not be a requirement for promotion if you do a good job of teaching I feel that should not bar you from being promoted because many of the institute like even our institute like they're forcing parry are how to publish one or else no appraisal yeah that's a very sad state and it turns into people publishing nonsense in or even worse plagiarism in meaningless places it turns into a exercise where even the people who are doing it don't want to do it you know they know it is useless but they're being forced to do it that's really sad solution the solution is as a fact can I do something so that the solution is that yeah one thing is to make noise about it in various fora so that the noise reaches back to people who made this decision in the first place and they review their decision and change it UGC or AICT or whoever I think who made the decision UGC AICT sir AICT implementation with 5th person or compulsory yeah as they did the 6 pay implementation they made it compulsory yeah so I think if you have contacts in AICT keep reminding them that this is a bad idea and I think it's also important that people from IITs and other places make a noise also because you say they will say you are lazy you don't want to do a PhD if we say it you know it may carry more so I just put it up to get your feedback thank you any other teaching finally I put it here because not as contentious went back to the beginning there so there's not much here so but there is something important which I want make especially in the second point there so first of all you know it's a big problem all of us know it that our education is turned into road learning go mug up something reproduce it in an exam and you will pass and somebody showed me a question paper from their university and on a hundred mark paper there were options where only 10 marks required you to actually solve a problem of some kind or do a design or write an SQL query or something the other 90 marks I think were all you know just remember something and come and reproduce okay from the book or notes or whatever so how many of you feel that your university's question papers are like this how many of you feel the university's question papers make the students think when they answer it good I'm it's good to see that there are at least some quite reasonable number of universities which make students think but I have seen Mumbai University is any of in the second category from Mumbai yeah that is okay that is something which you have to do anyway the question is what is the hard question easy question mean and many university including Mumbai from what I have seen turns it into something which you can just memorize and reproduce anything which involves solving you know writing a query anything which requires thinking is thrown out essentially and there's a huge pressure that students will fail you should not put those kinds of questions in the paper and it's a disaster in some sense so the universities which are autonomous can escape from this sorry how many of you who raised your hands from autonomous sorry blooms taxonomy analyze remember understand so party contains the remember and understand question and part be apply and the party you are having analyzed and so I think autonomous colleges that's a good thing so if it's in your hands I hope everybody who's from autonomous college is following these principles yeah any comments on yeah if you're in a university which is following the road learning thing again make some noise you know it's not easy I know that there is a lot of commercial interests at work there's a lot of pressure from students but if we don't do something about it who else is going to do it all the other interests don't care the only people who care are us I had several teachers from Mumbai University lament about this that this turn into this situation where they want to teach their students something the students don't care because they're going to tuition classes which prepare them for the road learning so this is a teacher who is enthusiastic wants to teach and the students don't really want to learn because that's not what is asked in the exam and they were quite sad about this I think if that is I'm sure it's happening in many places so please make some noise and do what you can to change it okay so that's the other part you know laboratory assignments and so on this is a more practical point there so cheating in assignments is a fact of life it happens everywhere it happens in IIT our students are as bad or worse than anybody else okay but having seen that we have been trying to do certain things to reduce this it's actually the people who are most upset by it are not just faculty but other students the good students get very upset by this because they are seeing people cheating and getting away with it and an honest student gets a lower grade as a result but it really bothers them even more than it bothers us faculty so what can you do about it I'm sure there are many things which can be done I just put a couple of points here for software assignments there are tools for plagiarism checking which we mentioned a few I will put up more links so we can use those I many of us use these tools to check if the things have been copied the second thing which many of us do is make sure that for anything there is a viva you can do it for every single assignment but at least for projects and things like that students should be able to come and explain what they have done if they have copied most of the time they have no clue what is going on in there you ask them a few questions even a five ten minute viva it just shows up okay some there are some smart students who copy something read it carefully understand it and get away with it but at least they have shown they can understand it it's not you can't catch every single case but it at least catches the people who don't understand what they're doing so that's something we take a lot of efforts on for projects and so on and one more thing which I found useful is I don't just say submit a final project I say submit a project one month from the start it has to be an incomplete project if you submit a final project at that point I will get suspicious you know few people are very smart and can do it but most can't if you submit nothing there and finally submit a polished product at the end obviously I will get suspicious but most people don't do such thing they do build part of it and submit it and you can see the progression and it's clear that they have worked on it not just copied it so I'm sure there are other ways so I'll stop talking here and let you talk so yeah why are we taking the project we put this review system like we have two or three reviews like one with our guide and there's one person from department go and check whether the project is up to date or not and then second they have to submit the undertaking kind of thing from cross signed by the guide that okay this much work is already done and then second level we had again the review and on final submission we check you take three or four review in concern and then award the marks this is one thing I feel like it's going to stop this play is a man with the coping go every time they have to show the progress in a step by step manner yeah yeah that's important yeah I agree hello sir from the beginning of this workshop and earlier also we have got this impression that nowadays most of the contents are available online and even whenever we are taking classes the slides or prepared course materials we upload earlier than the class so there are at least 10% people in the class in the lower side they are always interested to explain everything in the class but all other students they are always very much interested in practical works and ready in problem solving yeah and there is a door for those 10% people personally myself always in a dilemma that what to do take that explanation in the class then this good 90% people and most of the cases they are bored so that's a very tough point to get myself teaching that how I should be giving knowledge to those people just please show a light on that yeah this is a hard problem in all of us face this we have very smart students we have very poor students so one of the things with Professor Farrakh was talking about I'm not sure in this workshop but somewhere else what he does in his course is to have extra lectures once a week for the weaker students either he teaches it or TAs are there so class help outside class because if you slow down the class too much then you have a problem so this is very important and good for the introductory courses for them you know later courses third fourth year courses we don't do a good job of this I mean so we teach somewhere in the middle and the projects usually give opportunity for the very smart students to do something interesting that keeps them engaged with the course the fact that they can do something and the bottom end you know case by case so yeah others can shed light on this what they do sir my question is related with the teaching learning methodology it is in the professional studies especially in engineering how can we ask students or promote innovation process in specially UG because in the university curriculum there's no place for the innovation it's not research it's pure innovation the new idea yeah that's a good point I mean we confuse research with just papers I mean innovation is a much better innovation yeah so I think that is very important so in fact I think you know PhD is very tightly coupled with this thing called research but innovation is something everybody can do even a undergrad student can innovate so I think the system should actually reward this kind of innovation if you can build a you know an application which serves a need I think that should count for something so for your promotion for example supposing you build a system which does something interesting and useful that should count as much or more than a research paper published somewhere that it could be a software it could be other kinds of innovation so yeah as far as student innovations go we find that almost every course we have a project associated with it and they do innovate so the project is in many courses open-ended they can come up with ideas for what to do in the database course people have done excellent projects they built huge applications they they've written in in a course they're doing five other courses but I've had groups which write a 70 page system requirements specification and 10,000 lines of code and it all works beautifully nicely designed so students really some of them and they learn about new tools although all the new tools which I learned about in Derek eventually it came from students who first found out about those used it in their projects and showed that it actually does something useful after that we took it up and made it standard part of the course. Sir more relatively for the colleges where we do not have the campus facility like the colleges in metropolitan cities like Mumbai the student used to travel a lot so light from right from 9 to 6 or 9 to 5 students have do not have much time to work on that so if at the campus facility or hostel facility the picture would be much pleasing yeah so this is the point we're trying to do something for this that's a good point it's not easy to solve but yes yes yes yeah the other thing is students are able to work with each other at any time they want they go home they are cut off but these days with the internet and phone and things it's actually a lot easier for them to work from home and to cooperate so I think that is partly solving our problem okay so let's wrap up then I'd like to thank you all for coming and staying although so here especially for staying till the very end of the session