 The topic of a project is OER repository platform. I am Sunakshi, he is Ankur, Vibhav and Harshwar then. Open educational resources. What is an OER? OER are freely accessible resources that are used for teaching, accessing, learning purposes and for research purposes. They are released under creative common license. The question arises, why do we need a platform? So the requirement of a project is to create a unified platform for searching the contents which is user friendly. We have made the platform in a software called dSpace. Now, Vibhav will continue. Good afternoon. dSpace is an open source repository software platform that we have used to develop our platform, dSpace. A lot of people won't be familiar with dSpace, so let's get some idea about dSpace. dSpace was developed on Java Spring Framework and it uses various Java technologies for various functions. It provides interface for administration, deposit, ingest and search and access. Now we'll look at the various technologies that dSpace uses. Now, talking about user interface. dSpace has two user interfaces. It supports two types of user interface, JSPUI, XMLUI. We are using XMLUI for our project. It has solar for indexing searching. It contains for database. It supports Oracle MySQL database, also Postgres SQL database. This is dSpace architecture, sir. Whatever fields we have circled, we have worked on those parts. There is an application there, business logic there, storage there. On applications there, we have worked with WebUI. We have configured it. By default, we were using XMLUI. We have configured it with a theme created by Atmire. We have used Mirage theme. Also, we have used Bootstrap to configure the UI to change various stuff. In search field, in business logic layer, we have modified search filter to make it more educational. Make our website more educational. Also, we have worked with authorization and submission workflow manager. This is the current hierarchy of dSpace. By default, dSpace has this hierarchy. It has top-level community, second-level sub-community collection. Inside collection, we can have items where we submit the digital contents. From sub-community, we can further create collections and sub-communities. And this line goes on. What we have done is we have treated this as a discipline or department. This will be our course. This is lesson. Inside lesson, we submit items. We have modified the hierarchy like this to suit our needs more. So what happens by default in dSpace is dSpace gives default read policies to everyone. So if user has to contribute or he has to add a module, he has to ask for the permission from administrator. Now, let's say we have 100 users, 200 users, and there are like two administrators and everyone wants to submit. It will be very problematic for administrator to grant rights to everyone separately. So what we did here, we changed the default policy of dSpace from read to read, write, edit. Read, write, add. This way, any user who has an account can submit a course. This created one more problem is that the problem of security. Let's say I create a course and he creates a course. I should not be able to modify his content or he should not be able to modify my content. So for that, we have changed the, whenever I create the course, he will just get the read, write, and I'll be the admin of my course. If I want, I can add him on the submitter or reviewer or any other role, but I will have the admin write over my course and I can just read his course. Now Ankur will... Coming to the metadata schema, this whole dSpace system is based on tagging. And this tagging is based on the metadata schema. And this, at initially, when we installed dSpace, there were some 67 fields which were provided at initial for addition of any new item. And those fields were not enough for the assets which were uploading, like the PPTs, the PDF, and the scanned documents. So we just, like, we found a learning object metadata, which were more oriented towards learning resources. And from there, we identified some nine metadata tags which were subject, which defined what the item was related to, topic, interactivity type. It defined whether the topic was of practical or it was theory-based, learning resource type, context, format, difficulty, if the course is difficult or is easy, and the learning time. Like, after changing the metadata schema, we were required to reflect the schemas while submission. So these were the schemas which were added by us. Difficulty level, interactive type, intended, end user role, which usually tells the viewer that who's, like, for the digital assets, who are the targeted audience. And initially in dSpace, like, dSpace works on solar in the background for searching. And by its default, there were some five fields upon which items can be searched. And those five fields were not enough. Like, those fields were usually topic, author, the date of its issued. So anyone who wants to search any learning resources, it was difficult for them. So we modified our search engine so that it can include more five metadata tags. Only you increase the number of metadata or any algorithm you have changed. Which algorithm you are using right now? Solar uses inverted indexing and it uses research in its background for searching. So have you tried some lattice search or something, which is basically used in the educational field, very popular? So at present the system, like the deal, solar was provided in the system itself. You improved the search, that is why I'm asking what improvement you have brought. One thing is that you have increased the metadata. But that doesn't mean that you have improved the search. You have improved the metadata. Search improving in the sense? So because even if I'm coming on the system and I want to search, like... So I have added the more five with that five. Like, I will give you from five to ten. So those were more relevant. Initially you would require those five but to make it more accessible. Like the extra five were added. Sir, the point was to make it more educational wise. See, sir, let's say C programming course. It's taught in class 9th, 10th, 11th and even in first year, sir. So we have added one more field which says which kind of course it is for school level or high school level or engineering level. So collection of metadata is manual or automatic? So it's a manual. While we are submitting an item, we have to provide the metadata schema. So then as far as this slide is concerned, I am saying that just you have increased a few more metadata but there is no intellectual work in this. Just increase the metadata, people have to fill in manually. So no sir, the main system was like when dSpace was initially configured on our system and it provided some set of 67 fields. Like those were not learning resources oriented. So we needed it to be more... You customized it. Yeah, we needed it to be more towards our system. Like it was like for our general sense. Like when you're uploading a file and you describe the file like what's the date or who's the author of the file, what's the type of the file and giving this metadata... But it is only the changing labels. There is no such logic available for this, no? So this is required because like at present our system loads of files like in our internal server and we are thought of importing data from other systems. You will be more impressed if you have created the metadata automatically, if you have brought some intelligence kind of thing. It's okay, customization is also work but it is a laborer's work. It has no intellectuality available. He is saying assets, I am saying learning resources. So it is just a label change. The content is same. And he has to fill the content manually. So there is no automation available in between. We changed the metadata keeping in mind that like in future we wanted the system to import data from other resources. Like MIT OCW and other... So whether there are some APIs or something you have written to pull the content. Yeah, there are sword server in this. The space to access the data. We were not able to configure it but if the system is... We are more interested in your work. What is available is okay. What I am saying that you have created any API or something to map the content of the OCW to the disk space or something. Some bridging you have done then we are very happy. If available, it's okay. That is available. The main sense was that like this is an open source system and there were no available documentation for anything. So configuration itself was a... That is the challenge before the open source things. Many times documentation is available. Even if it's changing its little source code like we have to go through something on thousand files. What is the language they have written? It's Java. The back end is in Java and the front end like it has two types of front end. And database? Database is Postgres. And just database was just for the authorization like it shows the user documents. Then the content he put into... Content is done by tagging by solar. So then this all metadata and other things where they are going to put? It actually... 5 storage. XML. XML or a text file? Sir... Where the hashtags or the... Your metadata... Solar has a JSON file like it has all... The tagging of whatever metadata tags... That is the best practice actually because it is the fastest search rather than the databases. Solar was the best search engine available at like for metadata. These were the new types using that which we can search for. And initially like there were no... Like this system was just for like... You can upload a file and you can download a file. And like there were no... Anyway like you can view the file like media files like videos, mp3, audio and image. So we like change the source code of an integrated plugins for each of the files which we mostly require. Which codec it will support? It supports like for mp4 and WebM mp4.ac codec. Which player you have used? Sir I am actually using a Web VideoJS plugin which supports... The native player or something else? So it's a VideoJS plugin, JavaScript plugin which implements both flash and HTML5. Like if the browser is a legacy one it will fall back to its flash... So it is a HTML5 built in native player? Yeah it uses its own codec for streaming the videos. Okay. And that was required because if we were using HTML5 inbuilt tags and this system was not able to play the videos as the video being of last size and these space were not providing those size in the actual way to render the videos. So we were forced to use a VideoJS. Like let's say a file is one jv and it starts downloading like user won't... like user maybe doesn't need to download that file like you can see it first and then decide what does he want to do. That's fine. That I understood. Good. And even PDF streaming plugin we used because mostly our system consists of scanned documents which are most like on 200 plus mv. So video streaming plugin means what you have used? I have used PDFJS. PDFJS. So again it is a flash dependent then. No it's HTML5. HTML5. It's HTML5 dependent. It's actually the PDF plugin used in Mozilla. So the mandatory thing is that user should have Mozilla only? No. It's a plugin which is actually integrated into the Mozilla. Okay. But it's an offline plugin which have integrated it which can be used in any browser. So somebody has to download the plugin? No. There's no need of downloading. Just I can show you the demo. So like automatically the PDF streams and you can see the content. And this like it doesn't have to require like even in legacy browsers it supports. User interface and we'd like to show the demo. So sir this is our user interface looks like. So there are the first hierarchy as you told you. Departments, then courses, then lessons and inside lessons we submit the item. So currently we have two departments that we have made. Computer science and mechanical lessons inside course also. So. Suppose I just want to add my open educational resource to your department. I'm not a course. So you can submit that in a department. There is the provision that without. I'm not a course. Sir there will be a like right now it's not there. There will be a like something like others or something where you can upload everything that's not related to that. I don't understand when you're talking about open educational resource. Repository platform. Yes sir. Where does the department and course come into it? Sir we were basing it more on college. On what? College level. For what? Sir like. There is OER repository by government of India. Yes sir. Okay. It has no department. It has no this thing. It is topic. Yes sir. What we did. Open educational resource. Who the hell cares when it's done by IIT Bombay or IIT Madras? Sir what. Why the hell do you care? Sir but the department what we meant here is not the any government. It's like your content might be related to some mechanical department or computer science department. So it's a topic. Yeah it's a topic. Yeah it's a topic. Why course? Sir course like. Sir we can call it module sir. It is like a module. Why module? Sir. You don't understand. Sir. Because if you don't need to. I want to write something or pointers. Yes sir. I want an educational resource. Sir there is. Where? How the hell do I put it? Yes sir. There is two things. You don't need to create a course. You can just have a lesson. Lesson as in you can just say pointers and you can upload your content. You don't need to create a course for that. That's what I am asking. Yes sir. So that is also there. Like in the computer science department. Yeah. If you have a standalone lesson. I am not a department. Sir the course is related to that department. Sir not inside department. Inside the topic. Inside the topic. No. No. What about cross disciplinary? Come on. I mean I don't understand why you are rigid. If you are coming here to open the repository system this is rigid. So you. Wanted to have. Heat engines in aircraft. Where will I put it? This is the best practice. Correct. You could have the labels. User defined. Labels. User. Okay. Exactly. Yes sir. Yes. Formation technology department. Okay. Can I put it under two departments? Same course. Under department. I am talking about resource here. I am not going to talk about anything. I have got an educational resource. Yes sir. Can I put it under three departments? That is what it is. One is to be. I don't understand anything else. There is an educational resource. All the structure you are building is supposed to tag that resource. It doesn't have to belong to anyone. Belonging is my problem. Okay. I don't mind tagging it as anything. Okay. So if we make that user defined. Interdisciplinary is common. Yes sir. No. It is going to become more common. What? Like there is this course on data structures and a lesson on linked list. So here this is a video which can be streamed. We have added those plugins for image, video and audio as well. And apart from this that search. Understood. So if you want to search, show advanced filters. If I want to search all the videos that I have uploaded on this portal. So I will give the type video. No. Keep it video. I want to see if you are smart enough to add I. Are you? I don't think so. It will. No. He is not smart. No sir. I asked him he is very smart. So we got this search is a string. Substance of all the tagging because you see it contains. Yeah. Otherwise it will drop down. These are the ones you can use. So there is one more thing. Let's say sir PDF content or docs page. And like it can. No. All this is given by a d space. Correct? You have not done anything. No sir. Okay sir. What we have done is let's say like what was not there by default in d space. Instead of title I can drop something else. No sir. Like we can search in the content. Like if I need to search a word and if it's in present in any of the PDF documents. So that's searching the words will give me the occurrences in all the PDF documents. That you added. It gives the occurrences like if I am searching link list. And I have a PDF document with the lesson of link list. There is one question now. What happened basically in such kind of systems when you search. And you will go inside the PDF and pull it out. There are two questions. Two issues which comes many time. First is that whether I should have a right to see that. For example if I search Mr. Outage balance in the bank. If it is available in his repository. But whether I have got the right to see that. But your search will pull it and show it to me whether it is taken care first. Second thing is that many time the jurisdiction is not defined. Then it will bring so many search results like Google that I cannot get the appropriate thing what I want. So how you have control both the issues. Ranking. Yeah. Sir the ranking thing is already supported by the solar. Like it ranks each of the documents according to the frequency of the words you are searching. Like relevancy. Like the word you have searched it has all the text relevance. Like it calculates the probability of that letter and creates a ranking of that items. So it gives a more appropriate search results. What about the rights of the user. So it's open source. So like that at default we are allowing every. Open source also there is a right. If he is admin he has got separate right. I am a user. I have got separate right. He is a teacher he has got a separate right. Yes sir. Like I am a normal user right now sir. Like what I can do is. No no I am asking the rights for search. Rights for search. Yes if I am civil engineering student that I have got only rights to search civil also or I put in civil engineering search and then I will get computer also and mechanical also and everything which are irrelevant. Sir you can search in that particular like there will be this. No give me a specific answer don't play with that. Yes sir. Give the answer. Here you can search a particular department like right now. Whatever answer. That also I have to give. It should take from my login from which department I belong. So but if the civil engineer also wants to search computer then. Then I specifically mention but I should have the rights. See what I am saying that I cannot see his financial things. Yes. The reason is that I am not empowered. Forget about the general thing. I am not worried about that. But which I didn't have rights whether your search will dig it out. Normally I didn't have the navigational right to go there. But search will many time pull it out. Sir but if you upload on the open source repository it will be available. Yes sir it is an open source repository. See the purpose of the project is to search any kind of. So you can answer like there is no rights provision. Yes sir. It's okay. Answer another basic question. Yes sir. You have this space. I am asking for open education resource. Yes sir. What is special apart from adding five tabs. Special as in sir that was not there before. Apart from adding five tabs. That was not there before. Apart from adding five tabs. Yes sir. And giving this artificial boundary of department and course and whatever. Actually forgotten that's an educational resource. The educational resource is a separate resource by itself. Exists by itself. And it has got a topic which is to be used for searching. So that plain thing has been lost somewhere. And I think that is the most serious flaw you are done. Because when I come in I am not interested in anything. I am interested in learning about a topic. So what are the resources in that topic? Learning about a keyword. What are the resources in that topic? And I expect you to figure out that a topic belongs to computer science. Even they say that department is immaterial for user. So I think you have lost that. But because of the metadata schema we can uniquely identify each and every item belonging to a particular department. So there is no. But I think your requirements are wrong. Your requirements should be how to classify large amounts of open educational resource. That was the second question I was asking. How do I know that what you have done is scalable. That's fantastic word of it. So the display system is highly scalable. Because across the globe like thousands of systems are using d-spells alone for storing all these digital assets. And they are mostly digital assets are like around in TVs. So like it has a capability to be scaled. Like we didn't get the resource to scale. Scalable in terms of response time. Yeah. For search. For search if you are asking like the solar search we are using here is like almost efficient. Yes. For some reason this d-space animal. Okay. Things that only five writers is enough. Okay. So according to me d-space itself thinks that it is not scalable beyond five. Why can't I take it like that because you are not tested. So d-space is. So d-space was not like it was not constructed like keeping in mind that it will be meant. It was not meant for educational alone. Like it was as a repository for anything. Like across institutes it is like customized according to their requirements. Sir like sir I am getting your concern sir. Why five to ten was my first question. If it is educational. Yes. Okay. It has no other data except educational. So it's a d-space instance only for education. Why can't I. So d-space is not only for educational. It is not but you are configured it for education. Yeah. So. If you want to search like a particular set for. I don't want to get biography. It is not educational for me. Why are you putting it there. d-space is meant for something else. If it is meant for education then all I need to do probably in my view is to take those like you are done with a community is equal to department. This is equal to class. Similarly to the I will change title. Title is equal to department. Why I need title. I suppose corona. Why extending five to ten is a risky proposition. Yes sir. I don't know. Anyway. Yes sir. Okay go. Okay sir. Thank you sir. Thank you sir.