 Good afternoon. Today, I am going to give a talk on model for rapid large scale development of learning objects in multiple domains. I will start off my talk with what are learning objects. There are various definitions available in literature as to what learning objects are. For the purpose of our research and for our Oscar repository, we have adopted Lallier's definition since it has education as its purpose. According to that definition, a learning object is the smallest independent structural experience that contends an objective, a learning activity and an assessment. On the screen, you see a sample screen shot of an Oscar alone. The home page containing the list of learning objectives constitutes the objective part. The learning activity part comprises of the interactive animation part, the glossary part which contains definitions of keywords coming in the animation and the references section that contains references to books, research papers or URL which the user can use to study the concept further. The assessment part is comprised of a set of five multiple choice questions for self-assessment of the learner. Why learning objects? Learning objects have become valuable educational resource in recent times because they can fit into different instructional settings. Be they be classroom teaching or distance education programs or self-study purposes. They are ideal for spreading higher education to different parts of India. Consequently, the number of learning object repositories have increased in number. On the screen, you see a screen shot of an Oscar learning object repository. The goals of Project Oscar? Project Oscar aims to create a repository of open source courseware of learning objects in various domains like bioscience and civil engineering. The education level targeted is the tertiary or the college level and the production rates stipulated for us is producing 100 LOs in a year. Project Oscar is a part of enemy ICT initiative that is National Mission on Education through ICT funded by the Ministry of Human Resources and Development Government of India. How are these LOs produced? There is a DERP in literature containing detailed documentation of the production process. The three models which we have studied are TIL, MEDOVA and FET because they have well documented processes. All these adopt the synchronous model. In this synchronous model, there are four types of team members, the subject matter expert, SME, the pedagogy expert, the interface designer and the coders or animators. The entire team is located in one place and there is extensive face to face communication between the team members. The result of these synchronous models are good quality LOs but produced in restricted domains like physics and physical chemistry. The production rate is also around 100 LOs produced in approximately 4 to 5 years. When we adopted this synchronous model and tried to scale up along numbers, the problems that we faced was while working simultaneously in multiple domains, the team size increased. Because the team size increased, it led to space and infrastructure constraints. To solve this constraint, the team had to become geographically dispersed. This geographical dispersion hampered the face to face communication between the team members. When we tried to scale long domains, finding faculty in each domain who would be willing to devote this much time to the labour intensive task of producing a LO proved difficult. Also, the developers not having domain knowledge aggravated this distance in communication problem. The conclusion that we arrived at was the synchronous model will not work under the following conditions. A, if we are trying to produce LOs in parallel in multiple domains, B, when there is a time limit to the production of learning objects and C, when a large development team is involved. The broad research questions facing us at this point was what model do we adopt to meet our objectives stated earlier. An offshoot of the question number one was can an asynchronous distributed development model replicate the efficiency of a synchronous model. To answer the second question, we came up with this production framework. The LO production cycle comprises of four stages. The stage one comprises of the concept proposal stage in which a domain expert proposes a concept. Project Oscar does a check to see if a similar learning object exists in the web. If it does not exist, the concept proposal is approved and it passes on to stage two. Here the proposal is elaborated into an instructional design document, the IDD which is approved by an IITB faculty. Then it passes on to the third stage which is the design stage that is carried out by Project Oscar. And the last stage is the development stage where the developer in consultation with the SME creates the final LO. And on approval of the SME, this LO is uploaded in the repository. We tried out various asynchronous models based on this methodology. The names of the four models are given on the screen. The variable that we changed for each of the four models was the person who is responsible for carrying out a certain task. In the online distributed model, in stage one the concept is being proposed by a member of the online community. This was the advantage of this model. The IDD would be approved by the IITB faculty and this is where we face the biggest difficulty. There was a problem with the standard of the content and each IDD from this model would go for a minimum of four to five rounds of review by which time the IDD writer would lose interest and stop responding. The lesson that we learned from this model is we can develop an LO from this model only if the developer had themselves domain knowledge. We moved on to the individual faculty model. In this case, the concept is being proposed by the IITB faculty and the IDD is being written by their students. Generally a faculty proposes a couple of concepts. The advantage of this model is the content quality was ensured to a great extent. But the main problem we faced was there was a gap in communication between the SME and the developer. The SME assumed a lot of things as common knowledge which the developer not having domain knowledge was not able to assume. So in this model, the lesson we learned was there is a need to template the instructions coming to the developer. The next model was the fully outsourced model. In this case, the concept is being proposed by the IITB faculty as before but the IDD is being written by the vendor SME. The development is also being done by the vendor SME. In the previous model, the development was in-house by Oscar animators. The advantage of this model was the production cost was low but the overall development time that is ALO plus the IDD development time increased substantially since it took the vendor to get approved one IDD one month. So the lesson we learned from this model was the content writing should not be outsourced. It should be restricted to the educational institute. The last, the most successful model is the domain owner model. In this case, the concept is being proposed by the IITB faculty but the faculty is now a committed stakeholder in project Oscar. So they are responsible for making a minimum of 20 learning objects of which a minimum of 10 should be mapped to a course. The IDD writing is being done by the faculty student. In this case, we had made it mandatory for the student to attend a two-day instructional design workshop to introduce them to pedagogy principles. The stage one and stage two content generation phase is taking place in synchronous mode. After the approved IDD comes to Oscar, the designing takes place and the development is being outsourced to a vendor. So, the advantage of this model is the review time for both the IDD and the learning object reduced and instructional designing training was being given to the IDD writers. The limitation of this model is there is a time lag between the content generation stage and the starting of the development stage because you need a substantial number of IDDs in your hands to do the outsourcing. The detailed process of this model is given in the right hand bottom corner. If we come to the comparative analysis table of the four models, we see that the domain owner model has emerged as the most successful model. Here the mean production time that is the total time for the IDD and the animation production paid to 0.375 months. Though the cost increased compared to the fully outsourced model, but to make a tradeoff between the mean cost and the mean total time taken for production, the domain owner model comes closest to the ideal case. Here we have taken the ideal case to mean as the mean cost, minimum cost taken among all the four models and the minimum mean cost time taken across all the four models. To conclude the talk, the domain owner is the most successful model because it seeks to harness the expertise of both the educational institute and the expertise of developing animations of the e-learning industry. Also, the social impact of this model is enormous. Not only are these learning objects great vehicles to promote higher level education to different parts of India. Also, students and research scholars who are writing the IDDs are becoming aware of pedagogical issues right at this stage of their careers. With this, I conclude my talk. Thank you.