Oral-History:Ming Lin

About Ming Lin

Born in Taiwan, Ming Lin and her family immigrated to the United States in 1980, settling in California. She graduated from Sunny Hill High School and earned her B.S., M.S., Ph.D. degrees in Electrical Engineering and Computer Science from the University of California, Berkeley.

After spending twenty years at UNC Chapel Hill and holding the John R. & Louise S. Parker Distinguished Professor of Computer Science, Lin joined the Department of Computer Science at the University of Maryland, College Park as the Elizabeth Stevinson Iribe Chair of Computer Science. Lin has authored or co-authored more than 250 papers in refereed technical and scientific publications and has edited, co-edited, or authored four books. She is known for her work on collision detection as well as the Lin-Canny algorithm, a method to execute fast collision detection in dynamic simulation. Her research interests include computer graphics, robotics, and human-computer interaction, with focuses on physically based modeling, sound rendering, haptics, algorithmic robotics, virtual environments, interactive techniques, geometric computing, and distributed interactive simulation.

Lin is an IEEE Fellow (2012) and a Fellow of both ACM and Eurographics. In addition, she has received many honors and awards, including the NSF Young Faculty Career Award in 1995, the Honda Research Initiation Award in 1997, and the IEEE VGTC VR Technical Achievement Award in 2010.

Lin is active in IEEE and other professional organizations, and has served on boards and committees and held various editing posts at many journals and magazines. For example, in 2015, she served as a member of the IEEE Computer Society (CS) Board of Governors, a member of Computing Research Association-Women (CRA-W) Board of Directors, the Chair of the 2015 IEEE CS Transactions Operations Committee, and a member of 2015 Executive Committee of IEEE CS Publications Board. She also served on the Steering/Executive Committees of IEEE Virtual Reality Conference (Chair, 2010-2011) as well as the ACM SIGGRAPH/Eurographics Symposium on Computer Animation; the Executive Committee of IEEE Computer Society Technical Committee on Visualization and Graphics; and several other technical advisory boards for U.S. government agencies, industry, and the international scientific research community. For three years, she volunteered as Editor-in-Chief of IEEE Transactions on Visualization and Computer Graphics (TVCG), 2011-2014. In addition, she served as associate editor and guest editor of many other journals and magazines, including the International Journal on Computational Geometry and Applications; Proceedings of IEEE; IEEE TVCG; IEEE Computer Graphics and Applications; IEEE Transactions on Haptics; and IEEE Computer Animation and Virtual Worlds.

About the Interview

MING LIN: An Interview Conducted by Peter Asaro, IEEE History Center, 23 January 2015

Interview #807 for Indiana University and the IEEE History Center, The Institute of Electrical and Electronics Engineers, Inc.

Copyright Statement

This manuscript is being made available for research purposes only. All literary rights in the manuscript, including the right to publish, are reserved to Indiana University and to the IEEE History Center. No part of the manuscript may be quoted for publication without the written permission of the Director of IEEE History Center.

Request for permission to quote for publication should be addressed to the IEEE History Center Oral History Program, IEEE History Center, 445 Hoes Lane, Piscataway, NJ 08854 USA or ieee-history@ieee.org. It should include identification of the specific passages to be quoted, anticipated use of the passages, and identification of the user. Inquiries concerning the original video recording should be sent to Professor Selma Šabanović, selmas@indiana.edu.

It is recommended that this oral history be cited as follows:

Ming Lin, an oral history conducted in 2015 by Peter Asaro, Indiana University, Bloomington Indiana, for Indiana University and the IEEE.

Interview

Interviewee: Ming Lin

Interviewer: Peter Asaro

Date: January 23rd, 2015

Location: College Station, TX

Education, Computer Science, and Robotics

Asaro:

So if you could just start by introducing yourself and telling us your name and school?

Lin:

Right. My name is Ming, Ming Lin. I was born in Taiwan. My family and I immigrated here to California in the 1980. I went to high school here in the US, Sunny Hill (sic) High School in California. And then I went on to pursue my undergraduate degree. I spent two years at UCLA. That's close to home because I’m actually a Southern Cal girl. And then I transferred for my last two years to UC-Berkeley, so I eventually got my degree in electrical engineering and computer science at UC-Berkeley for my bachelor degree. I have always-- don't ask me how I know, but I always know that I want to pursue a PhD, and I think that was in large part why I have requested a transfer from UCLA to UC-Berkeley. Even requesting transfer was not easy. So I think I wanted to explore if I liked Berkeley, and I was really happy they accepted me, and so I went to Berkeley. I think Berkeley just always had the programs that I-- you know, they have a lot of faculty in my area of interest, and I was very happy to-- for the opportunity to transfer there and then to continue my study there. During my two years of undergraduate study, I really liked Berkeley, and Berkeley is such a large campus that they just so vibrant, they have so much to offer that I felt-- when I was looking at grad school, I applied to five places, the top five in the country, and I got accepted everywhere. And Boston was too cold-- sorry, Stanford. You were a competitor to Cal. Couldn't go to Stanford, so I ended up going to Berkeley. Yeah, but I mean there were really-- it was a hard decision. I really liked it there. So I actually end up getting my master and also my PhD all from Berkeley.

Asaro:

How did you decide to go into computer science?

Lin:

Oh, gosh, I think-- I mean, actually, I'll be honest. I don't think I know that I want to do computer science. All I know as an undergraduate student-- I mean, actually, now I'm a faculty, right? I look at all these incoming graduate students. They already know what they want to do. It just amazes me. As an undergraduate student, I just know I was really, really good at math. I enjoy physics. I do well in all my science classes. I was the top in all my science classes, but I did not like the biology and the chemistry as much as I love physics. And so I knew that I want to do something that was related to engineering, something that would have the mathematical component, something that involved logic. So I-- when I went to UCLA-- UCLA was pretty close to home. It's a-- it has really good engineering program, and it has a lot of other things like the liberal art educations. I was chatting about yesterday over dinner table that I really enjoyed it, that respect. So I would say it was a great first two years because I get some of the really, really foundational kind of well-rounded college educations in a humanity side while I was at UCLA. But I also explore my interest in engineering, so I actually started by taking very general engineering courses. I had the engineering courses. I had the computer science courses. I couldn't decide actually whether I want to do really EE or CS. So EE/CS as a major at Berkeley was a huge attraction because they allow me to continue not making any decisions, right? So I just continue, went on to EE/CS, and I continued to take courses in both computer science and EE, and it allowed me to have that blended perspective of what's going on. And undergraduate, it's-- to me, it's like a breadth education, a general education. And since I know I was going to pursue a PhD even when I was at UCLA during my first two years, that not having to make that decision regarding specialization into a graduate school was a really nice option, you know? And so I continued to take a lot of math courses over the summer. I work, also, and so it was just-- because I lived close to home. So to be able to have that flexibility during my first two year was great, and then be able to continue to have the option when I was at Berkeley was even better. So I don't think I really focus on my area of research and concentration until graduate school.

Asaro:

And how did you get interested in robotics?

Lin:

It was one of those things that you take classes, and the robotics was offer at a graduate school level. I don't think it was offer at undergraduate level. And I remember I took that course from Ron Fearing, Professor Ron Fearing at Berkeley, and it was just really, really interesting. And I'd like what I study there. I thought it was a really nice combination of everything I enjoy. The math, the engineering, the hardware, the software all kind of fall into one place, and I think that was one of the reason I enjoy that class a lot. So I just decided that-- and I had a fellowship my first year, so I have a lot of choice to kind of shop around. And so I shop around regarding what I want to do, and I decided that robotics was what I wanted.

Asaro:

And so who did you end up working with, and what was your thesis project?

Lin:

So I end up working with Professor John Canny, who was just hire not too long ago. He was an ACM Visitation Award Winner from MIT and he was a just super nice, super great guy, smart guy. He was one of the person I really, really have worked with and really, really have tremendous respect for, and I still do. I mean it just-- he is quiet, he's reserved, but he have great idea, and so it was just a pleasure and a joy, and he's just a tremendous-- tremendously nice person. So I actually-- I look around, and I thought, well, you know, the stuff that he does seems really cool, and I want to be able to be part of that, that research program, and luckily, he's saying, "Sure." You know, he was new faculty looking for student, so I was just very lucky in that regard that he took me on as his advisee. And I started working on robot motion planning. That was one of his thesis area. I thought it was very interesting how I started working because when I talked to him, he gaive me his thesis, and it was really, really hard reading. I remember looking through and spending summer reading that book, his thesis, which was turned into a book, and it was so hard because there's so many terminology I have never ever seen before. And I consider myself, taken-- for someone of my level, I have taken a lot of math course. I have topology, I have complex analysis, I have a lot of other things.

So for students in my major, I have a lot of math, and that book has so much more math than I have ever know, and I remember marking his book and writing down those page number, "What was this?" And I went back-- when he came back from his summer vacation, I sat down and I chat with him, trying to ask him questions, and he just smile at me like I ask all these question. Lo and behold, it was kind of intimidating. Now, when you read a book like that and then you say, "God, how am I going to get to this level in five years or four years, right?" And there's just so much I did not know. And later on I found out from other people, and they say, "Oh, there are probably only about five to 10 people who understand John's thesis in the whole world." So I felt a lot better that I did not know as a graduate student who are starting in my research career at that time, but I kind of knew why he was smiling later on because he's probably like, "Okay, you passed the test," or something. So it was fun. I started reading a lot of paper, related to geometry, motion planning. So we started working on problems.

John came from the MIT AI labs, was a LISP guy. Had to learn a new programming language, but program language is one of those things that you just-- once you know the concept, just pick up. So I started working the area. It was fun, but then I-- I liked robot motion planning, but I also found it a little bit more abstract. And as part of implementing the robot motion planer, I need to develop a collision detections library for it. And it turned out that the best-known technique out there are all quadratic in terms of the complexity of the object. And if you need to do this every single time stepped, having to deal with a quadratic complexity, it just not going to be-- the result is not going to be anything that's going to be close to really usable. So I end up diverse-- I mean I end up diverging to a different direction. I started spending lots of time working on the collision detection problem, working on the distance computation as the object move, as the robot moves in environment. And I end up publishing multiple papers, which was already at that time getting pretty good traction and attentions. I have people calling me, sending me e-mail, inquiring about the paper I publish. So lo and behold, like I think my third or fourth year, I switch my direction. I had a conversation with John, and this is where I have to say John was just fabulous in that regard. Imagine that you're really passionate about this direction that you want to go, which was motion planning, and it's not a "I don't like it," but I have seen myself making a lot more contribution on a collision and distance computation problem, which has tremendous utility, not just in motion planning but in dynamic simulation, in animations, in manipulations, in collision avoidance. And so I see I am charting <inaudible> a little area for myself, and I sat down with him and we had a talk, and I said, "John, I really would like to switch directions to a very closely related topic. It's going to be collision detection and distance computation-- what do you think?-- for my thesis, and it's still that the motion planning will now become secondary to my thesis. It will become an applications of collision detections and distance computation algorithm, which I have already developed." And he's like, "Okay." And I'm like, "Thank you!" So that was-- at that moment, I see like the end of a tunnel. I can see myself finishing up because I-- so I was really happy, and I continued to work on that area, and then I finish in the expected time, like five years or so.

Asaro:

And when was that?

Lin:

Around '93.

Post- Graduate School, University of North Carolina and the Army Research Office

Asaro:

And where did you go after you finished your PhD?

Lin:

So I had-- very, very interesting. When I started my PhD, I always thought that I wanted to go to a research lab. Academia was never ever in my considerations. And what happened was I end up interviewing for a few industry job, and I also had a summer internship at Digital System Research Center, and DEC SRC-- DEC, by the way, no longer exists today, but DEC SRC was this very well-known research center located in Palo Alto, where there are a lot of famous people when I was a graduate student. I was just going to go, "Wow." When I went there and worked there, I just-- looking at all these extremely famous people, and then I also know they were losing their job. And it started to sort of dawn on me like there were just a lot of instability in the industry. You want to do research, and it's a very practical question where it comes to you. If you want to do research and you're worried about tomorrow, it's unclear, like are you going to be able to continue-- just imagine you started working on a really exciting project and then somebody gonna come and say, "Pull the plug." They say, "I’m sorry. You can't work on this anymore." So it was sort of a very-- it was sort of kind of changing moment that kind of make me think maybe industry life, it's not so much cut out to be the kind of places that I might like to be. So I had almost a very last-minute switch and say, "Okay, well, let me try academia," and it was very late in the season. Robotics was not terribly popular at that time because there was so much hype regarding AI in the eighties, and some of that hype did not quite pan out in the eighties.

So the early ninety was not exactly the best time to be a robotic researcher. So at that time, I seen some of my friends having to postdoc, and it was also during my last year. I said, "Well, I do a lot of these simulations. I do a lot of collision avoidance. Why can I not apply this to things that are very closely related?" So I apply some of the work I have done to computer animations. I published some paper in those area. So when I was looking for jobs, I also try out some of the faculty positions, largely in California. So I limited very much my search to California, and I have a couple offer-- I have a few interview. I end up taking the one in one of the really nice places. That was Monterey. It's beautiful, and it's in California, five hours' a drive from home. I'm very close to my family even though right now I mean the place is as far as I could possibly get from my family. My family's still living in California. So I took the position. I went to Naval Postgraduate School, and it was great. It was hard but different because most of the people don't know Naval Postgraduate School is actually a graduate school only for predominantly navy and marine. They also have other DOD personnel who goes there to pursue their graduate degrees. So it was very much research focus. So it was great because I get to-- at least at that time, I thought that was great. I get to do my research, focus on graduate educations, and that's where I end up for a year because I had a two-body [ph?] problem. So the very next year, a long-time very, very good friend of mine, who was also the office mate, and I, we had our relationship change, and we decided to get marry. He was already at University of North Carolina Chapel Hill, and so I move out. Decided to take visiting positions at another university within the UNC systems, and I was a visiting positions, so I came to North Carolina-- or I went to North Carolina.

Asaro:

Which school was this?

Lin:

That was NCNT, and that was also a very interesting experience because that was a visiting positions, which, by the time I arrived, they changed it to a tenure track positions. And it's an HBCU. I don't know if you know why it is. So I had a really interesting perspective to see what it is like to actually work in minority students, what their issues are, and that was a really, really good experience for me just to be able to see all the different perspective that I think a lot of people from the typical university don't see. So I would say my career has-- that was only my second year after I graduate from my PhD. I already changed job twice, and because it's very different and to me was always a temporary visiting position. I have started looking for another position when I was there, and I look around, and given my two-body problem, my constraint is very much to the local area. So a friend of mine was at the US Army Research Office in the Research Triangle Park, which, by the way, was closer to Chapel Hill than NCNT.

NCNT was an hour away, so I have to drive and commute one hour every day each way, which was two hours, and often when I come back, it's in the dark. So I didn't really enjoy the driving. Even though that they have turned it into a permanent position, I felt the commute was too much, and there wasn't very much option living in between. So I've been looking around for something that will be closer to Chapel Hill. And so a friend of mine was at US Army Research Office, and she's like, "Ming, we are looking for a founding program manager in computer science, and we're just starting this program.

Why don't you apply?" and I was like, "Okay. It's going to be a huge step away for me because US Army Research Office is not exactly a research lab. It is actually a funding agency just like NSF except it's a mission-targeted funding agency, and it's for Army. So I interview for the job, and I got a job, and I moved to Army Research Office, which just abbreviated as ARO. It was my third job, and I barely graduated for two years. It was great. I went there, and I will say if you look at life as experience, it was another very interesting experience. So I actually work at a research-- I work at ARO for about two year-and-half. What prompted me to leave ARO was the uncertainty. They are constantly talking about relocations because Army Research Office is the only funding agencies among all the other agencies, including NSF, including DARPA, including ONR, including AFOSR that is actually outside of DC, and there has always been a desire to relocate ARO back to DC. And that was a constant chattering when I was there. And, you know, it's that fear of having to be told that we have to move again, and I was just thinking, you know, maybe I should preempt the move and think about where I want to be. And so while I was at ARO, ARO had this really, really nice program, which is called a Staff Research Program. And you get to spend one day a week at one of the affiliated University and to be able to continue to do research. So I was actually an adjunct with UNC Chapel Hill already. And right-- also right before I left NCNT-- when I was at NCNT, I actually apply for the NSF Young Faculty Career Award, and I was one of the three lucky one out of some 30 people or so that got an award. And so before I joined ARO, I told ARO, I said, "Look, I already got this grant, which is very prestige, and if I leave my academic position, I have to give up."

So ARO said, "Well, why don't you apply for a staff research grant, which is not going to be comparable in size compared to what you have, but at least it will allow you time and funding for a student to support your student." So I was actually to also get-- I apply for the grant, I got the grant, so I also have funding to support students while I was at UNC as an adjunct. So it worked out really great. So I actually continued to do my research, and I would say, you know, some of the work that I have done that was sort of a follow-on to my PhD dissertation and that turn out to be some work that had been very well cited and actually have been transitioned to industry were the work done during that time. So around '97, when I decided to leave ARO, I already had built up a portfolio. We look around because this time my husband has to be tracking, and we started looking around. We have actually-- this is the time when I want to come back to robotics and do more work, and we look around, and we had a few really good options.

Actually, everywhere we interview, we got the job. The hard one came down to the decision between Michigan Ann Arbor and UNC, and being a California girl, even though I really, really, really like what Michigan has to offer, we end up staying at UNC because I thought I would minimize the move. And that's a really-- I think that was a major decisions, I think, in many way for our career. I would say that the focuses are very different, I think, between the two department. Had I gone to Michigan, I will say I would do far more robotics than I had been at UNC. So being at UNC, UNC has a lot more virtual reality, a lot more graphics program. It's considered like the top-- the top program in those area in the nineties, all the way up into 2000s, into 2000s. So I still maintain my connections. I still work on motion planning. I still work on collision detections. I still publish a lot of papers in robotics, particularly the more algorithmic side. So I continue to work on those area, but I will say I probably would have done substantially more had I move on to Michigan instead of being at UNC because Michigan has very, very large presence in robotics at that time.

Robotics at UNC

Asaro:

So who else was working on robotics at UNC when you got there?

Lin:

It was just me and my other half, yeah. And my other half was also John Canny's student.

Asaro:

What's his name?

Lin:

Dinesh Manocha.

Asaro:

What areas was he working in?

Lin:

Well, he was a grad student there. He actually was working on pretty theoretical stuff. He was working on computer algebra, and he had application of his thesis on computer algebra to inverse kinematics. So he was working on essentially a computational solver, a real-time solver for inverse kinematics using a computational algebra technique.

Asaro:

And you worked for your PhD and then at the postgraduate school, and you-- NCA and TEN [ph?]NCNT and to UNC. Were you mostly working on the theoretical sides, or were you really working with robots or--?

Lin:

No, I was working mostly with simulations. Yeah, yeah. I was working very much mostly-- that's one of the reason I say I was working more on the algorithmic side. So it was very much a lot of simulations and mostly on the geometric algorithm, motion planning, collision detections, and a lot of our algorithm actually we also--- So during that time when I was an adjunct at UNC and later on, we put-- we probably were the first group. So I was saying that Dinesh actually move on to robotics and motion planning partly because of me 'cause I mean it just-- it's great to have a collaborator, and so he started looking into motion planning problem. He started looking into collision detection problem, and we had a lot of-- and he certainly had the expertise. I mean there's no question, yeah. Dinesh was very, very strong analytically. I'm not saying that just because he's my husband; I'm saying that as just being fair. And he's a very quick learner, so he really pick up a lot of these things very quickly. So we had a group that we had work on, so I would say we, as our group, was probably one of the first group, if not the first group, to start to release our code on the Web so other people can download it and can use it, and we have a lot of downloads all over the country, in fact, all over the world, that I will say a lot of people in robot motion planning had-- who were active in research during that time have used our library or have downloaded our library or their students have downloaded.

Asaro:

What was the library called?

Lin:

We have multiple one, and the first one was called the I-COLLIDE, and then after that, we have released it a library-- I, as in interactive. Who would have known I'd become like iPad? I will say we were before the Apple, yeah. We had a name I- COLLIDE because it stands for Interactive Collision Detections, COLLIDE as in collide, and it was a collision detection library. RAPID was another library. And then we went back to the COLLIDE naming system. We call it V-COLLIDE because that was compatible with VRML. So those were the three major library that we have release in the nineties. And around late ninety, early 2000, we release another library which we call a PQP. And each one of these actually all stand for something, right? PQP was Proximity Query Package. RAPID-- I don't remember what RAPID stands for because it was so long, but something that was related to very fast, rapid, accelerated proximity interference couriers or something. I can't quite remember what it stands for. But they were acronym to stand for some sort of library name. So that was some of the library that we have release, and amazingly, I will say probably over 50 company around the world-- most of them are very big-- that have actually incorporated our library into their software systems. I will say probably some of these companies-- I can't name them because confidential informations-- are very, very well known. Some of them-- one of them was-- total surprise to me-- was related to medical imaging. They use it for collision avoidance for medical imaging devices. And we have CAD/CAM companies, automobile company, aircraft design companies. I mean just all kind of company, robotics company, that have used the software. So we continue to work in the areas still today.

Asaro:

And did you consider creating a start-up or anything like that around your software?

Lin:

Not really. I mean we thought about it, but we didn't feel that-- we feel like there was some really-- and we have game developer, also, license that some of them just probably use them for free. Our interest was not money making; it was just trying to do something which is impactful and useful because that's what the fun is, right? I mean to be able to do something-- someone actually can use it and can make a difference. I mean that was always the goal, and so being in academics and sort of kind of not really think about the potential of what that revenue will be, we didn't really-- I mean we thought a little bit about it, but we didn't take the jump. Let me put it that way. Yeah, so we thought about it; we didn't take the jump.

Haptics and Virtual Reality Conference

Asaro:

And your work is-- we came into the 2000s. Did you stick in the realm of virtual reality and simulation, or did you move it more towards robotics?

Lin:

So when I returned from Army Research Office back to academia, even though I was working one day a week at UNC, it's very different kind of work being at a Army Research Office. So there was a transition moment. It was sort of kind of a very painful transition and moment for me because it's like I have to get back to writing proposal again. I have to think about-- and I remember my first proposal I got good review but was not good enough to get funded. I had a conversation with a program manager who funded me in 1994 when I got my career, and he give me some really good advice, and I thought about it, and I say, you know what? What I did, what I wrote as a proposal three, four years ago was good at that time because I was a young, fresh PhD. But no one was considering me as a young, fresh PhD at that time anymore because I'm four-year postgraduate. I'm supposed to have a new research directions, and I can't work on the problem which I'm really good at. I need to shine some new light, think about a larger picture of where I want to go. So I took that into consideration. Just right around the time, SensAble Technology contacted me because of my work on collision detections, and SensAble Technology was a spin-off company from MIT developing this device called Phaantom, which is a haptic device. And most of the people don't even know what haptic device is. And when they contacted me, they say, "You know, one of the things that we have to do because we have to create this sense of virtual touch, so what we need to do is to run the simulation in real-time, we get the feedback"-- and what a haptic device essentially is is essentially a robot arm but use in the reverse. You guys, yeah, both of you probably didn't hear our talk, so Allison Okamura give a really very excellent lecture in the symposium yesterday and talk about haptics.

It's really engineering the sense of virtual touch. And the way that this particular device works is essentially like a robot arm used in reverse, where the users will use this robot arm to kind of track its positions and orientations, and that will be-- so the virtual proxy of the user's hand in this virtual environment. And the user can grasp or pick up a virtual object, which really-- it's only have its digital representation. It doesn't really have a physical metaphor. I mean it doesn't have a physical presence. And so one of the things that you want to do is you want to be able to feel the object that was being manipulated through this hap device, and in order to do that, the device need to generate forces to push your hand back so that you feel like you are feeling the surface's object and interacting with object. And you have to do all this in 1,000 hertz, which is a extreme computational challenge. Even though we were like the first group that was able to do collision detection interactively, 1,000 hertz, that's like two order of magnitude performance improvement that we need to have, and so that's why they contacted me. And I met Ken Salisbury during that trip. They came down to see me and UNC and wanted to see what we have. We had some discussion. He got me interested, and Elaine Chen, which was a VP of Research that came with him, and we had a really great chat. And I started-- and I was very honest. I say, "I'm so sorry. I don't have a solution for you because what we have is interactive collision detection. It's very fast, but it's at least one order of magnitude off."

But it got me thinking about what can we do differently, and so I-- that actually-- that visit, it was pivotal in a way that it got me interested in haptics. It got me think about what can we do, how can we extend the algorithm, how can we design new algorithm to support haptic rendering, force feedback, and so I started working in haptics. I actually work in the area for a good, I would say, seven years, and we designed some of the, I would say, the quickest, the fastest algorithm in the area for haptic rendering. I had a really, really great PhD student, Miguel Otaduy, who is now, I will say, a young rising superstar in Spain. He has been faculty there for a while. He's reasonably well established. He got some EU superstar- equivalent awards for junior faculty, so he's doing quite well. So he was the one that I have worked very closely in that particular area. We have couple books that we have published together. One of them was Joohi Lee with Multiple [ph?] Researcher, who I consider giants in their field in the haptic domain, and Allison [ph?] was actually one of the contributors. So it was great. It was a great seven-year, eight- years' run, and that's related to robotics because it's extension of robotics used in a very different context. I will say haptic, I think, originally probably started from the tele-operations because you can manipulate an object and you can tele-operate another haptic device remotely. There's collaborative haptics, so they're kind of very similar to tele operations. So it was really, really-- it was really, really a interesting period. And also through haptics, I also got more into virtual reality. Even though just because I was there, if I don't have something that I feel like I’m part of and contributing to, it wouldn't matter if we have one of the best virtual reality group or cluster in the country at that time. So I end up-- really end up going to the Virtual Reality Conference, and so I’m still involved actively with IEEE Virtual Reality Conference, and I'm actually currently sitting on the steering committee for IEEE Virtual Reality.

Asaro:

So how much overlap is there between those two communities?

Lin:

I think there is substantial overlap. I mean there are some people who go to both conferences, and I think that's an area where there could be more synergy. In fact, we are-- we, as myself and Steve LaValle and Paul MacNeilage, we are organizing a conference, The Boundary Between Virtual Reality and Robotics, at ICRA this year, and the idea is to get people who are doing very similar research but looking at it from perhaps a slightly different perspective to come to know what was going out in the other area, like try to bring some-- our researcher to a robotics conference and try to get a robotics researcher become aware of what's going on in virtual reality. So we're going to have a full day of invited talks, we're going to have hands-on demo, participation by the attendees, and so we're very excited. We're looking forward to it. And we also get a boost very recently by the acquisitions of Oculus by Facebook. I don't know if you know about that? Yeah. So there is definitely, I will say, an insurgence of interest now in VR partly because of Oculus because now it is actually possible to buy a commodity device, right? A head-mount display can put on, and it can interact with the virtual environment for 400 bucks. So it's real interesting. There are really interesting question regarding what do we know in VR that we can bring it over to robotics and what can't robotics researcher have already known. And there's really, I would say, a significant crossover beyond just haptics, but other sort of simulation technique that are very, very much integrated; tracking, for example, full-body tracking. People are now looking at Kinect as a commodity, a little commodity, right? Just a couple hundred bucks. You've got a cheap tracking device that can immerse you in the VR environment, but you can also use it for robotic application, so there's certainly a lot, a lot of crossover between the two.

Other Projects

Asaro:

And what are some other projects that you've worked on <inaudible>?

Lin:

I work a lot in what I call physically based modeling and simulation and animations, and that's just sort of kind of another branched off area for me. I also has been working on collision avoidance, which, by the way, is very much related to robotics, and part of that was, again, sort of kind of going back to my root on motion planning because I work in motion planning for many years, and I have been publishing paper in motion planning all the way up until 2000, even as early-- I mean as lately as 2010, I think, was the most last motion planning paper I have publish. So one of the things that we have done being at UNC, which is also one of the few academic places that actually designed graphics hardware, and we also have designed some of the head mounted displays, so we have a whole entire generation sequence of a hardware, graphics hardware, which can also be regarded as a supercomputer because the design of graphics hardware take into account a lot of parallelism. And so being in a place like UNC, it kind of gets you to think about a lot of things, which is supercomputing and parallel computing on your desktop machine in the nineties when no one is even thinking about that. So we have designed a lot of parallel algorithms that were run on graphics hardware even in the late nineties. So we already were thinking about how can we take that ability of the graphics hardware to do real-time motion planning.

So we were already doing real-time motion planning using graphics hardware by late nineties, early 2000. We have some paper that was published in ICRA that was looking at taking advantage of graphics hardware. I mean some people have picked that up, and I’m not sure how much there is until more recently. Now, people start thinking, "Oh, you can actually use graphics hardware to do a lot of things." And so we have done whole bunch-- a whole entire sequence of algorithm which we call general purpose computing and graphics hardware, and we have organized workshop in that area and was sort of prelude to some of mini-core computing because we start thinking, you know, how can we possibly do this kind of high-performance computing on your desktop so that your desktop machine, it's a supercomputer. And that concept has gone far beyond desktop computing. It's-- your cell phone today has a mini supercomputer, right? It has graphics chips in there, and we have used that ability to look at a lot of-- not just robot motion planning but also robotic simulation, richer body dynamic simulation, articulated body dynamic simulations. So we have-- even looking into the use of graphics hardware for designing robot motion planning algorithm for developable robots. So things like-- and that was great because it's a coupling between what I do for physically based simulations and robotics, so it's like part of my-- it's all my passions sort of kind of combined in one. And we've been looking at how can you possibly design simulator for medical robot that are deformable and that are articulated and can go through human body, and we can do all that in real-time pretty much ahead of everybody else.

So that was sort of part of my work at UNC, as well, which is really looking at physically based modeling simulations, simulation of soft tissue, simulation of articulated body, like robots, like human, and we have also looking at how can you take advantage of some of these simulation technique integrated with robot motion planning. So we have looking into optimization-base robot motion planning very, very early on that will integrate physical constraint with geometric constraint with your motion planning constraint with a biomechanic constraint for the robot, for the human, you name it, and we have worked quite a bit in some of those area, as well. And more recently, it's going into fluids so we are also now looking into how can you do real-time fluid simulations. And then we started some discussion, research discussion and possible project with S.K. Gupta at University of Maryland looking into how can we design robot motion planning algorithm and manipulations for robot manipulator to manipulate container with fluids.

So if you want to apply fluids, like oil, for example, paint, to spread paint, how can you control and apply the fluids in the desired locations that you want? And they're a non- trivial problem, right? So we are looking at a combination of physically based simulation of fluids combined with decades of robot motion planning algorithm and manipulation algorithm with real-time monitoring and sensing and modeling of an uncertainty, as well as machine learning because we certainly want to learn from human demonstration, but it's far more complicated because some of these-- I think there are a tremendous promise in terms of demonstration-based algorithm in terms of machine learning, but a lot of these learning technique also depends on their input data set, which is essentially based on how many different kind of human demonstrations that you can give it as a input. And we believed that if you have high-fidelity simulation model, you can generate far more possibilities and examples beyond just human demonstrations, so a combination of both human demonstration plus this physical models can be a really nice combinations to be integrated with machine learning.

And, also, I believe-- that's one area which I've been looking at-- was this idea of integrating the huge amount of data that was all around us that we sense from camera, from all different type of sensors, that we have all these data that we have capture. And a lot of time, these are data we'll capture with human in the loop. Can we learned by combining simulation and these data what are some of the intrinsic simulation parameters that we need to improve the simulations? And when we reach that level, then I think that we have done truly data-driven modeling because the early day when people are talking about data-driven modeling, it's a lot of capturing of motion data and then play it back or trying to figure out how to interpolate it. But decades and centuries of physical model and mathematical model that we developed are not for-- just for academic interest. There actually are physical laws and mathematic reason. These are the governing law of basic principle how we can describe the dynamical systems and the physical environment that surround us, right? But doing simulation, it's hard because a lot of time, you don't know what are some of these simulation parameters that you need to use. And my joke is a lot of people say graduate students is cheap compared to full-time researcher, so they'd have graduate students sitting there in the lab and they tweak the parameter until they works, and that's why there's always a question, how do we know your simulator is actually producing the right result, or it's just simply because your student's sitting there and tweak and tweak until the simulation match what you observed.

So it's a really interesting question. So the question of how do you come up with a matrix, how do you determine those simulation parameters are critical. And I think the very fact that we got so much data, the huge amount of data that's out there, will be a great way for us to determine what should be the appropriate simulation parameters automatically. And this is what I refer to and what a lot of people refer to as solving the inverse problem, so that's the first thing. The second thing is that it offers essentially a platform of test suites, where we can validate our simulation results. And so in that regard, we have-- if we can combine simulations, all these captured data, and machine learning, I think the potential is tremendous. And if we can combine human intelligence in there somehow, I think the future of robotics and AI is just unlimited. I mean there are people who are-- there are definitely philosophers and researcher who are questioning-- we had this conversation yesterday-- about the limit of AI, and I am, for one, that believe it's unlimited, that anything is possible.

Collaborations with Ph.Ds. and Postdocs

Asaro:

So speaking of graduate students, who are some of the PhDs and postdocs that you've trained that have gone on to do work in robotics?

Lin:

Miguel Otaduy is currently in Spain, so he's still working at haptics, and he is, I guess, advisor of IZ Think Alike [ph?]. He's independently working a lot of data-driven simulations. We haven't collaborated recently because I always believe the students need to have their own identity, and so we haven't-- we are very close. I mean we chat and a lot of time we still have e-mail exchange, but I haven't really collaborated with him that much. I mean there are some project that we have, some joint paper that we have written since graduating but knowing how much we think alike, he independently pick up his own project that he's working on. So he's working definitely on data-driven simulations. Haptics, he's also pretty well known in that area, so he's definitely working on haptics/robotics area. I'm trying to think. One of my postdoc who went back to India, and he is working on the kinematics algorithms in applying to molecular docking, nano manipulation, nano simulation area, so that's Stephane Redon. He is the head of NANO Modeling NANO Simulation NANO Sciences Group INRIA. Oh, gosh, I have other students. And then there is Stephen Guy. He is at University of Minnesota, Twin City (sic). So one of the things I have not talk about was one of the application that we have looking into was also a spin-off from some of the work that we have done on collision detections. Collision avoidance and motion planning is actually crowd simulations and also modeling of swarm robots. And Stephen is one of the students who is working in that area, so he's currently assistant professor. He's been at University of Minnesota for about a couple years. He's looking at largely multi-agent simulation, modeling, and collaborations, so he's working also with Maria Gini. I'm sure you have met her. You interview her? Yeah. So trying to think of anyone else <inaudible>. I have a lot of students. Not everybody's in academia. Oh, yes, another postdoc.

That's Jur van den Berg. I am not 100-percent sure whether he has resigned or not, but Jur van den Berg was one of our postdoc who work in collision avoidance for a couple years, and he had a faculty positions in robotics at University of Utah. He's very well known, even for a junior faculty, for some of the work that he has done with us and also some of the work that he have done in collaboration with other folks at Berkeley.

And when Google X was looking for an experts on collision avoidance and motion planning, they make him what I understood was a irresistible offer. So they pull him away from Utah, and he was supposedly on leave working at Google X. So I am not 100-percent sure whether he had resigned or not, but very likely he has. For a little while at least, I know he was a on leave on Utah. So that’s another post-doc. Oh, gosh.

Another post-doc Vivek Kwatra, I don’t know how you would call him. He does video processing and tracking so he’s kinda cross of vision graphics. He is also at Google Research. So we work on that area as well. We look at how do you do video processing and integration with reality using sorta a combination of vision technique and graphics technique. So we work on that area as well. Oh, boy. I should have been more prepared, right? But to remember all my students and all my post-doc and I’m trying to-- oh, Yu Zheng, of course. Yu Zheng is-- he just graduated last year and he is now assistant professor at University of Michigan Dearborn, and he work in grasping. So we also had work on grasping algorithm, optimal grasping. And, in addition, he also work on humanoid robot. He spent some time at Disney Research. He was a post-doc there for a little while so he is continuing to work on humanoid robot and service robot there at U Mich. And who else? Oh, gosh. Oh, and I don’t know how to contact but I have another student who work on sorta multi-agent simulations. And he also work on this idea of, if we have sensor data, that we get from the ground sensor or from the camera can you use that essential data to reconstruct the motion of these ground vehicles. And so we called this motion reconstruction, traffic reconstructions, just purely based on sensor data, and we have been able to do this in real time. And he is currently at Google as well. That’s David Wilkie. Gee, okay. I’m slowly counting. I have a lot more students than just the robotics, as you can probably imagine.

Asaro:

Yeah.

Lin:

Yeah, so I’m only giving you the robotics ones.

Asaro:

Okay, and similarly, who are some of other faculty or research institutions you’ve collaborated with over the years?

Lin:

I have worked with lots of people. I know, at some point, I kept track of it until ‘75 and, at one point I know I was beyond 100 and I lost track of it. So...

Most Influential People

Asaro:

Who are some of the ones that have most influential on your thinking and <inaudible>?

Lin:

God, that’s a hard one, ‘cause if I left out anybody they are probably gonna be mad at me. I mean, certainly, my advisor, right, is very influential. But my colleagues at UNC, Fred Brooks. He had this idea <inaudible> and he’s one of the touring award winner. He talk about we do research to ensure we have an end user. It’s a really nice interesting philosophy that I think certainly have an impact on me, on my colleague Henry Fukes [ph?], who I have learned a lot about augment reality/virtual reality from. We have, I would say, minor collaboration. I don’t think we have collaborated on any paper but just sorta general projects/discussions. My colleague Russ Taylor, who I am very, very grateful because he was my colleague, my peer, but he has done some tremendous work. He developed the first Nano Manipulator from haptic devices so that you can manipulate object that is a million times smaller than its actual size. And Nano Manipulator is unique. It’s a first of its kind. Just imagine that you literally are shrunk a million size. Or the other way to look at it is, “Honey, you blow up the viruses a million times.” So you can see something million times bigger. You can manipulate it. You can feel the forces. He is the go-to guy when I started working in haptics and so I learned tremendously a lot from him. And I am very sorry I am not at his retirement reception today. I wanted to be there but-- he’s retiring. He wanted to do something that is impactful for the mankind. So he decided to retire from his faculty position. He’s my age. So my hat’s off to him and wish him all the best and I’m gonna miss him.

So those are some of my colleagues but I certainly have collaborators all over the world. I have learned a lot from Ken Salisbury, who have left MIT, went to Stanford. I have work with a lot of people at INRIA. Some of our most recent work on cross simulations. And I have close collaborator at Tsinghua University. I actually have a secondary appointment, which is more a courtesy appointment at Tsinghua University in China. So this is-- I’m going backward in terms of chronological order. Oh, and I finish up that. So I would say that we have also a research collaboration and discussion with, I would say, Jean-Claude Latombe on motion planning. I would say Jean-Claude has also a tremendous influence, even though we have never collaborated on papers. We have lot of discussions. Davids Hsu, his PhD students, we have a lot of interaction. Lydia Kavraki. All that discussion really help us to think about problems that we did not think about ‘cause they work on motion planning and they need proximity current libraries. And definitely, Stephen Neville at UIUC who is my co-organizer for the workshop. So those are the one I mention, just mostly people in robotics and <inaudible> area.

Chinese System

Asaro:

Great. You mentioned that you have an honorary appointment in China. I wondered if you had any insights on what’s been happening...

Lin:

The Chinese system?

Asaro:

...in robotics in China and Taiwan and <inaudible>.

Lin:

Yeah, well, let me just say this: since I’m on camera. I really hope U.S. will wake up to the very fact that I do not think we are way, way behind. ‘Cause I have some insight to the Chinese funding and the kinda resources they have. We have a small fractions of what the Chinese researcher have in China and it’s scary the differential of what the Chinese government is putting into research and educations. If you go to some of their workshop in robotics-- I’m just gonna focus on robotics. The Chinese government is putting in so much funding. I mean, this is interview, they targeted us a way for them to accelerate their productivity, their manufacturing capability, and the ability to compete in technology, in automation technology across the board. So once the Chinese government have identified the <inaudible> area, they just dump tremendous amount of resources into it. And, despite the fact that we do have National Robotics Initiative, by NSF, with multiple agency join force, I do not believed the funding level is anywhere comparable to what the Chinese government are actually putting in, based on what I have seen. And I think perhaps the studies do, at some level within U.S., to really take a look at what’s going on around the world, even compared to Europe. I have a lot of European friends and I know they are going through some turmoil in terms of economic hardship and this E.U. problems that they have to deal with. But there is a commitment at the E.U. level and at individual country level. So this is double funding. As a union, they are putting in resources for doing research and what I am seeing is that I’m seeing American researcher, who are well-establish, who are train here, who’ve been faculty here for years are now leaving U.S. going back to Europe because they have guaranteed funding for six students for the rest of their life. Versus I spend lots of time thinking about what I need to do next. I spend lots of time writing a proposal. I spend lots of time worrying about managing funding, whatever little I have. And I spend lots of time thinking about where I’m gonna get my next grants. And the funding rate, I think, are inflated at some of the agency, because they are counting some of these small c grants as into the computing of the acceptance rate. But even when you do this inflated acceptance rate, it’s extremely low compared to everywhere else. I mean, we are a small fraction and I think that a lot of people forget it’s not just research. It’s research and educations.

We don’t do something to change that, we don’t invest in our future, the U.S. is not gonna be the leader in technology. Forget it. I mean, we are still basking in the glory that the U.S. government has invested in research and education from the '60s from the '70s, and it’s kinda trickling. You think about people like Steve Jobs. You think about people like Bill Gates. These are the product from the education and the research infrastructure that we had put in place. And we still have some of that influences but the very fact is I believe that we are losing human resources in a tremendous rate. ‘Cause I know there are faculty who are leaving for Europe for better position. There are faculty who are leaving for position in Asia, simply because there’re just simply a lot more resources to do the kinda research you wanna do. And the living standard is improving in Asia and the living standard is always very high in Europe. So, despite the turmoil that they have, they still have the research resources that I think a lot of faculty here in the U.S. just simply don’t have. And I have to say robotics is doing relatively well compared to other area in research, in science and technology and that got me worry, ‘cause I have kids and I worry about where is the future for the U.S. so I don’t know. I mean, that’s definitely one thing I have seen, especially when you go to a robotics conference in China that was local to just China. And the amount of technology and the kind of application that they are thinking about, it just mind-blowing. It’s like seeing the Japan 10/20 years ago, and they are using robotics in the most innovative way on daily basis. I think you guys have seen the slicing noodles robot, right? But it’s more than that. It’s how to massage you, how to interact with you, healthcare robots, household robots, that kinda thing. It’s a robot everywhere. So they are really thinking about things far beyond than some of the things that we have, and they have the resources. I mean, that’s really-- I think it’s gonna be what makes the difference.

Asaro:

Historically speaking, do you see that China was sorta late in coming to robotics as a research area?

Lin:

Oh, yeah. Yeah, China is late-coming. I mean, but I am also-- China is definitely late-coming. I would say that China came up in just the last 10/15 years or so. I was the editor in chief of an IEEE journal. I am currently the chair of the IEEE Computer Society transactions operating committees and I see that we have a lot more submissions and we say majority of our submission are coming from China. And they may not be accepted as the higher acceptance rate like in U.S. and Europe. I would say that is largely due to their language problem. Not necessarily because of <inaudible> research quality. And so just imagine that, at some point in time, that if they have learned-- China is also trying to teach their students how to improve their language skill, especially writing in English and speaking in English. Just imagine that they are become better writer and better presenter. They are able to articulate a research better and they’re certainly will have more publications. I think, in terms of publication, they’re already number two in terms of the number of publications. And just traditionally, they’re publication has not been as well-cited. I think partly because Chinese names are very confusing. I mean, even my last name Lin. It’s very common. There are tons of Lin. There are tons of Ming, and so it’s very confusing for Westerner to figure out which is which.

And so there is lesser a main name brain that’s associated with the research that’s being done there. But I think there’s a lot of interesting work that’s going on. It’s not getting recognize by the outside world and-- I mean, outside world meaning the community outside of China. But, within China, there’s tremendous amount of activity that was going on. I go to China at least couple times a year and so I see a lot of this what’s going on and what’s happening there and that kinda get me a little bit concerned about what should U.S. do differently.

IEEE Activities

Asaro:

You mentioned some of your roles in editing? What’s been your relationship with the IEEE and the leadership of the IEEE?

Lin:

I’m currently also-- so other than being a former EIC-- I just stepped down-- and being the current chair of the IEEE Computer Society Transaction Operation Committee, I’m also a member of the board of governor of IEEE Computer Society. I am on the steering committee of IEEE Virtual Reality Conference. I was a member for IEEE VGTC. That’s one of the technical committee. So those are the involvement. I mean, I’ve been organizing IEEE conferences and ACM conferences, so.

Asaro:

This is the first time there’s been an all-female <inaudible>.

Lin:

Yes, it’s tremendously exciting. I cannot tell you. Although we are 100 percent sure it’s gonna be all-female yet. Yeah, but it’s tremendously exciting to see this is happening. I think this is the first time. I think we-- I hope we are making history here. So I think it’s great, yeah, it’s great. There has been some members saying that-- I think there are excellent women researcher as well as men, but generally speaking, there has been some study that say that women, a lot of time, tend to hold back if they feel they’re in the minority. And that might be personality. That might be-- that probably would be the same for men. I mean, we haven’t done that experiment, trying to put one man against four women in the same group and we’ll see what the man would do. But generally speaking, in many of the professions, that women is the minority. And there has been study that show that, until the representation of women has reached beyond 30 percent or so, that they tend not to voice their opinion as much. So having a all-women organization for a conference, it’s-- I think it’s a really interesting experience and I am looking forward to it. I’m looking forward to the conference.

Challenges for Women in Robotics

Asaro:

What are some of the challenges and maybe some of the solutions to getting more women involved in robotics?

Lin:

It’s hard. Okay, I mean, I think this is, generally speaking, a hard problem because-- and I do care. I think I do care a lot more now than I used to when I was a student or when I was a young assistant professor, and for very personal reason. ‘Cause I have two daughters and I watch them grow. And it has always been my two kid-- my daughters, they play chess and they-- thanks to, also, a male mentor, they learned how to play chess in elementary school. And I could see that, even at very young age, that they feel they were outnumberred by boys of the same age and it was discouraging for them. But I keep telling them, “You enjoy playing chess, keep playing. Ignore who’s sitting next to you and just keep playing. Just have fun.” And they also like math. I guess, maybe some of that’s in the gene. They like math. They do very well in math and maybe it’s the emphasis that we have. I think that we, being-- both my husband and I are technology oriented. We always think that the science, the math are important. So both my daughters have been very good in math but competition-wise-- you have going to all these competition, the Science Olympiad, the Math Olympiad, all these competition. You don’t see that many girls and it make you wonder what is wrong. Why aren’t that many girls? And I think a lot of that-- having been born in a different country, and having grown up, spent some of my time in a different country, and having gone to an all-girl catholic middle school, that I have never, ever feel that somehow girls was less capable than doing the math and science. And I have never, ever witness the differential of somehow women or girls are less capable. So I don’t think that you have that kind of differential in other country like, for example in Asia, as much as here. And there’s a lot of culture issue, I think, that’s associated with a society that is making the girl thinking that it’s somehow that’s not the area they want to get into. And I think that’s a talent wasted. I’m not sure what would have happened to my daughter if I had not encouraged them. Because, when they’re out there and they see so few of their friends, it does feel lonely and I know there were time when my daughter wanted to quit chess or they wanted to quit just doing whatever and I told them, “No, no, no. You should not.” I mean, I joke with them. “You are unique. Enjoy the uniqueness. You are special.” I try to make it spin a very positive spin, ‘cause I know they like it and I know they enjoy doing it and I know they are good at it. So why should they not continue to pursue because there are so few women or girls. So I try to encourage that but it really does make me think, because they keep saying, “There’s so-- I’m the only one,” or “There are so few girl doing this.” So I think numbers matters.

The environment matters and parents matters. So I think the family need to, the parents need to, encourage their daughters to think about that. And I don’t think this is a issue of equality. I think that it just simply makes a better society. Women and men are-- let’s just face it. We are different, okay, and we have very different perspective and we have diff strength. And having both gender equally involved in the same area, you get a diverse perspective. It’s the same reason why we want diversity in workplace. So I would love to see more women be involved in robotics. I would love to see more women involved in science and technology and I don’t exactly have the solution for that. I think a lot of it, it just a culture issue.

Advice for Young People

Asaro:

The question we usually wind down with is what’s your advice to young people who might be interested in a career in robotics?

Lin:

Ah, explore it. Get your hands on. Explore your interests early on and, if you think something is fun, go out and-- there’re a lot of robotics toys. My daughter play with a Lego robot when she was still in elementary school, very early on. She’s taking robotics course right now in high school. And just play with it. I mean, get your hands dirty with robots. And the other thing I have to say, I’m gonna put in a plug. Computing is so critical and I don’t necessarily want my daughter to be a computer scientist but we encourage her to take a computer science course very early on. So she took a computer science course online, ‘cause it’s not offerred in high school. And that’s another things that I think we can do better. We have math, we have science for elementary school. What happened to computer science? Hello, we are living in the 21st Century. Computer science need to be taught very early on and, if it’s taught very early on, it become a second nature. You think computationally and you think a different way, and it becomes a second nature to you. So for that reason, that’s why I recommended my daughter to take the online computer science course. And not necessarily because we want her to be computer science but we look at it just as everything else. Like statistic, like math, like physics, like chemistry, like biology, like humanity, like English, like history, like economics, like music, like art. It should be part of the elementary educations, middle school or K-12 education. It should not be some sorta luxury courses you take in college. It really should be just there as part of the K-12 education. So I would say those are critical. You need to be able to have all these general education knowledges, so you can make a decision what you want. If you don’t know, then how do you know? And I do know one thing. I had a freshman who approach me and she asked me the question. She say, “I really like your course and I’m thinking about majoring in computer science.” This was a course which I taught as a first-year freshman seminar and she was very concerned. She just started taking a programming class in college and she asked me, “Am I too late? I don’t know if I’m gonna be able to compete with everybody else.” And that’s the fear. That’s the fear she has and that could very well be the detrimental effect that will stop her from taking it. Because she fear she cannot compete with everybody else. So I told her, I say, “Nope,” that “I think you should continue.” And I told her, “I started learning programming the same time as you did. I can do it. So can you. I got my PhD. If you ever aspire to, so can you.” And so I say, “No, it’s not too late.” But she talk to me and maybe some other girl would have talked to some other faculty to ask that question. But, if some student never ask that question and they just quit, just imagine the potential talent that we would have lost that was never explored. So that’s one of the reason I think that it needs to be taught much early on, so they don’t feel like they are in a disadvantage situation when they get to college. ‘Cause that’s, for a lot of them, it’s too late because they think that the young boys, or the boys, or the men in class already have that very early on. I have asked a lot of my student, my graduate students, when did they start on learning how to program and the women usually much later, the boys, much, much earlier. We’re talking about 11, 12, 13, because they played games. They got interested in game. It got them to think about programming. So they just start doing it very early on. So there was the game affect that I would say that have impacted the boys compared to the girl. And also, girls don’t play as much of those traditional shooting game, explosion game, sports game. They don’t get the same kinda stimulant. So I think that is what-- if you ask me, I think that’s what makes some differences that we have seen. I don’t know. Who know when we gonna change that? All right.

Asaro:

Great.

Lin:

All right.

Asaro:

Thank you very much. Thank you so much.