table of contents


Eric Roberts on Computer Science

After receiving his Ph.D. in Applied Mathematics from Harvard University in 1980, Eric Roberts taught at Wellesley College from 1980-85, where he chaired the Computer Science Department. From 1985-90, he was a member of the research staff at Digital Equipment Corporation’s Systems Research Center in Palo Alto, California, where his research focused on programming tools […]

download transcript [vtt]
This is KZSU, Stanford.
Welcome to entitled opinions.
My name is Robert Harrison, and we're coming to you from the Stanford campus.
When I say coming to you from the Stanford campus, I mean the campus of Stanford University,
in the Golden State of California, on the western edge of this republic that goes by the name of the United States of America.
Land of the Free, home of the brave, as stated in the national anthem, they sing at our American football games.
Where is the mute button?
It would be nice if the first thing that came to people's mind when you mentioned Stanford University were entitled opinions,
but we're not there yet friends, we're not there yet.
For now, the first thing that comes to mind is Silicon Valley and the great Cambrian explosion of computer technology and digital gadgetry.
Those of you who follow this show on a regular basis know that I don't drink the Silicon Valley Cool-Aid,
my allegiance is to the three-dimensional world, not to the sorcery of computer screens.
But let's give credit where credit is due.
If you're not listening to this show on KZSU's live radio broadcast, then you're listening to it thanks to computer technologies that were probably created in Silicon Valley.
And those technologies, in many cases, were first incubated at this university.
Our topic today is computer science at Stanford.
Stay tuned in title opinions coming up.
The guests who joins me in the studio today is Eric Roberts, who is something of a legend here at Stanford and in the field of computer science education in general.
Here's some preliminary bio.
Eric Roberts received a PhD in Applied Mathematics from Harvard in 1980.
And from 1980 to 1985, taught at Wellesley College, where he led the effort to create their computer science department.
In 1990, he joined the Stanford faculty, and from 1990 to 2002, Professor Roberts served as the associate chair and director of undergraduate studies for our computer science department.
And in that capacity, he was the principal architect of Stanford's introductory programming sequence, writing textbooks that are used in those courses.
He has written six computer science books in all, which are used at many colleges and universities throughout the world.
Eric Roberts also serves as chair of the ACM Education Board and several important task forces.
I think you get the idea.
My guest is a leader in the field of computer science at an international level.
And it's a pleasure to welcome him to the show, Eric, thanks for joining us on entitled opinions.
Thanks, Robert.
Well, I think it's fair to say that you've done as much as anyone at Stanford to make our computer science undergraduate program.
The most dynamic, vibrant, and successful major at Stanford computer science is now the most popular major.
And I have read that some 90% of Stanford students take at least one computer science class before they graduate.
So my first question to you is what were your goals when you set out to develop the computer science program here at Stanford over 25 years ago?
Well, the first thing that I wanted to do was to make sure that people have exposure to this dynamic, unbelievably fascinating and intellectually challenging discipline.
Because very few people have exposure to it in schools.
When people take what are often called computing courses in high school, they're seen as primarily career oriented, almost vocational sorts of courses working with word processing or spread sheets and the like.
But the intellectual character of the field is exciting and if people come in with a bias against it because of what they've seen before, they won't know whether this is something that they'll be excited about or something that they'll be particularly good at.
And so my goal from the very beginning was to offer an introductory program that would be broad audience.
I wanted to reach that level that we've reached today of having almost all Stanford students take at least our introductory course if not more.
And that exposure that we have of getting people to try it out has been effective in allowing people to discover their own talents and to pursue them and to go in directions that they didn't in fact plan.
So someone like Marissa Mayer who's now CEO of Yahoo could take our course for non majors and then come back and take the introductory course for me and decide, I really like this.
I'm really good at this and we see where that's led down the road.
Are these introductory courses about how to code or what is actually covered, I forgive my naive today but not having taken one myself.
Well, certainly they include coding as part of it.
You have to learn how to speak the language of the discipline in order to use it as a tool that empowers you to solve problems.
But what we like to think is that what we're teaching in those courses is problem solving more generally.
And it really isn't so important what programming language we use many different schools use different programming language Stanford has changed its programming language twice.
What is a programming language?
Programming language is the rules of syntax and semantics that allow you to express a set of instructions or operations that together constitute what computer scientists call an algorithm.
A procedure for getting something done. And nowadays that's always done in a style or language that is intermediate between what the computer understands and what humans are likely to understand and there's a translation process to move one to the other.
So people write in a language that gives them expressive power to code those algorithms and a form that resembles the human understanding of what's going on and then you turn that mechanically into something that the machine can execute.
While I've been at Stanford we've taught that introductory course in Pascal and in C and in Java and I worked with those transitions.
Those are all languages and I wrote the textbooks for each of those languages as it matured and we kept up with the technology of programming language design.
So the difference between the introductory course for non majors and those for majors is how much of it is a technical difference and how much of it is an introduction to what is computer science in general.
Well certainly our courses we now have two for non majors CS105 which is introduction to computing that is taught to give people who are not particularly interested in being able to use that technology as a tool teaches a lot about the context of computing.
It does teach a little bit about programming but it's really more computational fluency or computational thinking that's the goal there.
We also have a course called CS101 which is a prototype for the new computer science principles advanced placement course that's being introduced next year as a real AP course.
Interestingly the enrollment in those courses the courses for non majors has declined significantly in recent years as more and more students take our introductory course the annual enrollment during the three academic quarters in CS106A alone that will single course is more than 1600 students and the class size is only 1700.
So there are some graduate students taking it so it's not all great but it's penetration level is high almost all students take one of these courses and they want that additional skill set because they know it's valuable in a wide variety of disciplines.
Since the vast majority of the listeners of this program are not Stanford students unless they happen upon entitled opinions on KZSU.
You talked about CS106 that is a computer science 106 course sequence of courses that you yourself were responsible for actually programming know.
CS106 has been the number for the introductory course since 1967 which is well before my time but I've certainly designed their current incarnations.
And the last two incarnations in fact right and so is that a three quarter sequence is a year quarter C two quarter sequence and that is the course that is enormously popular that you were just referring and it's the first quarter sequence.
The CS106A is the first quarter 106B is the second and they're both extremely popular the annual enrollment in B is well over a thousand students and what do you attribute the popularity to?
Several things one as I mentioned earlier those courses are useful.
They're useful no matter what you're studying because knowing something about the nature of computation is important to almost every field today.
It's used in the humanities the whole branch of digital humanities uses computational techniques to analyze text.
It's used in social sciences to do statistical analysis and prediction having some understanding of what computers can do that notion that has now been termed computational thinking is critical and that course does a very good job of introducing people to those skills those ways of thinking ways of knowing that we value a university so people see that.
It's also important to know that those courses are taught by some of the most creative teachers that I've ever encountered.
We've hired them specifically for that purpose.
We have a staff of lecturers and professors on the teaching line whose primary goal is to teach those courses in the best most enabling most inclusive way.
That we can to try to make sure that everyone has a chance not only to see that material but to succeed with it.
We've worked very hard to make sure that the safety net is in place that there's an army of helpers so that when people get stuck they can find the assistance.
The lecturers are extremely popular among the best rated courses in an entirely unscientific survey that came out in the joint publication of Business Insider and
They picked the ten best professors at Stanford and eight of them were in computer science.
Probably wrong but we have worked very hard to get those faculty who can teach well and we put them in front of those introductory courses.
Eric, how do these courses work? Because obviously with 1600 students you have multiple sections or offerings of the same material.
I suppose these lecturers or professors who are so highly rated do they give lectures and then they have graduate students or other hired instructors that hold sections for the students.
The courses are all taught every quarter. So when I give you annual aggregate numbers we're talking about the third of that each of the quarter.
It's still large. We have 600 students plus in the current incarnation of CS106A.
There's a lecture who gives three lectures a week. The idea is to use that lecture format in the most creative way to generate excitement to use the fact that you have a large community of listeners and get them excited.
There's a performance element to it certainly. But that's not where we expect the learning to take place.
It's where some of the motivation takes place but the learning takes place mostly when students are engaged in solving the problems on their own.
And for that they need some kind of assistance and we do have small sections that go along with that.
We don't use graduate students. They're nowhere near enough of them. We wouldn't have enough graduate students even to consider that.
We use as most universities in the United States use undergraduates as our instructors in recitation sections for the computer science classes.
They're called section leaders here. They go through an intensive training program. There's a very rigorous selection process that's still accepting only about one applicant and four.
And those people in groups that we try to keep around eight so the student faculty ratio is eight to one in those sections help those students along and provide that assistance which is integral to anyone's mastery of the field.
That's been going on for a long time. I did it as an undergraduate at Harvard. I've heard President Hennessy say on more than one occasion that he would not be an academic had he not had the chance to teach in that mode as an undergraduate section leader at Villanova when he was an undergraduate.
It got him excited about the prospect of teaching. So when you arrived in 1990, how great a difference was there between now and then in terms of the major and the popularity of the computer science major undergraduate.
Well, there's been a huge increase particularly in the last seven years. In 1990, the computer science undergraduate major was only four years old. It had been started in 1986 and had of course, therefore very small graduating classes as we built up a cohort in the first year.
We don't think there were very many people who had fulfilled by accident the requirements for a major that did not yet exist. So we're putting it into place and by 1990, Stanford had decided that they needed to bring someone in to run that program because it was growing in popularity.
We did grow in popularity through the 90s and through the dot com explosion. And then when the dot com crash occurred in 2000 2001, we saw as everyone nationally did a decline in numbers that flattened out and then starting in about 2007 just started to rise at almost that more's law rate of doubling every 18 years.
And the rate of that expansion has led us to a position that is much higher than anything we've seen before. We will this year almost certainly get up to the level of a thousand currently declared majors in computer science. We were close last year. And of course, after June, you graduate some so we're below that now.
But as we declare new students were on a trajectory that should put us over a thousand. The next closest major is down in the 300. So we're almost three times the size of the next largest major which now is the engineering special programs such as product design and then human biology right under that.
We had the amazing milestone this year of becoming the most popular major for women. It's historically been hard to attract women by having such a large cohort and the critical mass of women.
We now have 30% that's almost twice the national average of our undergraduate population being female and you've been very committed to trying to get women to major. No, this has been something that I've worked on since I started my teaching career. I went to Wellesley very specifically to do that. Wellesley was not advertising for a computer scientist. They had no computer science department.
I wrote to Wellesley and said I was interested in making one and they thought that was a good idea and hired me as an assistant professor and then made me the chair of a department three years later. So it was and were you successful at getting absolutely wellesley women.
Absolutely we could tell there were some sociological studies done that discovered that you could see the appearance of the Wellesley computer science department in the employment patterns in Massachusetts that we were adding women to the pool of potential employees at a level that was measurable.
What can I ask what sort of strategies you use to have used to attract women to the to the major. A number of them first of all having as open a class as possible that broad audience approach is very attractive and adds to the diversity of the program. If everyone takes it then of course the ratio in the undergraduate class is the same as the ratio and the university.
So we'll have very good participation by women and underrepresented minorities and we haven't fact done just that. But we've also done a number of things in terms of the program.
One of the aspects that I think is most important is we've worked very hard to ensure that the section leader community, those undergraduates who teach the small sections reflect as much as we can the diversity of the
institution because it's been my experience that students respond most positively to role models that are close to them in age.
That very few people look to the front of the room see me there and say oh in 45 years I can be like Eric that's past the time horizon of any undergraduate.
But if you look across your section room and you see someone who looks like you and that person is a junior for example, I could be that person in two years is a much more reasonable leap.
So I've called this in the papers that I've written about it the stepping stone role model approach that you want to make sure not that the I and the prize target is visible, but that all the steps across that stream are visible so that we have maybe TAs who represent that diversity and assistant professors or graduate student as you move up the pipeline to make sure that that's always there.
We also for example have put in place a number of bridge programs so that people have the opportunity to work in smaller groups.
The engineering school has done that on a more general framework the summer science and Stanford science engineering academy see in the school of engineering brings
particularly people the target people who are at risk and bring them in for advanced work I think we do that with our sophomore college and freshman seminar programs.
We also have worked very hard to purge the curriculum of gendered stereotypes.
So for example my books never use any gendered program pronouns. How do you avoid you rewrite the sentence it's always easy you might repeat the noun it just it you can learn to do that.
I also for example don't ever indicate that I happen to think that the noun data is plural unlike most of my colleagues so I never use it in the context where you can tell it's number.
And the so you just work very hard but more importantly this question of of trying to keep out gender there just aren't first person shooter games or football metaphors in the classes we for many years use this our graphics assignment.
A quilting assignment designed by our lecturer Julie Zelensky who's just a wonderful lecturer and the associations that people make with that are much more inviting and open to women and that that's been true all along and the growth that we've seen in recent years is allowed us to reach that milestone where we became the most popular major for women.
So I have two questions one I'll start with the one on gender because there's.
The popular perception if you go to San Francisco where you have this all the startups and the computer technology in the high tech industry that and it seems like they're all a bunch of kids.
And that's male who are completely awkward around girls and that you go to the bars and there that there seems to be a complete dearth of women in that industry not perhaps this is just a stereotypical perception doesn't correspond to the actual statistics.
But it does seem like that particular industry is not overly hospitable to women or that women don't seem in general to be particularly attracted to it is that correct I think that's a correct perception and something that we need to work on.
I think one of the best ways to work on it is to ensure that the pipeline includes as many women who have been able in a more balanced program to establish themselves and to avoid some of the feelings of marginality or the imposter syndrome and we've worked very hard so that for example, you know the fact that Stanford is graduating now more than 150 women in a year.
And those people are mostly going to go out into that industry and the industry will be forced to change.
Do you know whether many of them find their post graduate entry into the world frustrating because of gender issues or not?
Of course varies from person to person you certainly hear stories of people who withdraw not even necessarily because of overt sexism that they find in the community that's there.
But one of the hardest problems to overcome is that the pace at least at most startups doesn't allow for someone to have much of a life outside of the work environment and even the big companies create amenities on their campuses to try to keep people there as much as possible that doesn't work very well if you're also, you know, trying to under the
dynamics of a ticking biological clock and so people will certainly women leave the field at a much larger rate and all of these issues need to be addressed absolutely.
My own view is that the best way to address at least part of them is by making sure that more women are enabled to join that workforce and make the changes from the inside.
So the other question I had before we moved to another general topic is this huge explosion in the majors and the success that your computer science department has in teaching these courses.
There's a lot of enthusiasm but in your view what is the breakdown between the enthusiasm that is generated by an excellent computer science program for undergraduates on the one hand and the
reality of the get rich syndrome that attracts students as well.
Well certainly we need to contend with that reality when the New Yorker wrote an article, "Dubbing Stanford Get Rich U."
There's certainly a reality behind that stereotypical perception and it's problematic.
One of the reasons that I've taught both courses in the history of computing and computing in ethics and that I've taught courses that are interdisciplinary.
I've taught courses that are cross listed in 15 different Stanford departments over my time here is because I want people to see this as a broader intellectual discipline, not as one that is primarily for commercial interest when
the dot com boom was at its height and people were going off and making these internet startups that would sell dog food on the internet.
It's occurred to me that at the end of the day those people were selling dog food and that wasn't to my mind what Stanford ought to be prepping its students for.
I mean there are higher educational, I mean hereditary academics so of course my own view is that ideas are important.
Well ideas and also science is important.
I mean we are after all in the engineering school so we do want to build things at the same time but yes science, ideas, philosophy, all of those things are essential and sometimes we lose sight of them.
I will say that I think that the problems of having most of the interest come from this finality as you described it of the Get Grinch Quick model was more serious in the late 90s than it is today.
That a larger group of undergraduates and CS today are in it because they find the ideas exciting and have a sense that it can improve the world and that's been very important for being able to recruit more women and more underrepresented minorities who want to give back to their communities.
And I think that's all very healthy. Part of that comes from the fact that the start up boom in the 90s was just so testosterone fueled and there was this Get Grinch Quick sense that was very enticing for a lot of our students.
And today a lot of the interest the renewed interest in computer science came from the financial crash of 2007 and a lot of the students that left economics or those things that were moving them toward the Wall Street jobs which disappeared decided that they should have some secure employment.
And one of the things that's true about computer science is that you have this assurance as a computer science major that there will be jobs.
There's so few people to fill the jobs today that if you get that it's security and security is different fundamentally different from a Get Rich Quick model.
And I'm seeing much more of that security motivation which cuts across more lines within the demographics of the undergraduate population.
So what percentage of the graduates go on to work in particular in Silicon Valley?
I don't know the exact percentage but I would guess 80%. And they are most of them are not in the start up but that they're hired really to do code.
For most of them are for the large companies. What we don't have and you say there's a lot of job security in as well.
I mean there may be instability in individual firms and so because it's a very cutthroat industry.
But overall people move jobs in this industry all the time and there's the sense that the shortage of people to fill those jobs means that perhaps we're producing a third to a quarter nationally of the number of graduates that we need to fill the
even the jobs that are currently on the radar screen of the Bureau of Labor Statistics.
But a person entering that job generally young obviously is their job security down the light. Is there an age discrimination because can you say that well I'm going to be here I'm going to be in this company for the next 40 years and work my way up or are you actually
you're hired to be part of a class of you know,
I've done a lot of work in the labor economics of the field and I think the fundamental problem is this lack of supply of talent.
And so people for example there was this mythology after the dot com collapse of 2002 that offshoring was going to make all the computing jobs disappear.
It didn't happen and it didn't happen because companies needed people wherever they were.
If there was someone in India and someone in the United States it wasn't that they decided that they would hire the programmer from India because they were cheaper.
They'd hire them both. Right. It's just there was the sense that we needed all the talent we could find and offshoring was really a way of trying to increase the supply of that talent that was not being satisfied domestically.
And I think that there is a security in the field age discrimination has been reported but usually debunked.
I've written a couple of articles around that issue that when people are laid off in a company restructuring it's often the people who there are people who are laid off who are hard to hire again.
But that's because no one wants to hire anyone that a company has been willing to lay off.
It doesn't matter how old they are. It just matters that some other company didn't think this was employed was strong enough to keep on.
And so they would rather hire a younger person who's untested at an even larger salary than the person that they didn't retire.
So I mean salary inversions are ubiquitous in the field. So you can't make that argument that there is some kind of age discrimination. The good people at every level are retained. You see this all the time.
Well I suppose if you really do want to get rich you can't just be a coder for some company but you probably have to think of some start up yourself and find associates or people that will be willing to take those risks.
And I know that the risks are huge because I think one out of six out of seven or whatever fail miserably and the actual success stories are small minority.
But there was an article in today's Stanford Daily which I actually told me about it. But I wasn't able to get my hands on before we came on air about faculty members or other people in your Department of Computer Science that might have been a part of the business.
If not conflict of interest how much their commitment to their own sort of let's say entrepreneurial projects which often entail start ups.
These are the duties of faculty members and so forth. How much of a role does money play in this whole equation?
Money plays too much of a role and I'm going to go back and just look at this question. There certainly is this culture in the United States of very much focused on that entrepreneurial capitalism.
We encourage in our own students I think culturally people to become those industry leaders if you look at who people know of as the shining stars in computer science they're not necessarily the scientists and theorists by any means.
There are the people who have made successful companies and so you look at the Zuckerberg or the Gates's or Larry Page and Sergey Brin from here.
And there's a problem with that because as most of those individuals will freely admit the shortage is not an entrepreneur's but in people the engineers who can make the ideas real.
And when Gates talks to Congress about the problems facing computer science before he moved mostly into the foundation.
He talked about the shortage of engineering talent and part of that comes from the fact that we don't encourage it or reward it in the same way.
And everyone is an entrepreneur and no one's building anything we have focused attention on the wrong we this is not efficient economically.
And I think we have a problem like that in our culture.
The problem with the Stanford faculty or other faculties as well is a complicated one.
One way to get your ideas tested and used is to follow this entrepreneurial route and a lot of faculty members are very excited about that when Sebastian Thrawn after winning the challenge of the desert race that are for the self driving car wanted to take that technology to the next level he couldn't do it at Stanford.
The resources aren't there he went to Google to do that.
I don't find that problematic in a certain sense.
The problem is that we don't have enough capacity behind those individuals who are for reasons that I think make a certain amount of sense.
And our positive reasons that decided to at least focus some of their energy commercially.
There's an enormous shortage of potential CS faculty.
My own estimate is that in the next year there will be ten faculty openings for every PhD applicant looking for those faculty positions so that nine out of ten faculty appointments will
Perforce go on a unfilled that there just aren't people there. And in that environment there's nothing that the university can do to say oh I'm sorry but you can't work part time at Stanford and concentrate on your startup because those people might leave and you can't replace them.
I just even Stanford would have trouble. We have not been doing as well particularly with junior hire so the number of our other institutions who are working very hard to recruit from that very rarefied set of people.
So we have to understand that this is not a perfect world but somehow in order to get things to work we have to find a way to attract more people into the teaching profession.
You know world where there's so much money and drive and social pressure to be entrepreneurial.
Is it fair to say that you are at the point where you're actively discouraging students from becoming majors in computer science because you just simply cannot deal with the overwhelming numbers.
We are not doing that and I don't think that any of the people who built this program up would ever do that.
We may be forced to and I think if that happens a lot of the people who have worked to build this program will become so frustrated that they leave.
I did not put this program together to limit it to the people who have the most background coming out of high schools who will be white and Asian males.
There's this enormous disproportionate discriminatory predisposition in the way that people have been taught in this country.
And so if there are limitations as there were in the 1980s when University of California at Berkeley had a 4.0 minimum threshold for being a computer science major.
That rules out most of the people I think I'd like to see join that field and so we can't allow that which means that we need to be able to hire more people.
The soundbite I've been using for the year is that computer science at Stanford is 2% of the faculty teaching 20% of the majors only one of those numbers is growing.
That's the problem we have to fix that and grow the faculty number as long as the student number.
Not shut down the student so that's your well one thing that you did propose a couple of years ago that you went around all the humanity department like the English department and my own vision of literature and cultures and languages.
proposing that there be some kind of joint major or at least major in mind or encouraging your own majors to think about adding a humanities component to their degree.
And this seems like a wise thing to do from many points of view. First thing it takes a little bit of pressure off of view but it also would imagine adds to their profile and their education and essential element that they might not get if they were just focused strictly on the scientific aspect of computing.
Well I say a few things I certainly hope that that idea can flourish it has not done so so far.
In part I believe because it was implemented in a way that was very different in its motivations from what I initially proposed.
My idea was that people at Stanford had multiple interests and they should be encouraged to pursue them.
Here to four we've discouraged double majors for example but if someone had a passion for music or English and felt for whatever reason maybe was parental pressures that so often is or peer pressure or just economic insecurity that they should get a computer science major why not let the major in some other discipline as well to take a music major so they could pursue that other passion.
As well as have that credential. I have many students historically who have been able to do that through the double major program.
But that wasn't what the university implemented they inter implemented a unified structure in which those two fields had to be integrated there's a capstone in which combines the two it wasn't that you could pursue independent directions.
You would have that capstone where there was a CS advisor and an advisor from the other major all of which served both to decrease the audience and to increase the costs on both departments which part of my original plan was that we had to share students because we had too many and so many of the more traditional departments had too few that's what I talked about when I went around to the departments.
And that's not where we got and so the uptake on that CS plus X major has been extremely small at the beginning of the year it was for total across all Xs with no duplication they were for independent so there's no sort of critical mass in any program.
And I think that there are a couple of people in the pipeline but I can't imagine that we will ever as it's constituted today grow to the point that it has any effect on the imbalance among the different majors as the original proposal would have had.
But there's still time to change it and maybe it will work better than I expect to knows I think that's the right idea I've also been pushing an idea of trying to retrain people with doctoral degrees in the humanities and social sciences who can't find academic positions so many of them who are extremely effective teachers.
If those could be given a master's degree in computer science they would be snapped up in the current market to teach the lower division computer science courses and teach something about technology in their specialty.
I could hire thousands of those people across the United States no question and I think people would be really interested in such a program so I think we need to think creatively about how we're going to reallocate.
The teaching capacity from this field where most computer science PhDs get sucked off into industry and don't come back to academia.
Find people who want to be an academia and can teach those courses would be a great thing.
So, finally Eric how do you see the future of the discipline and what sort of challenges as well as what potential breakthroughs are imminent that you could announce to us?
Well I don't know that I can announce any breakthroughs I am worried about the future of computer science education nationally and really internationally too.
We are not keeping up with demand when that happened in the 1980s we basically hit a wall in 1984 had to shut down much of the pipeline.
We didn't have the faculty so we just restricted majors and that led I think eventually to downturn in the industry because they couldn't hire about 1987.
It certainly led to an enormous erosion in the diversity of the program if you look at the numbers for women for example.
They fell after 1984 and haven't recovered until the last few years and most people think that that had something to do with a differential preference of women and men but the number of men fell considerably after 1984 just recovered faster.
It was that we hit this period of capacity collapse and I worry that we are edging towards a similar one that is going to need a lot of investment to overcome.
I don't know whether that will happen but I have been pushing for it.
And the investment would come from which sources the university or this industries?
Well many different sources ideally it would come from non-market actors it's a public good therefore it ought to come from the National Science Foundation or other public sources of revenue and state schools would come from the legislature.
There is some understanding that we need to increase our capacity and the STEM fields.
The fact that computer science is unique among those and being understaffed by factors of 3 to 5 has not in fact led to sort of a targeted investment in that area.
Universities will need to make investments. Industry will have to make investments but that of course is difficult because industry wants return on investment and there is a problem with a tragedy of the common situation that industry hires people with bachelor's degrees to work at their company.
When in fact it would be better for everyone if some of those went on to get doctoral degrees and some of them went on to teach.
But no single company can afford to pass on that person.
Everyone would like the cooperative approach that best of all possible worlds in the tragedy of the common situation would be for people to agree to let some reasonable fraction go on.
But for one company to pass means what happens if some other company hires that person that's the risk that no company can afford to take.
And therefore there's this feeding frenzy on the graduates none of them go on to recreate the next generation.
This phenomenon was called eating our seed corn.
So you're having trouble getting graduate students have trouble getting graduate students have trouble getting faculty particularly at the smaller institutions basically they can't get faculty.
And did I understand you find it just like to clarification did I understand correctly to say that someone could have a humanities major and then get a masters in computer science and get snapped up right away.
I believe that's true.
Can you actually start can you get a masters without any previous I think I can make that happen.
You can make that happen.
Well, what about you?
You're thinking about doing it.
Yes, and that would do a lot to a lay some of the anxieties that students undergraduate have about having a humanities degree if they feel like they can do that and then get a masters and have that job security you're talking about then everyone wins.
Remember what I was talking about though I suppose is that I was looking at PhDs in the humanities so that right and retraining them because those people have more teaching experience and understand a little about the research enterprise and have the credential that they need.
Oh, I see so you're PhDs in the humanities right that could then just get a masters and then join the academic line in ways that they are closed out of now because of numbers.
Well, I have Dylan Montanari here at my side he's our production manager PhDs going to get a PhD and he's very good with computers as you can tell because you take care of all the technical stuff.
So maybe I'm going to talk to Dylan after this show and say you might want to get a major gain of masters in computer science and see what happens there.
Well, Eric has been a very informative and certainly fascinating discussion and I compliment you on all sorts of things not just what you've done to create this vibrant program but also I know that you're very committed to bridging the sciences and the humanities and that you teach also courses in thinking matters where you try to make that bridge happen in your own thinking.
So we need a lot more of that here for sure.
So thanks again for coming on to entitled opinions.
I'll remind our listeners who have been speaking with Professor Eric Roberts from the computer science department here at Stanford.
I'm Robert Harrison for entitled opinions.
We'll be with you again next week.
Thanks again, Eric.
Thank you, Robert.
Take care.
[ Music ]