EE Students Opting for CompSci: Is EE Going Extinct?

[gleem]

Summary: Are too many EE students opting for computer science?

We've seen some people asking about the best opportunity in EE in this forum. It seems that too many are leaving the traditional areas of electrical engineering or at least one person thinks so.

Any thoughts?

 

[Vanadium 50]

I started poking around looking for evidence to support or refute this, but then got lazy. Why should I do the work?

Can you post any evidence to support your conclusion?

 

[fresh_42]

I don't think that it is a valid hypothesis. From people I have met in either area I can tell that they have fundamentally different personalities.

 

[gleem]

Can you post any evidence to support your conclusion?

Not my conclusion.

https://www.msn.com/en-us/news/tech...-entire-tech-ecosystems/ar-AAZH3x5?li=BBnb7Kz

https://www.theregister.com/2022/07/08/semiconductor_engineer_shortage/

 

[Averagesupernova]

I have no opinions on semiconductor device development but I noticed a trend 30 years ago that there was a tendency for folks to be interested less in how hardware works and more interested in software in general whether it was firmware development or playing with software on their computer that I knew would be old news in 6 months.
-
That being said, I have a relative formerly employed by Rockwell who was an ASIC designer. I'm sure he would have an opinion. And for what it's worth, what I've just said can be considered my opinion and personal experiences. Someone else's observations may directly contradict mine.

 

[gwnorth]

If it's true, all the better for my son studying EE.

 

[gleem]

If it's true, all the better for my son studying EE.

Yes but warn him of the dark side.

 

[gwnorth]

@gleem the dark side being...?

 

[gleem]

@gleem the dark side being...

Being the issue this post started with, EE's becoming computer scientists not designers of electronic devices or systems.

 

[gwnorth]

Not much of a worry there. Had he wanted to be a computer scientist he would have gone into CS. EE is definitely where his interests lie, especially circuits and signal processing.

 

[Averagesupernova]

Being the issue this post started with, EE's becoming computer scientists not designers of electronic devices or systems.

It's a slippery slope. That's for sure.

 

[Vanadium 50]

Thanks for the reference. The plot that seems to be driving this is this one:

You will notice that it sums to 100% by construction. So if you doubled the number of EE's and increasaes the CS grads by 10, you'd see the EE numbers fall, even though there are more of them. Conversely, if the number of EE grads dropped by 2 and the CS grads dropped by 10, you'd see this graph shoot up, even though there are less of them.

Over the last two decades, EE enrollment seems to be almost flat: up around 15%, which is not too different from the US population as a whole (17%). "CS" enrollment has zoomed up, but how "CS" is defined is unclear, A lot of places have opened up so-called CS programs that don't teach a whit of computer science: they have classes like "web design", "python", and "network security" rather than "data structures" or "algorithms and complexity" and the like.

This may not be a bad thing, but the fact that a bunch of people are getting classes that lead them to entry-level IT jobs has very little to do with the number of EE's produced relative to the number needed,

 

[MidgetDwarf]

Thanks for the reference. The plot that seems to be driving this is this one:
View attachment 304595

You will notice that it sums to 100% by construction. So if you doubled the number of EE's and increasaes the CS grads by 10, you'd see the EE numbers fall, even though there are more of them. Conversely, if the number of EE grads dropped by 2 and the CS grads dropped by 10, you'd see this graph shoot up, even though there are less of them.

Over the last two decades, EE enrollment seems to be almost flat: up around 15%, which is not too different from the US population as a whole (17%). "CS" enrollment has zoomed up, but how "CS" is defined is unclear, A lot of places have opened up so-called CS programs that don't teach a whit of computer science: they have classes like "web design", "python", and "network security" rather than "data structures" or "algorithms and complexity" and the like.

This may not be a bad thing, but the fact that a bunch of people are getting classes that lead them to entry-level IT jobs has very little to do with the number of EE's produced relative to the number needed,

Cannot find a source, but from general experience it appears to be three things:

1)Poor math skills of US students as a whole. Since EE has more math, many transition to CS due lower math requirements on average?

2)Coding is being pushed in middle/high schools. Which may spark young student interest...

3)The idea of being rich. CS is advertised as a field of study that practitioners can make lots of money. Look at all the questions asking about data science of PF.

 

[gwnorth]

EE has more math than CS? CS is applied math. Now if you want to talk about Software Engineering, that's a different story.

 

[MidgetDwarf]

EE has more math than CS? CS is applied math. Now if you want to talk about Software Engineering, that's a different story.

If you look at the requirements at the average university, it is less math required for a degree. Not everyone goes to MIT or one of the flagship schools.

https://ecatalog.calstatela.edu/preview_program.php?catoid=66&poid=29963
Here is one the lower state schools in California.

https://catalog.cpp.edu/preview_program.php?catoid=36&poid=9570
one of the better state schools in California.

Im too lazy to post links, but yes, on average, BS in Engineering requires more types of "math".

You are from Europe, so the programs in EUROPE, may require students to take more math classes, or the courses themselves rely heavy on applied mathematics. Not so in the states on average.

 

[berkeman]

EE has more math than CS?

Absolutely. I studied both in undergrad, and never once used complex Fourier analysis or solid state semiconductor diffusion equations or other advanced math in my CS classes.

https://math24.net/complex-form-fourier-series.html

https://www.google.com/url?sa=t&rct...cture24s.pdf&usg=AOvVaw20nb_z9_LIi5W35BPVhogZ

 

[yungman]

Ok, this is my opinion and experience, I have no proof or data to back me up. Also, my degree was in Chemistry, never have degree in either EE or CS. I studying everything all by myself. Please excuse me for the long winded writing below. I literally started out as programmer and changed into analog RF EE.

I actually started more in programming mainly in assembly language back in 79 to 82. I had every intention to become a programmer. But after I designed my first project, I found programming is very tedious, nothing hard, just a lot of variables to keep track. The bigger the program, the more tedious it is. More importantly, technology was transitioning from 8bit Z80 to 16bit processor. Meaning completely new instruction set to learn. I realized that if I get into software, it's going to be endless learning and learning but nothing really challenging, just a lot of book keeping type. Designing hardware is not much more either but you have to keep learning new device as they came out.

In 1983, I was working for LeCroy(digital scope), I got a chance to work directly with Walther LeCroy(the owner) to design a new data acquisition system, I got a chance to enter the analog world. That was scary. I studied day and night to keep up. That's the first chance I got into analog electronics. It was challenging particular we worked on high speed circuits of over 100MHz. I mastered it, then I change job to design analog IC, then to Siemens to design the front end of their Ultrasound medical scanner. I since completely change to analog design. Later on, I got into RF and microwave design.

My conclusion is, analog RF is a whole lot harder to learn and understand. Particular RF, you really cannot understand RF until you studied Electromagnetics. You cannot understand EM until you study calculus to at least PDE which I did all on my own. I think a lot of students find it's very difficult to study math and EM, I can see a lot of students get discouraged by that. Also, studying Semi-Conductor and IC design is NOT a joke either. So many EE I met really are NOT good in analog, not to mention RF and Microwave. So they kind of drift to designed hardware and stuffs that are a whole lot easier. Problem is they need to keep studying or else they'll be obsoleted in a few years. It is easy when one is young, but as a person gets older, have a family, mind starting to be forgetful with age, it can be hard.

Bottom line, People drift to software or hardware design because it's a lot easy to get into. For real EE that is into analog and RF, the learning curve is very very steep. People get discouraged and drift to software and hardware. That's too bad, but good for people like me. It's always quite easy for me to find a job throughout the years, You pay the steep price studying, things really don't change much throughout the years, sure you have new parts, but it's much easier to keep up. Those knowledge are not going to be obsolete. Put it this way, I retired in 2005, in 2015, the company decided to design a new generation system, they called me to contract with them. I ended up working for a year and half, they even commented I did not miss a beat. Can people in software or even hardware say that they can jump right in after 10 years?

 

[symbolipoint]

EE has more math than CS? CS is applied math. Now if you want to talk about Software Engineering, that's a different story.

That has the potential for members to begin arguing Which One Uses More Mathematics? Electrical Engineering or Computer Science? If members participating go on that route, that part of the discussion may become unproductive. Just my guess.

 

[symbolipoint]

I actually started more in programming mainly in assembly language back in 79 to 82. I had every intention to become a programmer. But after I designed my first project, I found programming is very tedious, nothing hard, just a lot of variables to keep track. The bigger the program, the more tedious it is. More importantly, technology was transitioning from 8bit Z80 to 16bit processor. Meaning completely new instruction set to learn. I realized that if I get into software, it's going to be endless learning and learning but nothing really challenging, just a lot of book keeping type. Designing hardware is not much more either but you have to keep learning new device as they came out.

All that is something that most or many people involved or studying Science, Technology, Engineering, or Mathematics really should read and think on. But I disagree with what you say about "not being challenging."

 

[yungman]

All that is something that most or many people involved or studying Science, Technology, Engineering, or Mathematics really should read and think on. But I disagree with what you say about "not being challenging."

To learn programming, it's not hard at all. How you use the program is a different story. If you program in some scientific field that involve high math or physics, of cause you have to know those, but that to me is NOT about programming, it's about the particular job.

I just learned C++ for the fun of it when my grandson(CS major in college) said he needed motivation, I said how about grandpa nipping on his heel! I just picked up and learned C++ in a book from cover to cover(at least two semesters) in less than 7 months. It's not hard at all. I was going to learn graphics and animation, then I found out the library used in the book is NOT around, all the stuffs used for that are useless. I have to use SFML instead. It's like you learn to use one, then later on, if another one came out that is better(more importantly, the old one is not around anymore), you have to relearn their custom commands. Just a lot of simple stuffs to remember and they keep changing. You are at the mercy of what is available. To me, that's not challenging, it's just pure pain! Try maintaining an old program when things you depending on disappear on you and you have to migrate using newer available ones.

Hell, I learned C++ from an old book, I did not find out that the code is quite different from the newer C++17! If I learn the latest one, a few years later, it might change. Nothing hard, just keep changing.

If you consider this is challenging, I guess you are correct. But to me, this is nothing compare to studying PDE and EM. That, is hard. But to the best of my knowledge, these don't change, it's just as valid today as 50 years ago. You learn them one time. You might have to sweat it out and pull your hair, but if you understand them, you got it.

 

[CrysPhys]

With respect to employment opportunities, there are also two big pluses for CS over EE (or more specifically, many EE specialties): the capability to work remotely and the capability to work on flex time. These have become important factors not only because of the pandemic, but also because they make it easier to achieve the delicate balance between work and personal life.

There are many families now in which both spouses have high-level professional jobs (I'll leave this category intentionally vague; it doesn't matter). In the past, a better career opportunity for one spouse often entailed relocating. That would then create a problem for the other spouse if they are happy where they are. And also a problem if there are children of a certain age. With remote work, relocation is not needed. And for other families, there's the issue of childcare. Remote work or a flexible schedule (or a combo of both) make childcare much easier.

Are there opportunities in EE that provide remote work and flex time? Sure. But if you're working in a wafer fab or a power plant, probably not (or at least less likely).

 

[anorlunda]

You will notice that it sums to 100% by construction. So if you doubled the number of EE's and increases the CS grads by 10, you'd see the EE numbers fall, even though there are more of them. Conversely, if the number of EE grads dropped by 2 and the CS grads dropped by 10, you'd see this graph shoot up, even though there are less of them.

Correct. That makes the conclusion in the thread headline highly misleading. We may have a shortage of EEs, but not so short as to threaten extinction.

Especially CS may have attracted students who otherwise might have chosen a non-STEM major. The two best programmers I ever hired were an archeologist and a carpenter. The carpenter went on to manage a silicon valley software startup. A decade later, both of them might have been attracted to start by seeking a CS education.

 

[gwnorth]

You are from Europe

I'm from Canada.

I will concede however that the amount of math required depends on the rigorousness of the program in question.

 

[phinds]

Absolutely. I studied both in undergrad, and never once used complex Fourier analysis or solid state semiconductor diffusion equations or other advanced math in my CS classes.

except that my CS was in grad school and it STILL didn't use as much math as my undergraduate EE studies.

 

[StatGuy2000]

Absolutely. I studied both in undergrad, and never once used complex Fourier analysis or solid state semiconductor diffusion equations or other advanced math in my CS classes.

https://math24.net/complex-form-fourier-series.html

https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&ved=2ahUKEwiLnI6vnov5AhUSKkQIHfE5DqMQFnoECAoQAQ&url=https://www.ucl.ac.uk/~ucapahh/teaching/3C25/Lecture24s.pdf&usg=AOvVaw20nb_z9_LIi5W35BPVhogZ

I believe the amount of math that CS students take varies considerably depending on where such students study.

For example, at the University of Waterloo, CS students are part of the Faculty of Mathematics (one of only a few institutions in the world with an entire school/college/faculty devoted to different branches of mathematics), and are required to take a variety of math courses as part of its basic curriculum. At my alma mater (University of Toronto), CS students also had to take a variety of math courses as well, very much on par with what engineering students in general take. They may not necessarily be the same math courses (as engineering students tend to take more analysis and differential equation classes), but in terms of rigour or level of difficulty, they are not that much different.

I should also add that it is very common for CS students in Canada and the US to double-major with math.

 

[StatGuy2000]

View attachment 304622except that my CS was in grad school and it STILL didn't use as much math as my undergraduate EE studies.

The level of use of math in CS in grad school depends very much on the specific subfield of research in CS.

For example, there are many fields of CS research that are heavy users of math (e.g. theoretical CS, AI, scientific computing/numerical analysis). Other fields are less so.

The same could also be said for EE research fields as well.

 

[StatGuy2000]

As for the original claim being made about EE students opting for CS, I think to the extent this is true (and the extent to which this is true will vary depending on country), it is largely driven by labour demand.

Certainly it has been the case for a number of years that the demand for CS graduates (or more specifically, demand for software development and related work) has been far higher than has been the case for EE graduates in traditional EE areas of focus (e.g. semiconductors, power systems, etc.) So it should not really be all that surprising that more students may express an interest in studying CS vs EE.

It is also worth pointing out that the line between EE vs CS is not always clear-cut. There is an entire field called "computer engineering" which involves a blending of both EE and CS. Furthermore, many research areas within EE (e.g. signal processing, control theory, robotics, machine learning/AI, etc.) intersect within CS as well.

 

[yungman]

I still think CS is a lot easier to get in than EE particularly modern electronics are faster and faster and you cannot get away by "winging" it. Even digital logic is so fast you need knowledge like control impedance and knowledge to layout boards, not just using simple logic when things are still in MHz range. Hell, if they are flunking EM class and/or semiconductor design class, I don't think they have a choice but to move more into "computer" design which is designing digital logic and firmware. Lots of people are in this category.

As for being able to work at home, EE can do that also. Might not be full time at home, but definitely part time. So are CS if it involves hardware, you still need to test on system. Case in point, I contracted for 1 1/2year in electronics design working full time at home. I design, layout pcb and sent the file over in one email!

Sure, now the time is good, if you can breath, you get a job. Time can change and change very very fast. One down turn and all the jobs disappears, those "computer" jobs can be very cut throat as there are too many people apply for one job. I've been in the field for a long time, I've seen the cycle over and over. I was manager of EE, it was so hard to find someone good in analog.

 

[StatGuy2000]

As for the original claim being made about EE students opting for CS, I think to the extent this is true (and the extent to which this is true will vary depending on country), it is largely driven by labour demand.

Certainly it has been the case for a number of years that the demand for CS graduates (or more specifically, demand for software development and related work) has been far higher than has been the case for EE graduates in traditional EE areas of focus (e.g. semiconductors, power systems, etc.) So it should not really be all that surprising that more students may express an interest in studying CS vs EE.

It is also worth pointing out that the line between EE vs CS is not always clear-cut. There is an entire field called "computer engineering" which involves a blending of both EE and CS. Furthermore, many research areas within EE (e.g. signal processing, control theory, robotics, machine learning/AI, etc.) intersect within CS as well.

One other thing I would like to point out is to what extent students who may otherwise pursue electrical engineering (EE) be opting to study other fields of engineering/science besides computer science.

For example, what are the rates of enrollment for, say, mechanical engineering? Aeronautical engineering?

These 2 engineering fields specifically have some degree of overlap with EE.

Also, how many potential EE students may be deciding to study, say, physics? Or applied math?

 

[Vanadium 50]

I don't think "CS is easier" is all that on point, especially since the person who is advocating it a) doesn't have a CS degree, and b) required thousands (yes, thousands) of messages helping him with simple programs. The one part that is on point is that it is a lot easier for colleges to starts a "CS" program than an engineering program.

Bad engineering, and people die. Society has an incentive to monitor these programs more carefully.

 

[yungman]

I don't think "CS is easier" is all that on point, especially since the person who is advocating it a) doesn't have a CS degree, and b) required thousands (yes, thousands) of messages helping him with simple programs. The one part that is on point is that it is a lot easier for colleges to starts a "CS" program than an engineering program.

Bad engineering, and people die. Society has an incentive to monitor these programs more carefully.

That's the whole point. We all here thinking as adult, put yourself in the students' shoe. They don't necessary know what they want, a lot of them goes for the path of least resistance. I know, I have a grandson in 4th year college! By the time they are out into the real world, it's too late and they are stuck. CS is easy to get in in school. It's the path of least resistance. Don't think like an adult looking back.

In that sense, I am glad I got to choose my path when I was 28, so I can see a lot clearer. I also made the mistake to be Chemistry major just because I was good at it. I hated it after 4 years even I got straight A's on all the lecture classes.(not in lab because I always hate labs). It's the path of least resistance. What did I know in my college days!

 

[phinds]

I don't think "CS is easier" is all that on point, especially since the person who is advocating it a) doesn't have a CS degree, and b) required thousands (yes, thousands) of messages helping him with simple programs.

 

[yungman]

I don't think "CS is easier" is all that on point, especially since the person who is advocating it a) doesn't have a CS degree, and b) required thousands (yes, thousands) of messages helping him with simple programs. The one part that is on point is that it is a lot easier for colleges to starts a "CS" program than an engineering program.

Bad engineering, and people die. Society has an incentive to monitor these programs more carefully.

It's the path of least resistance, I'd find it myself if I have to. Not that I am not grateful, but that's just facts of life. I studied PDE and EM by myself also, I did not ask that many questions on the forum as the books were written much more consistence and clearer. Books are just better. Just the path of least resistance if there are people that already have the knowledge. As I said, I am grateful to find help here, but I would have made it on my own.

Case in point, I am just starting to review C++ from my notes. All the names(terms), definitions, all the cin.get, setw(), St.append...thousands of terms, each has a specific meaning, NONE complicate and hard to understand, just so many of them. Then there are exceptions, people keep changing throughout the time. It is by no means HARD, just easy to get mixed up. Unlike in EM and physics, "curl" meaning is very specific, once you understand, it's clear.

 

[Joshy]

I don't think EE is dying, but I do think CS is growing much faster than EE, and just like others have said the demand for it has been high. Higher demand also offers a very attractive salary especially for entry-level positions.

I don't think CS is easier or that you can just wing it. I think there are aspects of it that you can wing like if you just need to fudge a microcontroller enough to do some basic tasks, then sure you can wing it; the same could also be said for a CS student wiring up an Arduino with some LEDs, or buying a BT/WiFi module we wouldn't claim that they can wing RF/EE work. If I got a CS career down the road and had some real constraints for adding some robust function to legacy code from a 15 year old product still used today worked on by at least 100 previous engineers, then I might not want someone just winging it. I have seen plenty of people wing layout too some places hire technicians for that work instead of engineers (current place I'm at they hire people from retail stores for a lot of the positions it's their stepping stone into white collar careers), or a blend of so that the more advanced layout work is given to the engineer or falls under their guidance. One place I was at would not even consider hiring expensive engineers for the layout positions, and this was layout for flight hardware. Application notes for the IC that is used and even contacting Application Engineers can be very helpful too some people simply copy the layout recommendations from the application notes / datasheets although I don't think this is a good idea but design might be good enough to get away with because performance issues does not always equate to functional issues. Unfortunately: I see a lot of winging in RF/mmWave design too and because so few people actually understand it that they can get away with being wrong or their stakeholders are okay with paying extra for an over engineered fear driven design. I'm honestly becoming a little bit more embarrassed whenever I might mention to anyone that I work on RF.

 

[CrysPhys]

Bad engineering, and people die. Society has an incentive to monitor these programs more carefully.

But software increasingly is integral to engineering. Bad software, and people die (Boeing 737 MAX, autos with various degrees of software control, ...).