The Declining Value of a CS Master's Degree (2013)
blog.regehr.org
blog.regehr.org
Basically your masters program is giving you a choice of 9-12 courses split between:
1. Algorithms .
2. Your choice of a primer in Java, C, Python or Swift .
3. Databases, Compilers, Networking, Comp Arch, Functional Programming OS, Distributed Systems .
4. IoS, Product Management, UX, Intro to Software Engineering, Advanced C++, OOP, Android Dev, Web Dev Etc. etc.
I really don't see anything on that list that is worth paying for a years worth of classes for. Half of that list you already take in undergrad, and the other 40% you should be able to pick up with a book and basic experience and not need a class.
Maybe 10% of those classes you would be at an advantage in the classroom?
Source: BS from UChicago
As is the case with the BS, you can learn all of the material from books and papers on your own without having to shell out thousands of dollars to a college.
The real money quote, though, really applies to both BS and MS degrees:
> I believe that the purpose has, in some ways, shifted to be a cash cow for the university in question.
Universities as a whole shifted their business model to focus more on revenue. This is a multi-decade trend that many people have discussed, and are currently debating in the context of the massive student loan bubble.
Nope, only PhD courses. In the US, Master’s programs are widely understood to be “professional” programs, i.e. geared towards private sector.
Bachelor's degrees are mostly focused on covering all the general stuff that no longer gets taken care of in high school, plus a couple of semesters of consuming the basics of your degree program. The Master's is basically an intro to research + some deeper dives into focused topics.
I think the article is accurate but not particularly useful. The value of all degrees is dropping over time as degrees inflate. It's not specific to comp sci. It's a reflection of our incredibly shitty primary education.
Universities simply don't have a choice. They can't keep the standards of admission they had 20 years ago, let alone 50. They just won't have any students at all. And each link up the chain of education has to pay for the sins that came before. High schools have to deal with kids who can't read. Colleges and Universities have to deal with kids who can't write. Grad programs have to deal with kids who don't know wow to research, and PhD programs have to deal with people who don't know how to produce new insight.
There are some exceptions, of course. But it's a problem the entire academic system is struggling with very badly right now.
Well... let's be realistic - this is true of undergraduate CS work, too. If you're going to college to learn how to program computers, you're probably going to be disappointed - the point is more access to experts in the topics who can check your work, offer guidance, and help you get unstuck if you get stuck.
The cause is simple, an MS is very advantageous when applying for an H-1B.
Not really. It's okay, I guess.
> Those who have the U.S. master's exemption have two chances to be selected in the lottery: first, a lottery is held to award the 20,000 visas available to master's degree holders, and those not selected are then entered in the regular lottery for the other 65,000 visas. Those without a U.S. master's are entered only in the second, regular, lottery.
That is definitely a non-negligible advantage.
I am not sure if you have any idea about the quality of STEM education in India but let me tell you it is far more pathetic than you can even imagine. Even I graduated from such an awful college and I know nothing about any area of CS. Nothing.
I am now 37 year old and working in non-software field but I now sorely miss the opportunity of getting good CS education and I wish I went through a decent college.
While your point may be a valid point (I have never been to US) but I do believe many enroll in those US college for good quality education, at least once. Hell, given a chance, I myself would enroll for such a course just for real education rather than H1B visa.
Anecdotally, the people on my team seem quite smart and capable, regardless of whatever quality of education they received, but we did have an intern locally who had a bachelor's from a school in India and was working on his masters at Oregon State University -- he knew so little about how to even use computers, much less write code that wasn't copy/paste from Stack Overflow, I was amazed that he was an actual master's student.
Our oldest son earned a Master's degree in CS from the same school where Regehr teaches and did the research path. He now works at Google. His program was the most difficult thing he has ever done and it prepared him well for his career.
H1B is a quota based lottery system. There is extra quota for MS graduates in US. They also have two chances to get an H1B while being in the US under an OPT work visa after studies. (2 years stayback visa)
Now after graduation, these students go to the body shops which are conveniently setup in those college cities. The body shops are usually run by IT managers in fortune 500 companies who is now a US citizen. (most of them came in the Y2K boat). The IT managers bill the body shops at very high billing rate, thus filling their own coffers.
For example: Take ASU, Arizona State University. Check the demographics of MS students there and the body shops around that area.
You get the idea.
More about the faux consulting companies here. https://news.ycombinator.com/item?id=4100451
I can top that. When I was doing an MS in CS at UT Arlington between 2005 and 2008, I was one of two US citizens in the entire program.
This was never the case in Europe, where if you have only a bachelor, people assume something went wrong. The target at university always was a Master's degree. The PhD, something you did for your mother, or if you wanted to pursue an academic career.
What? This is definitely not true in Germany, at least not anymore. Unless you want to go into research or teaching, it's commonplace for graduates to start working right away.
I think a much bigger problem that contributes to the devaluation of university degrees is the fact that students that don't really need a degree still get pushed to attend university - and that's definitely because of current recruiting practices where you really have no chance without at least a Bachelor's. I've met so many people in university who only really wanted to code and got no use out of learning things like automata theory, complexity, higher mathematics. Not saying it isn't important to learn these things, but you gotta differentiate between coders and computer scientists. And if coder positions require a computer science degree, well... Of course you're going to end up devaluing those degrees.
Edit: Adding to this, I attended university for 4 years before switching to a more practice-oriented college. In those 4 years I had one programming lecture (which only introduced three different language paradigms and didn't go into depth) and a system programming lab (admittedly, this was really cool). Everything else was mostly theoretical computer science and mathematics. Now, I chose this university because people told me of its good reputation and I felt like I had to attend a prestigeous school to get a good job, but it didn't really teach me any practical skills at all, so even if I had graduated I would've had a hard time landing that "good job" due to complete lack of experience.
I always have to explain that a Licentiate[0] is more akin to MS than Bachelor, which is a bummer on employment forms that don't expose it as option.
After Bologna, Portuguese universities started giving the option for former Licentiates to request their MS title, given the split into 3 + 2.
Which according to my knowledge also happens in other European countries where CS degrees used to be 5 years duration, before Bologna happened.
No, actually it's not. Germans are under this impression because it's considered equivalent inside Germany.
Looking at the contents of 3 years bachelor + 2 years MS, versus the contents of old Licentiate 5 years duration in Portugal, they are clearly equivalent in content.
Hence why after Bologna many universities basically renamed their Licentiate degree into something called Integrated MS, which means the students get a BS + MS paper at the end of those 5 years.
Also the reason why former Licentiates are allowed to basically acquire their MS title, by paying bureaucratic fees and in some cases presenting some kind of career report as thesis.
The actual program didn't change, just the diplomas you get in the end. Based on this, I'm comfortable saying that an MSc given these days is essentially the old 5 year degree.
I was studying physics in Germany right when the transition from 5-year diploma to 3+2 BSc+MSc happened. While most parts of the curriculum stayed the same, faculty took the opportunity to modernize the curriculum. For example, Computational Physics (an introductory programming course with a focus on numerical calculation methods and how to deal with precision-induced numerical errors) was upgraded from optional to mandatory, given that nearly all physicists have to write some code these days.
In CS, this means that you were taught courses ranging from basic Electrical Engineering all the way to Artificial Intelligence, including physics, algebra, calculus, signal processing, theory of computation, algorithms, databases, computer architecture, compilers, computer networks, robotics, etc.. That is, you got introductory-level courses on everything that makes "computer-stuff", and then there were a number of optional courses where you specialized.
I really don't think you can cover all that material in 3 years...
In hindsight, I don't think having a Master rather than a Bachelor influenced my job opportunities: but having a degree at all definitely did, at least for finding a job in established companies.
Most bachelor degrees are 7 semesters, diplomas were 8 semesters.
Ofthen they just cut out a practical semester and that's it.
In mathematics Diplom (Diplom-Mathematiker) was 9 semester; now BSc + MSc is 6 + 3 semesters.
In computer science Diplom (Diplom-Informatiker) was 10 semester: now BSc + MSc is 7 + 3 semesters.
Both for my university in Germany.
Also, education itself is much less a business here than in the US.
In countries where job market can be tough, such as Southern Europe or Poland - where I'm from, and where I've observed this first-hand (in Italy, too) - going to uni is a common way of deferring the unpleasant clash with economic reality.
It provides an excellent "alibi" for a few years' worth extension of parents' financial support. "We value education in and of itself" attitude certainly comes in handy ;)
You are simplifying it a tad bit too much.
In Italy Univesity course is not "free", it has a low cost (taxes and fees + books), but - unless you (with your parents) already live in the city where the university is, there are transfer/lodgement costs.
There was an article just the other day on the Corriere della Sera about the situation, I assume you can read Italian, otherwise it is not that bad via Google translate:
http://www.corriere.it/dataroom-milena-gabanelli/quanto-cost...
And - with the exception of very few specific and "technical" courses the issue is that once you have spent these 15.000-45.000 Euro you are not going to receive - if you find a job - an increased wage when compared to someone without a degree or with an undergraduate degree.
Of course the main underlying reason is that there is a high unemployment rate, and "on the market" an "average" degree has not much increased value.
Possibly something similar happens for CS, firms have "basic" CS degree as a requirement but do not value much an additional Master.
OK, so not quite free. I stand corrected, thanks :) In Poland anyway there's no tuition fee - at public universities, of course.
> unless you (with your parents) already live in the city where the university is, there are transfer/lodgement costs.
I should think this goes without saying... No free lunches for students, either ;)
To give you a rough approximation, for taxes there is a complex method of calculation, depending on the combined income of the family, loosely between 700 and 2,700 Euro per year, and the books may be anything between 300 and 1,000 Euro per year.
At my current university in the Boston area, for example, yearly tuition is around ~$50,000 USD, and that doesn't even include room & board; I'm paying an additional $1,400 per month for a private room in a 2bed/1bath.
With that in mind, the original comment definitely still offers a perspective I didn't consider before in terms of the cost difference in relation to the arguments in this thread.
What you point out as an issue, I see as a sign we have reached a good point in human history. (a good point, but not the end, of course issues exists, but as long as we don't mark them as "Working as intended" or "WONTFIX"...)
I'm an utopian of course. And of course someone can, and will, have a different opinion.
Hell, when I was studying for a masters, I felt that it was essentially a half-way course towards getting someone ready for a career in academia. Others also used it as a way to get a good university name on their CV so big companies would look at their applications.
This was never my impression during my university years, nor in working life (both technical and managerial).
If anything, it was the opposite, people I worked with would consider anything better than Bachelors in CS on a candidate’s CV (eg MS or PhD) as a signal that they wasted too much time in their life before starting with real work in the industry.
[edited for clarity]
I'm also in the UK, have an MSc (STEM subject), and FWIW have never come across this attitude.
Back in my first job (based on campus at CIT) they did prefer BSC's (with a first from the right uni, I think Harvard might just have made the list) for new recruits with the idea that they would do a masters /phd at the organisation.
So far no one (not recruiters nor employers) has so much as even asked me about my education. But I guess I have enough experience now that it really doesn't matter.
I'll throw in my 2 cents - this is not true at all for Estonia for example. It might be in some fields of study that I'm not familiar with but definitely not in CS.
Anyway, I've never had my lack of a Master's come up in Europe. Maybe it's because I'm American.
After I presented my BSc project my supervisor offered me support with PhD in computer vision and AI (area of my work and his specialisation). I had to refuse as during my studies tuition fees for PhD went up over 9x times (and for BSc almost 11x). On my first year I worked with people who paid ~£300 per academic year, on my final year I worked with others who paid almost £10.000 per year. "Something" went wrong.
Do they have sources to back this kind of snooty claim up?
Anecdotally, the only time I’ve come across this attitidude is when individuals heavily-vested in academia are attempting to stake out the moral high ground.
I started university in 2007 and among ~180 people studying CS with me, majority continued with full-time master studies immediately, and most of the others did it part-time (either immediately or after 1-2 years break).
These days it's different, it becomes more common to either finish formal education with bachelor degree, take a long break or continue in another, related major. 10 years ago it wasn't the case.
https://www.telegraph.co.uk/education/universityeducation/83...
The MPhil/MEng are absolutely real qualifications, requiring an additional year's study and a substantial research project.
If people are using it to be misleading, that isn't really Cambridge's problem -- anywhere it matters, people can ask for proof (the transcript), and there is a fair chance the person is being misleading in other ways as well.
[1] Example BA certificate: https://i.stack.imgur.com/VTWUb.png All undergraduate degrees are the Bachelor of Arts, regardless of subject (except Law, which is the LL.B.).
It is safe to say that the systems are a bit incomparable, since a 3 year degree in the states would be just as comparable to an associate’s degree.
The reasoning goes: the reasons why we need degrees are certification and reputation. By certification I mean that we only let people who we all trust practice medicine, as that kills people directly otherwise. So we restrict the people who can to those with the appropriate degrees.
The second point is more relevant to most other careers. A degree is nothing but a more reputable institution 'lending' reputation to its graduates. The reason this is necessary is because nobody's got the time to verify whether every person knows Computer Science - we're better off just knowing that people from University of X know CS if UoX says so. Even if you're going to personally interview every candidate for a job opening, still relying on degrees helps you filter candidates.
The thing is our interactions are recorded to the level that we can now reason about individuals in a way that was not possible before. Why do I need to trust UoX if I can mine your entire internet presence with reasonably comparable effort?
I thought from the title that, since CS is the field that is making this shift possible, we'd be the first one to suffer this influence.
As well as being an opportunity for someone to have time and freedom to study for a period of years university also provides a safe world to learn other soft skills and grow as a person. It's not prerequiste to develop those, of course, but it really helps.
That is why in Germany such degree programs (medicine, dentistry, veterinary medicine, food chemistry, law, pharmacy, teacher's training certificate, forestry) end with Staatsexamen (state examination), which is a special kind of degree that is much more tightly controlled by the state.
---
edit: Another thing is that I had a chance to dip my toes a bit into doing research without being in the PhD long haul. This was also very satisfying: reading papers is no longer intimidating and ended up being very helpful in my area (databases / systems).
Re: classes, yes the basic curricula always seems repetitive for folks who took a CS undergrad, but it's up to you to convince your advisor or graduation committee or whatever that you can and should skip basic requirements and instead tailor your major to more advanced stuff that you want to do.
The first year is identical for both, the second year is the specialization in one of the tracks. Everything's very project driven, you work remotely with other folks on almost everything you do IIRC.
Even the application to the program is somewhat similar to applying to a company: take home project, in person interview etc.
I'm doing mine right now because 1) it is convenient now with online courses, and 2) my company is paying for it. Credit to HN as well because I probably wouldn't have noticed the Georgia Tech OMSCS if someone here had not mentioned it.
I will enjoy having the certificate on my wall as another accomplishment, and GT is even a halfway prestigious school to have a degree from, but I'm under no illusions -- at this point in my career it has no real monetary value. It's purely for my own satisfaction.
I enjoy school, learning, and interacting with teachers and classmates so my personal value was high even if it didn't mean more money.
Same here. Back in 2004, I found myself surrounded by Indian folks who all has master's degrees from US universities, and here I was with nothing but my lowly little bachelor's degree. Since my employer had a tuition reimbursement program anyway, I went ahead and did a master's program - it was free, and what could the harm possibly be? More education is always better than less education... right? That's what made that Aline Lerner blog post such a gut-punch when it first came out - I had never considered the possibility that somebody could be evaluating me on paper and reject me because I had _too much_ education.
As someone who went this path, I'm troubled by the amount of assumptions and generalization being made here.
To my knowledge any legitimate MS program will require that you have completed an appropriate amount of foundational CS courses before starting the program. I obtained an MSCS after getting a bachelor's in a different field, and I had to take 6 CS classes to satisfy the prerequisites. Including the MS the total amount of coursework I completed was slightly greater than the requirements of the university's Bachelor's degree, plus some graduate-level research.
Granted, this is not as much coursework as someone who obtained both a BS and and MS, but from my understanding if you're not planning to pursue a PhD (or need it for immigration purposes) getting both is usually not a good use of time anyway.
It seems he is trying to specifically criticize programs that fall far short of what I went through, which seems valid. But it would be nice to make that distinction instead of writing us all off as inferior.
In fact, I went back for an MSCS degree ten years after finishing a BSCS, and even then I was required to take a couple of foundational undergraduate courses before they let me start my graduate coursework, since there had been enough advances in CS theory since I graduated a decade ago that they felt (correctly) that I needed to catch up.
Now OMSCS is riling up a few heads here and there which it will. At 7000 USD its a fantastic way for a working professional like me to advance my knowledge. You think these kind of MS courses devalue the degree? These courses are not THAT easy to do and this is 2018, there are fresh disruptive ideas in very field , why not Graduate school ? Hey with enough of these courses you will not need the evil H1B sith lords eating up the jobs here right? Its a Win Win. The market will then decide whos the more worthy MS student they wish to hire.
Here's my situation:
I have a PhD, but in a different, non-CS field.
Due to my research and experience, I have a ton of stats and programming experience across a range of languages, from lisp to C++ to javascript to python, to name just a few. I have published stats papers for example.
However, I kinda find when I'm looking at jobs, there's a heavy focus on CS grads, or people with finance-type degrees, or engineering degrees, or some combination. And to be honest, there are certain things out of a BS in CS that I probably am not that familiar with, especially low-level hardware type stuff, and network stuff.
So, I look at what's out there, and getting a BS in CS seems a little odd, for various reasons, in part because I feel like there's a lot of redundancy with what I have from my career and education already.
A MS in CS, on the other hand, is kinda not so strange to me. I might have to take a couple of preparatory courses, like in OSs, but doesn't seem too grueling to me.
So I read a thing like this, and it comes across as similar to the sort of gatekeeping that I bump into, like if you have a PhD in another field and then get an MS in CS, you aren't a "real" CS grad or something. I've supervised honors CS undergrads, and believe me, I feel pretty confident that my skillset is beyond many of theirs, at least in certain areas. At some point it just seems absurd to me.
I share some of the implicit concerns about cash cow MSs, that seem to be part of the current employment climate, not unique to CS, but a symptom of problems with hiring practices and also contributing to it. But it also seems misguided to me to start arguing that someone who completes a MS in CS isn't as skilled as someone with a BS.
There's just so much gatekeeping and pissing contest stuff that goes on in STEM, and it drives me crazy sometimes. It is maybe possible that that english grad actually understands programming, math, and algorithms at a pretty high level (I'm not an english grad, just making a hypothetical argument).
That is because too many people are insecure and evaluating who is good and who is bad is very subjective. We like to pretend it is objective, but the older I am the less I believe so. Moreover, programming is not one job, but multiple different jobs that require different skill sets, different aptitudes and different personality types.
Combine that together and you have perfect environment for bullshit gatekeeping and pissing contests.
Erosion of the value of the MS is not isolated to CS.
Forget some universities eroding the MS, entire countries are eroding it. If im not mistaken, Italy is a massive supplier of MS degrees.
But I'm not entirely convinced that it's appropriate to argue that the only "true" MSCS is one that requires a full bachelor's in that subject as prerequisite. That seems... overly restrictive.
What if you got a bachelor's in math or engineering and want to further specialize in CS? Should your only option be to complete another bachelors degree? What if you did a minor in CS? There's lots of people in these categories that can and do get an MSCS, even from very reputable programs.
Obviously coming from a related field will prepare you better than say a liberal arts degree, but where do you draw the line?
In my experience most MS programs have a certain set of prerequisites that generally equate to a minor in CS.
I’ve had too many technical interviews where the hiring manager will ask me questions about his favorite ORM or dependency injection framework and turn me down if I never used them before.
If I would have known that people in this field are so quick to disregard credentials and previous work experience I would have chosen a different technical subject to major in.
My take on a master's degree is that I treat it like a year or two of experience: experience that the poor fool didn't get paid for. Of course, a person with an MS is going to be a stronger candidate than one with just a BS, since they've had more experience. But I can take the engineer with the BS and give him the year of experience myself.
Perhaps it's different in Europe: they seem to value high education credentials more. I didn't care about the status of an MS and neither did most of my colleagues.
I think the only reason to do an MS in the USA is if you're a foreigner trying to get residency through an H1B visa. There's a separate quota category for people with an MS degree earned in the USA.
I came here from a 3rd world country that lacks a good education; Most of my CS experience and knowledge, at that time, came from the Internet, books, and personal projects. After working hard, I was accepted to this university as an MSc student.
I have never learned this amount of quality content as I've been learning in these past 2 years in this MSc program. Deep topics and subjects like programming language theory, low-level knowledge in distributed systems and machine learning, all this being taught by excellent researchers in their respective fields.
This was the first year, the course requirements part of the MSc program. Now I'm halfway through my research, and again, I've been learning tons. I've been writing very interesting code, building amazing things, writing papers, and working on my own ideas. I've never been this close to amazing scientists and engineers. This is priceless, in my opinion.
I've worked in the industry (before the MSc. program), it was super great. But I can confidently say that the MSc. program has made me a better engineer and taught me how to do proper research. I'm sure this may vary from university to university, department to department, and most importantly, from advisor to advisor. But that has been my experience!
I did a 4 year BS CS degree and graduated in 2005. I did calculus, linear algebra, real analysis, mathematical statistics, etc. until 3rd year and then the typical CS curriculum of compilers, distributed systems, programming languages, artificial intelligence, etc.
I've been working in industry for 12 years doing typical systems/software development, but got involved in more "Data Sciency" things in the last 2 years. I liked it so much that I decided I want to change career directions and move away from just "building software" to more "building software that runs on data as the fuel", like machine learning/deep learning/NLP related problems.
So okay, I have no real problem with doing anything that involves manipulating the data, building pipelines, or implementing things in Tensorflow or Gensim, etc it's just programming. Easy, peasy. I also understand most of the mathematics, and what I forgot was not too hard to "re-learn" very quickly.
The problem is it is difficult to apply for jobs as a Data Scientist when you only have a BS CS, even with real world experience. I'm not "formally qualified" as a Data Scientist.
So I enrolled for a 2 year course-based MS (1 year course work, 1 year dissertation). Basically it allows me to slap on my resume that (a) I have done university accredited courses in mathematical optimization, machine and statistical learning, deep learning, ethics, big data, etc. and I suppose the dissertation says that I can do more advanced work on data in an academically rigorous fashion. So basically I'm using the MS to change career trajectories 10 years after doing a BS CS.
I suppose you could do the courses through MOOCs but I appreciate the dissertation part as I am forced to read journal articles and learn how to do things more "rigorously" than I would otherwise in a plain old MOOC.
Isn't this a viable option for pursuing an MS? I don't think it's necessary directly after an undergraduate, but after a decade or so you might want to "specialize" into something that didn't really exist when you did your undergrad. BTW, I did this part time while working (you can kiss away your weekends and evenings).
None of which asked for proof.
I think my bachelors was too besides the social experience of college (I’m from a small town).
Most of what I learned was already familiar to me from developing as a hobby and I learned way more from studying stuff I was interested in than bubble-sorting and SAP case studying my way into boredom for an extra year.
Whether the degree was worth your time and money, and whether it was interesting, is an entirely separate question, of course.
Either he had very easy school or picked courses he already knew about or he splits what he learned into "important" (the ones he knew) and unimportant (new stuff).
You don’t need school if you don’t need a degree for your job.
Learning is my all time hands-down favorite hobby. If you put in the work you can learn all the things you learn in a CS degree on your own and in a much shorter time.
The problem is getting good text books since they’re expensive (when I was in school you couldnt google everything).
I realized I could social engineer my way to free books by appropriating college letterhead, pretending to be a professor, and writing to publishers asking for evaluation copies of books I wanted to read.
If employers verify the knowledge and not the degree - then the degree is worthless.
My degree is worthless. I’ve never been turned down for a job, never fired, and I’ve had a great career and have a great job currently. NO ONE has ever asked about my degree.
* in CS. Obvi it’s important for some careers like medicine.
I’ve been working as an embedded software developer for 7+ years at different #1 in their field companies without a degree. I’ve worked with a lot of newly graduates and the above is my experience.
The best technical solution is not always the best solution.
The rest seems pretty valid, although at this stage, I doubt if there's any well-founded way to determine whether a given degree will prove to be worthwhile in a purely financial sense, if there ever was. With rising costs it's probably just somewhat less likely...
I don’t have much interest in a CS MS. I don’t really see the point of it. If my degree goes well and after a few more years in the field I decide I want to specialize much further, I’ll just get a Ph.D.
My $.02
The reason many M.S.'s are failed Ph.D.s is that is easy to get funding for a Ph.D., but not for an M.S.
This article gave me pause, since I've never heard much skepticism expressed about a Comp Sci MS before. Is it really so overrated?
This is pure “appeal to authority” as the author is a CS professor-—better to discuss Aline Lerner’s research which he references
Discussion at the time (over 100 comments): https://news.ycombinator.com/item?id=5786720
But I have no hopes of instant value.
I guess it's good to have it when I'm older to get into a management position or something
This isn't really a brand-new thing either. One of my best instructors was a seasoned industry vet (former senior architect at a major firm, member of W3C and author of several of their proposals, etc.), and he had a bachelor's in Physics followed by an MSCS... from Stanford. Are we to say that his masters is not legitimate or that he is an inferior engineer because he technically did not complete a bachelors in the subject first?
To me it seems to be a problem of defining in general what the purpose of a masters degree should really be. In some fields it is used as an extension of a bachelor's in the same field, while in others it is used to supplement undergraduate education with skills in a different field, typically for the purposes of better rounding out the student's knowledge in preparation for pursuing certain career paths.
For instance, consider a professional degree like the MBA, where you can come in having a bachelor's in literally any subject.
What the author is noticing is that universities are increasingly offering the MSCS as something more like a professional degree instead of solely as an academic-track stepping stone to the PhD. And clearly he does not like this.
This issue seems to highlight the dual-purpose nature of universities as vehicles for training both practitioners as well as researchers. On the practitioner side, I think there is immense value in being able to supplement a bachelor's degree in one subject with a different set of skills obtained in a second degree. E.g., many people take the track of getting a science or engineering bachelors and then an MBA afterwards. The combination of those two disciplines can provide a very powerful skill set.
What seems to be happening is that the increasingly tech-centric nature of our economy is causing the skills offered by CS to start to make a lot of sense for some people as a form of secondary training. Much like business, technology intersects virtually every industry, and those who have training in a specific subject domain in addition to CS can be highly effective problem-solvers, much more so than those trained in just one or the other.
Sure, perhaps we can argue about whether an "MSCS" is really the most appropriate vehicle for such people. But being one of those people myself I will say that my experience does not at all reflect what the author claims (see my other comment re: coursework). The reality is that there are an increasing number of people who want to supplement their undergraduate training with CS, and this is a good thing. The author is correct in noting the distinction between this demographic and those who are on a research/academic track, but I'm just not convinced that we can definitively say that an MSCS must always require a bachelors in CS.