although my views on Computer Science grads are more nuanced than this article on The RegisterI am too frequently disappointed with CompSci grads who don't know algorithms, or how computers work or who know no language other than Java.

QuoteKings College students/victims told me that they do operating system internals in Java, and no they weren't jokingC'mon, they were joking.

If you go to a decent school, computational foundations are part of the curriculum. I've wasted at least 2 semesters of my life doing proofs on complexity classes and other such bullshit.

QuoteOriginally posted by: DominicConnoralthough my views on Computer Science grads are more nuanced than this article on The RegisterI am too frequently disappointed with CompSci grads who don't know algorithms, or how computers work or who know no language other than Java.And I don't see nextgen CS knowing them either, apart from the fact that hardly anyone CS these days.There are different kinds of algos 1) numerical/application and 2) numerical-less/universal aka symbolic.CS tends to lean towards 2).In the olde days it was a given that you could program.CS is so theoretical if you don't have the engineering apps to back it up.

they teach Python, Java, SQL and HTML in the regular high school in US. things are cool over the ocean, my European friends, we'll pick up the slack

This sums up one of my pet peeves:I also curse the stunningly naive idiots that created most of the internet protocols because they are so pathetically easy to abuse.

QuoteOriginally posted by: halikIf you go to a decent school, computational foundations are part of the curriculum. I've wasted at least 2 semesters of my life doing proofs on complexity classes and other such bullshit.It's too bad you feel that way because it's an extremely important and practical topic.Unless you are writing code that will live a very short time in a very well-predefined environment, you need to understand how it will scale. What if we double the number of data points? What if it runs for a hour, a day, a week, a year? How big a server farm do I need to handle a million users?

QuoteOriginally posted by: Traden4AlphaUnless you are writing code that will live a very short time in a very well-predefined environment that sounds like a typical quant code

QuoteOriginally posted by: DominicConnoralthough my views on Computer Science grads are more nuanced than this article on The RegisterI am too frequently disappointed with CompSci grads who don't know algorithms, or how computers work or who know no language other than Java.Ok Dominic cannot resist. I agree with you that java / C# jobs are may not pay as much as the more exotic ones. But then there are plenty. On the other hand you could invest large amounts on time on CUDA ( I am) but the risks are high. By itself it is useless - you need to do numerical programming in it. But what if the future takes another course and CUDA falls through. Well the knowledge will be worth it. But the way I see it currently the future of computing looks directionally uncertain. It takes a lot of time to pick up exotic technologies and they may not do as well. Alternatively a lot of these exotic jobs go to Phd's , so if you do not have it you are out of luck.What is your take on something like this gaining traction. I am interested too as the type of work that is available in java , C# world is not interesting and the the type of interesting work you speak of is overly competetive where if you do not have he right qualifications you stand no chance of even getting an interview. Your response to this query will be very appreciated. Especially do you see as F# development combined with a MS in math/stats , with good knowledge of R and solid grounding in concurrent programming with CUDA likely to be valuable going forward (1-2 years in the future). In other words would the market for those skills gain a critical mass so that having the "right qualifications" is not a pre-requisite for getting an interview.

I have come across too many of these computer science grads who are very knowledgeable from a technical standpoint, but they insist on writing spaghetti code which is not properly documented. They need to pass a law which forces people to maintain their own code for 2 years after it goes into production. That is the only way to get some discipline in coding.

QuoteOriginally posted by: traderjoe1976I have come across too many of these computer science grads who are very knowledgeable from a technical standpoint, but they insist on writing spaghetti code which is not properly documented. They need to pass a law which forces people to maintain their own code for 2 years after it goes into production. That is the only way to get some discipline in coding.There are many reasons for this-1. Pressure to deliver yesterday.2. Bosses who do not care about code quality or worse still do not understand the value of code quality. These reward the fastest not the most disciplined coder.3. Politics - Even if you care about code quality if (2) above is your situation and you do not have influence in a given place you stand to lose if you insist on spending an extra day or week on refactoring code.Often times you cannot create perfect code the first shot. Refactoring when you reach a quiescent state between project phases is essential. But if you have another team member who does not give a shit having more influence with your boss , good luck with it. Its a political battle not an intellectual one.The problem is that in the recent times two types of bosses exist in the IT world-1. One whose deep expertise does not go beyond SQL.2. Another who thinks becoz he can write procedural code in Java - he is a fucking object oriented genius. The educated CS boss is not very common place in IT. Enough rant for the day. My boss btw is a CS graduate so the above does not represent my situation but it sums up the situation in my last job.

(double)

QuoteOriginally posted by: CuchulainnQuoteOriginally posted by: traderjoe1976I have come across too many of these computer science grads who are very knowledgeable from a technical standpoint, but they insist on writing spaghetti code which is not properly documented. They need to pass a law which forces people to maintain their own code for 2 years after it goes into production. That is the only way to get some discipline in coding.Indeed. For some reason it is almost impossible to get CS-trained people to document. Is this because they are code-driven and less design/engineering driven??What I also notice - not always - is a preference to re-invent the s/w wheel rather than employ a well-tested pattern. Why is this?

QuoteOriginally posted by: CuchulainnQuoteOriginally posted by: CuchulainnQuoteOriginally posted by: traderjoe1976I have come across too many of these computer science grads who are very knowledgeable from a technical standpoint, but they insist on writing spaghetti code which is not properly documented. They need to pass a law which forces people to maintain their own code for 2 years after it goes into production. That is the only way to get some discipline in coding.Indeed. For some reason it is almost impossible to get CS-trained people to document. Is this because they are code-driven and less design/engineering driven??What I also notice - not always - is a preference to re-invent the s/w wheel rather than employ a well-tested pattern. Why is this?Probably because they don't know the well tested pattern. In France a CS degree is about programming, designing new language and so on. So they know a lot of programming language but almost nothing on numerical methods. Another kind are engineer degree on numerical scientific computation, it's the contrary they know only a few programming language (example C++ and Matlab) but are far much more aware of the maths behind numerical problem solving. It's two different kinds of degree I think.

QuoteOriginally posted by: CuchulainnQuoteOriginally posted by: CuchulainnQuoteOriginally posted by: traderjoe1976What I also notice - not always - is a preference to re-invent the s/w wheel rather than employ a well-tested pattern. Why is this?I reckon it's really due to the fact that if you use someone else's idea work you'll end up getting less marks for your project.Also re-use is a real world concern involving money according to how long the project takes to be completed which is something that most undergrads are unaware of.