Scholarships, College, etc.

[syedhs]

Either will do ... personally I think a well crafted math heavy, hard core CS course (and I don't mean just Java/c# and basic data structures) will be more useful with less noise than CE.
I don't know if they still do that but in my time ( early to mid 90s) I had to write my own mini OS complete with BIOS and direct hardware access (I still have a boxed copy of WATCOM 10 with DOS/4GW ! ) which is more than enough to grasp everything a game programer will ever face ( at least in terms of understanding hardware.)

For that matter, I think it is true that if you study computer engineering (and execute such project), then the knowledge is fully utilized and well grasped. Mostly because you are probably doing low level programming such as assembly and knowing the C code and how it relates to the hardware (AX, BX etc) is so fulfilling. You can still complete the project satisfactorily without knowing the hardware, but it still better if you know the hardware.

Back to the topic (it is off-topic in the off-topic sub-forum ), I really think you dont have to go to very high standard university like MIT, CMU etc. But going into the mid-height (10% percentile) is sufficient.

[_tommo_]

At MIT they should still put you through the ringer, from what I understand. Ditto for Stanford. Neither are "Javaschools"

http://www.joelonsoftware.com/articles/ ... hools.html

However, a good low-level (down to the etching, IMO) understanding of computer architecture goes a very long way towards understanding the "why" of optimized software.

It depends on how you use java
The figures in my course at the first programming exams are the same "good old ones" that they guy wants...
this because you can teach recursion and pointers also in Java, not only in exotic or dramatically old languages.

For example they often asked us to build recursively a non-shared-memory linked list, which requires to understand references and recursion just like you did in C.
The only difference is that you don't destroy your pc if you fail
Anyway i'm not defending Java (that i don't like) but if the courses are easier for sure isn't a java fault.

[xavier]

If you understand Joel's article, he says as much -- it's not the language that's the fault, it's the way that programming and algorithmic analysis is taught; in this, the use of Java is more a result of outcome-based education philosophy than anything else.

And you can't destroy a computer with Lisp either -- you just run out of stack and the program dies.

Another good read on the topic comes from Christer Ericson's blog:

http://realtimecollisiondetection.net/b ... 82#more-82

(Also, follow the links to the "Design Patterns Are From Hell" entry -- more good reads)

[RedEyeCoder]

<off topic>

I read design patterns are from hell, it was entertaining The author makes a good point, but I feel it applies mostly to programmers who are susceptible to concerntrating on the solution before the problem (which could be many of us at times).

Design patterns can be very handy. Even when I ignore them I find I end up implementing something very similar to what is called a pattern, though I may not be aware of its existence at the time. Using patterns to break large problems into smaller ones is also another great strategy as long as the problem has been thought about enough. I find patterns to be much like a language where the parts are movable and are interpreted and modified by programmers to express a solution. Maybe I am oversimplifying it.

</off topic>

[xavier]

I made that point in the comments to that blog -- I also agree with your takeaway from it as well.

[nikki]

If you destroy your computer when you fail then the OS sucks.

[jjp]

If you destroy your computer when you fail then the OS sucks.

No OS can protect you completely from non-functioning or malicious programs Well, maybe something that works like Singularity could. But none of the current real world OSs.