Please create an account to participate in the Slashdot moderation system


Forgot your password?
DEAL: For $25 - Add A Second Phone Number To Your Smartphone for life! Use promo code SLASHDOT25. Also, Slashdot's Facebook page has a chat bot now. Message it for stories and more. Check out the new SourceForge HTML5 Internet speed test! ×

Comment Re:Not bugs until they cause problems (Score 1) 266

In my experience, way too many programmers go for the obvious, short-cutting, direct, layer-breaking solution because it only requires writing a few lines of code. Out of an infinite set of possible solutions for the problem, they choose the one that saves them the most characters typed.

Experienced ones however will reduce the solution space and reject solutions that don't adhere to architectural criteria right from the start, and write something that looks a little more complicated but in fact results in a far more elegant, better scaling and maintainable solution.

Comment Re:A Very Old Performance Problem, Mostly Forgotte (Score 1) 266

So, I profile that code, I find, hm, odd, this loop is taking a lot of time.

It's accessing some array, wierd, array accesses are normally blazingly fast.

Oh look, it's a two dimensional array, that's something you don't see every day!

Let's play a bit with the code, hey, it's fast now that I'm looping through it differently... hm, I wonder how that thing is laid out in memory... oh, could it be that it is causing cache line misses / page faults / disk cache misses (yes, those abstractions are present at every level)?

Comment All it takes is a few tricks... (Score 1) 347

I've been in this business for over 20 years, and I think I can honestly say that seeing my own code from one year ago, every single year, would cause me to sigh and think: "I was so naive then..."

But anyway, I wrote this book, just follow it and you'll be a professional programmer after you finished reading it...

Comment Re:Because there's no such thing as one "performan (Score 1) 474

I think in part that the processes are now so complex, you actually needed the previous generation of CPU / GPU's to design the next generation. It be hard to skip generations if your current generation can't even run CAD software, let alone validate a design before putting it on a wafer.

Comment Re:One word (Score 1) 474

What would you parallelize though?

The code for UI interactions? If it responds and updates within 20 ms, my mouse will flow smoothly over the screen...

The validation logic? The business rules? The database access code? None of these benefit from parallelism as single thread performance is more than sufficient to return responses within *human* limits.

Parallelism is already employed at almost all of the stages where it matters, and it won't need to be written by the average joe programmer. They just use a library to play that sound or video (offloaded to GPU)... They just write a per request application, that scales by using a thread per request and if necessary multiple instances of the server. They just access that database, which then does a highly optimized parallel search for the results...

Teaching Joe to make use of the correct libraries and using things like WHERE and GROUP BY is a better investment.

Comment Re:W7-X is not a power source! (Score 1) 431

Wouldn't it be as simple as the containing vessel heating up, and you pump water along it? I mean, the plasma inside can be millions of degrees, and that would probably radiate to heat up the vessel as well when it starts producing energy.

A cubic millimeter of superhot plasma is not gonna heat up the entire containment vessel to melting point, but it might heat it up to say a nice comfortable 500 degrees orso...

Slashdot Top Deals

"In the face of entropy and nothingness, you kind of have to pretend it's not there if you want to keep writing good code." -- Karl Lehenbauer