dswp.de censored too ?

[natirips]

When I compare a typical racing game that runs with 10fps on lowest details in 640x480 and 16-bit color on a dual-core 2GHz CPU with an advanced graphics card and then remember that there were 3D racing games running at 40MHz 80386 with 8MB or RAM in same resolution and no 3D accelerated cards with 15fps, yes modern OO-based engines suck big time.

Need names and screenshots :P

I'm too lazy for screenshots right now, but I remember BC Racers ( http://www.mobygames.com/game/bc-racers/screenshots ) ran fine on a 386, while NFS-U (especially when drifting = especially on small maps o_O) ( http://en.wikipedia.org/wiki/File:NFS_U ... enshot.jpg ) had trouble on a 1833MHz Athlon (2500+) with Radeon 9200.

[Samtron]

When I compare a typical racing game that runs with 10fps on lowest details in 640x480 and 16-bit color on a dual-core 2GHz CPU with an advanced graphics card and then remember that there were 3D racing games running at 40MHz 80386 with 8MB or RAM in same resolution and no 3D accelerated cards with 15fps, yes modern OO-based engines suck big time.

Need names and screenshots

I'm too lazy for screenshots right now, but I remember BC Racers ( http://www.mobygames.com/game/bc-racers/screenshots ) ran fine on a 386, while NFS-U (especially when drifting = especially on small maps o_O) ( http://en.wikipedia.org/wiki/File:NFS_U ... enshot.jpg ) had trouble on a 1833MHz Athlon (2500+) with Radeon 9200.

I think this comparison doesn't work. NFS has much more polygons, details, effects and physics than BC Racers. NFS would never run (not even theoretically) on a 386, while BC Racers would run supersupersupersmooth on the AMD 2500+ (of course theoretically, it needs probably an emulator for the modern OSes). That NFS runs so slow on the modern computer is of course strange, perhaps the map is just bad or it has trouble with this hardware configuration and is rather optimized for Intel/NVidia.

In general I can sign SvaRoX' explanation: OOP makes programming much faster and easier. On the other hand OOP is not the best solution for all problems, sometimes strict procedural code might be the more efficient way.

[SvaRoX]

Natirips look at what you compare !


The first one is 2D game with poor (very) low-res "textures" whereas the second one processes gigabytes of data every seconds ! (yes gigabytes).
NTFSU ran fine on my Sempron 1800 MHz. If you had trouble with this game it's because radeon 9200 is crap, it can't handle enough 3D stuffs, textures and effects to run the game smoothly, and it has nothing to do with object implementation, only its lousy hardware...

[natirips]

I know where you're pointing, but that's still not a reason to ban games at schools. The other day I was in a school library and I couldn't open http://www.net.hr/webcafe/ to read a few jokes.

I know OOP is not total crap, it's just that I had to ventilate my feelings towards wrong methods of informatics teaching in schools.

We used Java to write 10-15 files of not-so-portable code where C could've done that in a single file without need for a VM, not to mention that stuff was usually much easier to implement in C. I often used to make my homeworks for OOP (Java in our case) class by first writing the program in C and then translating it to Java because it was easier (to me) that way.

P.S.: One more thing that has contributed to my point of view is: an average book on Java that I've seen teaches you everything but Java. I.e. you read 500 pages and you know everything about dogs, cars, space shuttles, aliens, politics, religions, ducks, pizzas and what not and you still don't know how to write a hello world in Java.

[Samtron]

Yes. Teachers love Pascal and Java, because they implement all the concepts of theoretical informatics and were early adopted by universities for teaching. And I know all the comparisons with iterators and apes and bananas and objects and cars and ships and it was also for me hard to start coding in the exercises. It's difficult to get the balance between theoretical concepts and practical knowledge, and that also depends on the appraoch pf peolpe and the program of study.

I have one prof who sometimes says "fuck OOP, let's hardcode the values in 32 arrays and use that, it's faster to implement and execute".

</end of nerdtalk for me today>

[SvaRoX]

Things much easier to implement in C, hmm I wonder what... Anyway programming is not only a matter of making things easier, it's also to produce readable, reusable and stable code (think buffer overflow). Reminds the "joke" : How many programmers are needed code an application in C ? 5, one people to code it and 4 people to understand the code six months later =)

I agree on Java books, since java is popular there are lots of crappy book to make programming accessible to beginners, then you read tons of pages about object and almost nothing about the code.

[natirips]

Things much easier to implement in C, hmm I wonder what...

A custom encrypting algorithm for encrypting a file (Java "byte" type is #1 most retarded variable type I've ever seen).