BounceBox Web Edition: a multiplayer online game made with Pulpcore.

[nouknouk]

Hi,

maybe some of you guys remember a (very) old post i wrote in ... march
2008. It was about an alpha version of a little casual game :
something like a pool game, but with a more 'arcade' gameplay.

The first version of the game didn't had any success: due to a lack of
'visibility' on the web, there was not enough people connected at the
same time.

Some time has passed since ; 1 year ago, I wrote again the same game
running on a totally different hardware: the 'Freebox', which is the
ADSL box provided by the french ISP called 'Free'. They indeed
developped kind of framework, allowing indies developper to write
their own programs and publish them on the 'FreeStore' (an equivalent
to the 'Apple Store' or 'Android Market' but dedicated to their own
box, connected on the TV).

Being the only one free online multiplayer game available on the
FreeStore, the game (BounceBox) quickly became a great success: the
game is only 9 months old, but more than 240k people registered and
more than 11 millions of rounds have been played so far.

Thanks to the success of the game, I setted up months ago kind of
'official website' ; and today I just released a second version of the
game : the "BounceBox Web Edition", which is directly playable in the
browser ... and powered by pulpcore.

You may want to have a try (*) ; text is in french, but the game is
easy to understand.

Here's the URL: http://www.bouncebox.fr/bouncebox-web-edition/


Cheers,

Nouk²



(*) Maybe don't try right now, as it's currently 3:00AM in France ;
there are more people online around 6:00 PM (french time).

[Gabo]

Nice game! i like it !!

[Andres Martinez Quijano]

Very nice! Congratulations!

Question: I had a mind a physics game like this one, and what popped
into mind was how to deal on the inconsistencies that might arise
between server and clients, for example, one client because of the
floating point architecture might render the trajectory in one way,
and the other client and/or the server in a different way, having
different boards... did you use strictfp? or how did you solve it?

> --
> You received this message because you are subscribed to the Google Groups "PulpCore" group.
> To post to this group, send email to pulp...@googlegroups.com.
> To unsubscribe from this group, send email to pulpcore+u...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/pulpcore?hl=en.
>
>

[nouknouk]

First, thanks for the support.

> how to deal on the inconsistencies that might arise
> between server and clients, for example, one client because of the
> floating point architecture might render the trajectory in one way,
> and the other client and/or the server in a different way, having
> different boards... did you use strictfp? or how did you solve it?

My specific case can be seen as the 'worst case scenario' as I have to
manage three completely different platforms:

- the last version of the 'Freebox' (i.e. the set top box) is based on
a x86 architecture and game is developped in Javascript and uses
SpiderMonkey as interpreter).

- the previous version of the 'Freebox' is based on a MIPS
architecture (still JS + SpiderMonkey)

- the Web Edition is mostly targetted for x86 platforms (PC, Mac), but
is developped in Java, not in JS.


Result: those three platforms give different results on floating point
computation (I have proofs ;-).

Thus I decided to *not* try to manage that problem and I use three
different servers, one for each platform. The drawback is that people
from one platform never play against people from other platforms. But
as I have lot's of players all day long (at least for the Freebox
versions, I'll look forward for the web edition), it's not a big deal.

Regards,

Nouknouk.

[Andres Martinez Quijano]

So between the different applet implementations (mac, win, linux) you
didn't have any problems whatsoever? I know there's a strictfp
modifier to force the JVM to do floating point in a strict way, but I
never used it

[nouknouk]

On Apr 19, 3:57 pm, Andres Martinez Quijano <tulsi...@gmail.com>
wrote:

> So between the different applet implementations (mac, win, linux) you
> didn't have any problems whatsoever? I know there's a strictfp
> modifier to force the JVM to do floating point in a strict way, but I
> never used it

I don't know. To be honest, I never tought there could be differences
between the same Java application running on different OS.

But if there is a 'strictfp' option, that means there's a reason, so
you're probably right.

Interresting question. If anybody has got the answer...