Re: [simspark-devel] rcssserver3d in simspark SVN, what's next?

[Hedayat V. <hed...@ai...>]

Hi Mahdi and All,

/*Mahdi <zig...@gm...>*/ wrote on ۰۸/۱۱/۲۹ 01:55:00:
> Hi
>
>   Good work Hedayat! On the new protocol you told about, how would you 
> propose the new protocol to be? Please tell what features it should 
> implement, and how would they be, and we can all put it in discussion 
> for its better performance and usability. What about the previous 
> protocol? Would it be eliminated or just be present in the simulator's 
> code? Any suggestions?
:) Thanks, but I'm not talking about a new protocol (agent-server 
protocol). My idea is to integrate our server with Player[1] (probably 
by providing a plug-in for player, providing virtual hardware), so that 
agents will use Player client libraries to communicate with the server 
using Player protocols. Since Player protocol is being used on real 
hardware, I think it should have good performance.
I think the current protocol should be present in the server while the 
new protocol is under development, but after that, there is probably no 
reason to maintain the old protocol, except for compatibility with old 
agents. (We might remove it completely after a compatibility period.)

>   And one thing more, what about the protocol between simulator and 
> visual tools like trainers? Currently ZigoratAssistant works properly 
> (on my PC for now). I'm currently creating documents for developers 
> and changing it to an automatic make system. The only part which 
> remains uncoded is the simulator protocol parser, which I'm not able 
> to write it, But I know some people has previously worked on this 
> protocol and have previous experiences in documenting it! I'm 
> currently composing a developers guide manual in LNCS format to 
> provide full information and steps for contributing to this project, 
> for both code developers, and for graphics developers or artists( 
> People who want to know how to create graphical objects like stadiums, 
> grasses, balls, goals, environments, etc and how to create visual 
> effects like shaders, composite materials, environmental effects, 
> etc). If anyone can give a hand on these matters, that would be 
> awesome. Tell me, and I'll send a most recent version of the manual.
Great work. :) For the protocol, I think you have some ideas, so maybe 
it's a good time to tell us about them, and documenting the protocol, 
and waiting for somebody to implement it!

>
>   I can see a good year for the league if everyone who can contribute, 
> would actually contribute.
:) me too!

Good luck,
Hedayat

>
> Cheers
> Mahdi
>
>
> On Thu, Nov 27, 2008 at 5:31 PM, Joschka Boedecker 
> <jos...@am... 
> <mailto:jos...@am...>> wrote:
>
>     Hey guys,
>
>     thanks for starting the migration to the new SVN!
>
>     I just wanted to tell you that I created a new mailing list for the
>     SVN commits:
>     https://lists.sourceforge.net/lists/listinfo/simspark-svn
>
>     Please let me know if anything doesn't work as expected.
>
>     Cheers,
>     Joschka
>
>     On Nov 26, 2008, at 11:13 PM, Hedayat Vatankhah wrote:
>
>     > Hi,
>     >
>     > "Yuan Xu" <xu...@in...
>     <mailto:xu...@in...>> wrote on ۰۸/۱۱/۲۶
>     > 08:00:38:
>     >> Hi Hedayat,
>     >>
>     >>
>     >>> Yes, maybe (I think they should talk about this to MC! But they
>     >>> are silent
>     >>> for now ;) )
>     >>>
>     >> Then, we need to ASK them.
>     >>
>     > OK, I've no problem with asking them!
>     >
>     >>> And about myself: I'll hopefully work on a new protocol for agent
>     >>> (and
>     >>> probably monitor) communication (most probably through Player
>     >>> integration),
>     >>> migration to CMake (I've found some comparisons in favor of cmake
>     >>> but none
>     >>> of them were up to date. If anyone has anything against it, please
>     >>> let me
>     >>> know) and implementing the Sync mode; in no particular order :P.
>     >>> (My main
>     >>> goal is player integration for now).
>     >>>
>     >>>
>     >> Good plan. I think the protocol is very important.
>     >> Although the S-Exp has some advantages, but it is too slow,
>     >> and especially it can not handle image data! This makes me in
>     >> trouble.;-(
>     >>
>     > It seems that it's priority is a bit higher than what I thought!
>     >
>     > OK, what about the others?
>     >
>     > Have a nice time,
>     > Hedayat
>