Java Server

[aleucht]

Hi,

I just wanna all let you know that a project Fibs has been introduced into sourceforge.com

The purpose of the project is to build an open-source clone (Java) of the FIBS Server

I'll upload the most recent source really soon, which can't do very much at the time. But multiple users can connect to the server already,invite each other and start playing. Missing are most other commands

So ,there is a lot do do and I wanna you invite you to join! If you intereted in participating in this project you have to register at sourceforge
                             

[webrunner]

I have looked at the pages which can be found here and here.

Just thought i mention the URL's  

Good luck on buidling the server.

[alef]

Out of curiousity, just how hard is it to build a backgammon server? Are we talking on the scale of making all the FIBS interfaces combined?

Surprised there haven't been more attempts, or is the complication having a longterm commitment to host and support it?

[klic]

Well, I can imagine that most people are shying away from the shere volume that's going through the pipe. Does anybody know how much gigs are going through fibs per month? I'd really like to see how this compares at todays rates...

[MadMatt]

Writing a FIBS-clone isn't that difficult and I'll wager it is more difficult to write a good FIBS client, due to the fact that the client has to deal with the whole user-interface issue.

In addition, writing a client for FIBS is difficult due to that "feature" of FIBS ignoring commands if they're sent before the previous one has been completed. Though that's perhaps more of a problem for bots who need to send a lot of commands in a short time. For example, when a tourney starts, TourneyBot needs to send notifications to every player in the tourney. TourneyBot handles the situation by listening for the response for the previous command before sending a new one. Since there's a possibility of something going wrong and a message or a tell not being delivered, it has to listen for every possible response to a sent command (that is, not just tell succesfully delivered, but also player not being there, player gagging TourneyBot etc.). That was a pain to implement even with the very small subset of commands that TourneyBot uses (it's mostly just message, tell, shout and time) but now imagine a real client... All of the FIBS commands, with all of the possible responses... bleh. I imagine most client authors simply ignore the issue because it rarely happens, however a really well written client would have to deal with it.

Anyway, my point was, it wouldn't be so difficult to write a clone. However, that would be a futile effort in my opinion. What makes FIBS popular is the community that has formed around it.  For 10+ years, people have been playing on it, writing clients, bots and webpages for it, organizing tourneys on it, discussing it on a board like this... And that's something that you just can't code up. It would take years for the new server to get a fraction of the popularity of FIBS.

I believe burper's idea to improve FIBS via a local proxy is the next best thing to actually altering marvin's codebase. Writing a clone has little sense now - FIBS was really written with plain telnet access in mind, its protocol is awkward for clients to parse and is severly lacking in features. So if anyone is considering rewriting FIBS, I would suggest to write a new server instead, one without the here mentioned technical shortcomings. Note however, that what I said about community still stands.

Cheers,

[burper]

Hi Axel,

 Thanks for posting here. Although I mostly agree with Mat (and fergy who has expressed a similar opinion), I am glad to know the project is still on your mind and apparently making progress. I was looking through the source code and it looks good, but the sf.net site looks long abandoned.

 Where I think this would be nice is on those rare occasions that FIBS is down. If you really have an update that works well for playing games,  perhaps fibsboard.com could host it as an alternate site? It might provide a nice transition for the community during development.

 I must say, that it feels weird putting time into fibs/clip compliant client-side stuff not knowing wether fibs will be around when it's done or not. An open source clone at least gives some hope that it's not a complete waste. If you update the code, I might just have to join in and help out!

[mucki]

Hi all together,

i think Matt is right when he writes that nobody will beable to code the community. Some one else wrote that it makes no sense to implement the old protocol on basis of telnet. I think this is right too. I do not know if the current fibs has lots of problems i  a technical point of view but a modern version on an open, maybe more effective protocol seems to make sence. Maybe with a failover with several servers. ;-)

On the other hand i started programming Java some weeks ago and maye it would be good study for me if the fibs is just cloned in Java. ;-))

Have fun,

Mucki  

[phr]

Hello everyone,

firstly, I want to point out that the Java Server Clone is a great thing - and I appreciate the effort already put in.

One disturbing thing about the real FIBS has a little bit to do with the protocol (although not exactly): The port number 4321 is really a pain in the ... . Reason is, that most (at least the good ones) firewall configurations won't let you out through that port. That makes it impossible to play with the usual FIBS clients from company/University intranets, etc.

I just started to code a client/bot combination for myself and am really happy to have a server clone for development that can talk to me on localhost or whereever on the intranet. I did not try it yet, but i will shortly.

A suggestion for the further development of a 'new' server might be that a new server supports all the old stuff of original FIBS, but offers other access on top of that. Be it other protocols (Http/WebServices or whatever one might think of) or additional features for clients.

This way we could save the community etc., be compatible to the old clients but also be flexible and open for future developments.

I feel that we have a lot of developers on FIBS, including myself. What do you all think about moving the 'clone' to a brand new fibs server that replaces the original in the future?

Greetings,
phr

[burper]

You might also check out FibsCompanion and/or TDTelnet (also on sf.net).
If you have a machine at home that can host TDTelnet, and it's properly secured, you *could* use it to proxy a port 80 telnet session to fibs at 4321, assuming your home firewall is open.

As far as local testing, I created something out of TDTelnet for ChouetteBot testing that  runs a script of server output for controlled test cases. If interested, let me know and I can post it somewhere. It's specific to Chouette/Box bot stuff of course, but might give you a framework for testing.

[adrian]

 Keep on the good work my friends, an alternative is always welcome.
I`m really sorry that I can`t program even a VCR.