jacobian, new computer bridge protocol (play against robots free)
Peteris
I would like to announce the creation of jacobian
http://peteriserins.com/2011/08/14/introducing-jacobian.html. From the
blog post:
"Jacobian is a simple, incomplete, but working protocol for computer
bridge programs. With the host program, you specify a deal,
vulnerability, contract and the four players by specifying their
executable programs. The host then leads them trough a game of bridge
(currently, only the play of the hand is implemented). That’s it."
The actual software components are available here:
* host program: https://github.com/Pet3ris/jacobian-host
* GiB wrapper: https://github.com/Pet3ris/jacobian-gib-wrapper
* GUI: https://github.com/Sedols/jacobian-GUI
Comments and contributions much appreciated. What do you guys think
about this?
--
Peteris Erins
Carl
If I understand this correctly, it is a program that lets computer
programs (or people) play one seat position per computer/human, and
Jacobian manages the four playing units.
This has already been done (Table Manager, Ian Trackman, Blue Chip
Bridge), if I have this right. There is an existing protocol for this.
If I have this wrong, then .... never mind.
derek
> On Aug 14, 4:35 pm, Peteris <peteris.er...@gmail.com> wrote:
>
>
>
>
>
>
>
>
>
> > Hello,
>
> > I would like to announce the creation of jacobianhttp://peteriserins.com/2011/08/14/introducing-jacobian.html. From the
> > blog post:
>
> > "Jacobian is a simple, incomplete, but working protocol for computer
> > bridge programs. With the host program, you specify a deal,
> > vulnerability, contract and the four players by specifying their
> > executable programs. The host then leads them trough a game of bridge
> > (currently, only the play of the hand is implemented). That’s it."
>
> > The actual software components are available here:
> > * host program:https://github.com/Pet3ris/jacobian-host
> > * GiB wrapper:https://github.com/Pet3ris/jacobian-gib-wrapper
> > * GUI:https://github.com/Sedols/jacobian-GUI
>
> > Comments and contributions much appreciated. What do you guys think
> > about this?
>
> programs (or people) play one seat position per computer/human, and
> Jacobian manages the four playing units.
>
> This has already been done (Table Manager, Ian Trackman, Blue Chip
> Bridge), if I have this right. There is an existing protocol for this.
>
> If I have this wrong, then .... never mind.
That's my understanding too, and the protocol certainly exists. iirc,
it's how they play the World Computer Bridge championship.
Peteris
My reply got lost somewhere. I'll try to reiterate.
That's correct, there exists another protocol. It had the following
shortcomings for our purposes:
1) not really open, single author
2) no publicly available implementation, to the best of my knowledge
3) operates over TCP/IP. This is an implementation detail, but it
scares off people who might otherwise have contributed robots
4) no free clients for the protocol are available, again, to the best
of my knowledge
5) no individual bidding and play modes, which is absolutely mandatory
for analysis of hands
derek
Well, you carefully don't mention what it is. If we assume
http://www.bluechipbridge.co.uk/protocol.htm, which is what the World
Computer Bridge championships use, then:
>
> 1) not really open, single author
Possibly - but plenty of people are writing programs that work with
it.
> 2) no publicly available implementation, to the best of my knowledge
> 3) operates over TCP/IP. This is an implementation detail, but it
> scares off people who might otherwise have contributed robots
Huh? If it _doesn't_ work over TCP/IP it'll scare off a whole lot
more. What use is a network protocol that doesn't work with TCP/IP?
> 4) no free clients for the protocol are available, again, to the best
> of my knowledge
WBridge5 (http://www.wbridge5.com/) and Bridge Monitor (http://
www.wbridge5.com/bm.htm) are free. Not Open Source, but free.
> 5) no individual bidding and play modes, which is absolutely mandatory
> for analysis of hands
But you don't even _have_ bidding, which those implementations do.
I'm all for open standards, but what's the use of publishing a
standard for something that all the decent Bridge programs have
already implemented differently?
Ian T
comment :
> It had the following shortcomings for our purposes:
> 1) not really open, single author
The source code is not open because Table Manager is part of the full
Blue Chip Bridge commercial package, which is available for purchase
from our website and elsewhere. Nevertheless, we are prepared to
provide assistance to any bridge program developers who want to
interface their program with Table Manager.
> 2) no publicly available implementation, to the best of my knowledge
Please see 1) above
> 3) operates over TCP/IP. This is an implementation detail, but it scares off people who might otherwise have contributed robots
Please see comment from Derek on 16 August. TCP/IP is used because
it's a very simple protocol for this purpose, requiring only the
sending and receiving of text messages.
> 4) no free clients for the protocol are available, again, to the best of my knowledge
It depends what you mean by "clients". There's a difference between a
simple interface application to Table Manager and a bridge-playing
robot driving that interface. As mentioned above, we can help
developers with the interface.
> 5) no individual bidding and play modes, which is absolutely mandatory for analysis of hands
That was not considered to be a necessary part of the design (as
confirmed by all of the participants in the World Computer Bridge
Championships).
As indicated on our website, the protocol was a draft for approval by
the WCBC participants, but since its implementation in 2005, no
further enhancements have been suggested to us. If others think that
there are shortcomings, we would be happy to consider suggestions for
improvements.
Ian Trackman