Gib source code
Tim Posney
and releasing it under GPL, He kindly responded and suggested that I
post the response to rgb, so here it is.
I''m not sure how many people would be interested in playing with the
source code. Perhaps people could post expressions of interest to the
list.
regards tim posney
"I've been contacted recently by someone who wanted me to relicense the
GIB source code under GPL. For a variety of reasons, I don't feel
that I can do that. Nevertheless, I've gotten so swamped with other
responsibilities that the code isn't getting the attention it
deserves, and I'm happy for other people to play with it. So if you:
(1) Are not associated with any competing bridge product,
(2) Will sign a fairly standard nondisclosure, and
(3) Agree that I get the (nonexclusive) right to use whatever
improvements you make,
I'll send you a copy of the GIB source to play with. I'll send you
the source for the bridge engine only, not the user interface. (I
don't want to start giving the whole product away for free!)
Matt Ginsberg"
Alex Martelli
> I recently contacted Matt Ginsberg regarding the source code for GIB
> and releasing it under GPL, He kindly responded and suggested that I
> post the response to rgb, so here it is.
>
> I''m not sure how many people would be interested in playing with the
> source code. Perhaps people could post expressions of interest to the
> list.
I'd definitely be interested, particularly in studying the differences
between GIB's approach and that used in Bo Haglund's library (which *IS*
GPL'd now). Finding the _time_ to do so is another matter, of
course;-). But I do fully understand that Matt, having commercial
agreements with BridgeBase, doesn't want to (or cannot) just opensource
his work. So, the conditions he's requesting for selectively sharing his
sources appear to be very reasonable ones (I assume that the NDA will in
fact also contain restrictions against any redistribution, including of
binaries; and that Bo Haglund's free-software DD library does not count
as a "competing bridge product", since it's not commercial).
Thanks for exploring this possibility and for posting Matt's response!
Alex
Travis Crump
> Tim Posney <tpo...@gmail.com> wrote:
>
>> I recently contacted Matt Ginsberg regarding the source code for GIB
>> and releasing it under GPL, He kindly responded and suggested that I
>> post the response to rgb, so here it is.
>>
>> I''m not sure how many people would be interested in playing with the
>> source code. Perhaps people could post expressions of interest to the
>> list.
>
> I'd definitely be interested, particularly in studying the differences
> between GIB's approach and that used in Bo Haglund's library (which *IS*
> GPL'd now). Finding the _time_ to do so is another matter, of
> course;-). But I do fully understand that Matt, having commercial
> agreements with BridgeBase, doesn't want to (or cannot) just opensource
> his work. So, the conditions he's requesting for selectively sharing his
> sources appear to be very reasonable ones (I assume that the NDA will in
> fact also contain restrictions against any redistribution, including of
> binaries; and that Bo Haglund's free-software DD library does not count
> as a "competing bridge product", since it's not commercial).
>
Don't be silly, of course it is a competing bridge product. That's like
saying Firefox doesn't compete with IE.
Travis
Greg
design notes I was wondering if it might be possible to extend the
carding methods.
The basic idea would be that each defender takes the same set of a
million deals (i.e. they agree the same random seed in advance) that
correspond to the auction and dummy, and uses this as a reference to
decide which information is most useful to exchange, probably via
several discards. i.e. "That 9 of clubs means he's got lousy diamond
spots under the AQ he just showed, but no heart honours, and about 75%
sure to have exactly 3 or 4 spades, next discard to confirm." Well,
perhaps a bit less tricky than that.
You can obviously extend this idea to the symmetric relay sequences.
No idea if you'd get much percentage out of it - and if it got
complicated it might not be much fun to play against.
Can I ask, did Matt give any indication as to whether it would be
particularly easy to hook up a licensed GIB front end to a tweaked
engine ?
mghmaine
"Greg" <solomo...@hotmail.com> wrote in message
news:1149185802.7...@c74g2000cwc.googlegroups.com...
I must say this is sounding more and more like a computer discussion
group than a bridge one.
M. Hopkins
>
Todd Zimnoch
> I must say this is sounding more and more like a computer discussion
> group than a bridge one.
And sometimes it sounds like an etiquette forum, but etiquette is
directly relevant to the game of bridge too. It's not like we're going
into the details of pruning heuristics for our min-max searches, or the
ideal hash function for caching solved positions. Though if we were,
the discussion would probably have greater value to the bridge
community.
-Todd
anon
>mghmaine kirjoitti:
>> I must say this is sounding more and more like a computer discussion
>> group than a bridge one.
>
>And sometimes it sounds like an etiquette forum, but etiquette is
>directly relevant to the game of bridge too.
> It's not like we're going
>into the details of pruning heuristics for our min-max searches, or the
>ideal hash function for caching solved positions.
Please!!!! Beats the heck out of beating the heck out of recalcitrant
explainers!
> Though if we were,
>the discussion would probably have greater value to the bridge
>community.
Jim
usawa...@yahoo.com
Alex Martelli
At the level of published algorithms, none -- Bo's draft paper quotes
Matt's (and its precursors), and the program has zero-window search,
Ginsberg's partition search, the usual transposition table etc etc.
Without actually studying Ginsberg's sources, I cannot identify (nor
therefore study) the differences -- which is why it would be interesting
to do such study (having said which -- I've barely scratched the surface
in studying _Bo's_ sources either, so far... barely more than was
necessary to make them portable to Linux and the Mac... the problem is
definitely not a lack of interest, but, rather, one of time!-).
Alex
Tim Posney
interested parties go to www.otsys.com/gib which has a copy of the NDA
you will need to fill in and the actual source code.
Mmbridge
news:1149298817....@u72g2000cwu.googlegroups.com...
> Hi, After seeing that there were several responses, Matt has asked that
> interested parties go to www.otsys.com/gib which has a copy of the NDA
> you will need to fill in and the actual source code.
I have filled out the non-disclosure form and faxed it to On-Time Systems,
but I must confess I am not sure what the next step is since the form
neither asked for my telephone number nor my E-mail address.
Mmbridge
Tim Posney
Sounds like you have detected a flaw in the registration process.
I would send an email with your return addresss and Matt will send a
username and password.
tim