[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: current development

From: Sarah Payne
Subject: RE: current development
Date: Wed, 29 Jan 2020 09:28:31 +0000

Hi Joseph. Before responding to your last let me repeat that I don’t have anything like the skill or experience necessary to contribute real expertise here, only enthusiasm for the game & a great deal of admiration for the gnubg project, so please make allowances for that. But re your query about analysis of gnu with xg: could this be the next step for planning an update / new release?


For sure, I would expect gnu still to have some playing strengths against xg. That’s what is so impressive: how it’s still as strong as it is with no new recent versions. Let’s not forget we’re only comparing it to what’s currently considered to be the world’s best.


Intuitively? I think gnu still has something xg doesn’t, ‘a different animal’ is not a criticism per se. But as you point out, intuition is nothing. Also it doesn’t seem necessary to wait another year for the new version of xg. This kind of comparative / competitive development will always be ongoing / open-ended.




Sent from Mail for Windows 10


From: Joseph Heled <address@hidden>
Sent: Monday, January 27, 2020 9:23:36 PM
To: Mary Hickey <address@hidden>
Cc: Sarah Payne <address@hidden>; pviau <address@hidden>; Timothy Y. Chow <address@hidden>; Michael Petch <address@hidden>; Øystein Schønning-Johansen <address@hidden>; Chris Yep <address@hidden>
Subject: Re: current development
Sounds like a good idea, but why bother with BOTS? just play against the current GNU version and analyze with XG. Perhaps this has already been done and just needs publicizing?


On Tue, 28 Jan 2020 at 10:04, Mary Hickey <address@hidden> wrote:
Hi Joseph,

I was referring more to the interface than the playing strength, but you are right that the general perception is that XG is the gold standard regarding playing as well as onscreen viewing and printing. Maybe that perception needs to be addressed, though since a new version of XG is rumored to be coming out soon, it might be better to wait and compare GNU to that instead. 

One way to produce some data, which isn't exactly the same as evidence, is have a bunch of us play vs. whatever bot is considered to be the best representation of GNU at the various servers, then chuck the matches through XG at a comparable level, say 3-ply rather than world class. Has this already been done? And if this idea makes any sense, which bot at FIBS and also at Backgammon Studio best represents the current GNU? And what level of XG analysis is to be considered comparable?

After we find where the bots differ, we can roll the positions out and see which gets the nod. I'd trust either bot to roll them out well on good settings.

I like this method because it's not difficult either to understand or to do. It won't detect all the flaws in either program, because rarer positions aren't as likely to be seen even in a large number of matches (...duh...that's why they're rare LOL) but we might learn something anyway.


From: Joseph Heled <address@hidden>
Sent: Monday, January 27, 2020 3:24 PM
To: Mary Hickey <address@hidden>
Cc: Sarah Payne <address@hidden>; pviau <address@hidden>; Timothy Y. Chow <address@hidden>; Michael Petch <address@hidden>; Øystein Schønning-Johansen <address@hidden>
Subject: Re: current development
I am very happy to see (reasonably priced) for-pay services for BG players. It means the game is still alive, which is far more important than any one BOT, free or not.

 But perceptions vs. reality is one of the issues we are talking about, right? would you be able to put something, based on your experience, that will show that XG and GNU-BG are not that different in terms of playing strength, or "prove us wrong" by showing why XG is superior?


On Tue, 28 Jan 2020 at 08:19, Mary Hickey <address@hidden> wrote:
Hi Joseph,

It doesn't have to be the end, but from the posts I'm reading, it appears many programming person-hours will need to be directed just toward catching up with XG and other tools now available elsewhere. For example, Pierre talks about quizzes, but you can take quizzes all day and night at backgammonstudio.com, play vs. other people, yak at the forum and also study an extensive library of matches for $24 a year. 

But since GNU is the bot that follows you in consultation matches there, and also is the engine running the FIBS-bots if memory serves, it has some immortality, don't you think?


From: Joseph Heled <address@hidden>
Sent: Monday, January 27, 2020 2:03 PM
To: Mary Hickey <address@hidden>
Cc: Sarah Payne <address@hidden>; pviau <address@hidden>; Timothy Y. Chow <address@hidden>; Michael Petch <address@hidden>; address@hidden <address@hidden>; Øystein Schønning-Johansen <address@hidden>
Subject: Re: current development
Nice obituary, Mary :)


On Mon, 27 Jan 2020 at 12:38, Mary Hickey <address@hidden> wrote:
To the entire GNU-bg Community,

I remain a GNU fan even though I'm One of Those People who stopped using it once XG established itself as the go-to backgammon bot. And having read this email exchange, I'm shouting out a huge "THANK YOU!" to this community for its contribution to the game, and its example of co-operation, trust and mutual respect among programmers from all over the world. 

I'm not a coder, but remember helping test the GNU-bots at FIBS and providing feedback regarding their practical play on FIBS to the developers. I appreciate the GNU community's generosity in permitting GNU-engined bots to play on any servers that want them.

Wherever this project goes from here, the co-operation and always respectful communications among the members of this community shine brightly in a world that needs more of those qualities. 

Sincerely yours,

Mary Hickey

From: Bug-gnubg <bug-gnubg-bounces+thehick64=address@hidden> on behalf of Sarah Payne <address@hidden>
Sent: Saturday, January 25, 2020 10:18 AM
To: pviau <address@hidden>
Cc: address@hidden <address@hidden>; Øystein Schønning-Johansen <address@hidden>
Subject: RE: current development

Hi Pviau, re position databases I’m guessing this is just the kind of feedback needed, particularly coming from someone familiar with xg’s potential. It would be a huge bonus to have a feature like this to launch a new version & get gnubg back on the radar of serious players.


But I’d suggest just one new feature / usp of this kind (& err towards keeping it simple). Develop neural nets / cross platform compatibility then get a new version out there soon as, before people forget how good the project is.


Completely agree re offline functionality. Also re UI: flat, simple, clean. Very little required there.



Sent from Mail for Windows 10


From: pviau <address@hidden>
Sent: Saturday, January 25, 2020 11:04:43 AM
To: Sarah Payne <address@hidden>
Cc: Øystein Schønning-Johansen <address@hidden>; address@hidden <address@hidden>
Subject: Re: current development

I’m an ex-avid-gnubg player, but ever since I found XG, I never looked back. And this is from a die-hard Mac user/evangelist, who actually only installed Parallels Desktop so that I could run XG on my Macs. Please don’t tell my friends.

Neural nets are of course key to take gnubg seriously again, and lots of competent people will end up taking care of this, I sincerely hope. But I would like to suggest another angle to evolve gnubg.

But first a comment on UI/looks. I also do not share the view that gnubg is much inferior to XG there. Heck, I consider XG’s UI is its only weak point. It feels like a 1990’s Windows app, because essentially its creator (who is a fantastic guy) never cared much to evolve his UI skills beyong what he learned :-)

So if gnubg needs to improve its UI (which it does), from a visual point of view the starting point should be something like backgammongalaxy (the web site). Simple flat color schemes, as little visual noise as possible, everything geared towards efficiency for learning yet clean and elegant. And no, no 3D will ever help anybody learn to play better backgammon.

Now for the other angle I mentioned above.

What I am missing in the various competent apps around, is a trivial way to build position databases, and study them. So this is a function which would make an app stand out, at least for a while:

- the database itself, with sections and filters etc
- easy to feed from various 1-or-2-click methods (flag a position while playing, while replaying someone’s match, copy XGid or any other id from another app and paste into the db, why not even from a screenshot of the whole board...)
- the possibility to batch-generate technical equity data on a position, a section, the whole db, using rollouts
- the possibility to define quizz-based studies and score them (on a section, on any hand-picked list)
- a history of quizz scores to show play quality evolution over time

All of this can be done today, but it takes a combination of tools and a lot fiddling. This has to stop :-)

Oh and finally: Android *AND* iOS have to be on the radar. Absolutely. Soon only dinosaurs will be carrying around a laptop everywhere they go. I was talking about myself there, but still even I may evolve one day...


On 25 Jan 2020, at 11:15, Sarah Payne <address@hidden> wrote:

Here’s where I’ve got to - hope some of it proves helpful.
I contacted the USBGF and UKBGF asking for any feedback / interest and / or recommendations re generating fresh input from a new generation of coders (also posted similar messages out to a few C programming forums).
Not much joy here I’m afraid. I think the main issue at USBGF & UKBGF (ie for professional or competitive players / club players or serious hobbyists) is how far gnu has fallen behind XG now (unlike Snowie in its day, XG is generally affordable, available for mobiles & a new version with enhanced neural nets, compatible with Mac as well as Windows for the first time, is due for release at the end of the year).
But ‘free to all’ is still a significant USP for GNUBG especially among younger players (and in other parts of the world) & GNUBG is still cited up there as best of the rest on more general games forums. But as far as I can tell (in the UK at least) there are no clubs / tournaments / forums ringfenced for younger players so hard to know how to tap into this for feedback / enthusiasm / new coding talent.
I also contacted Chris Bray, one of the UK’s leading writers and promoters of the game. He’s the one who filled me in on the latest re XG. In his opinion:
‘…gnubg always had creditable neural net engines but never had a friendly User Interface which held it back considerably. I always felt it was written by technical programmers with little commercial awareness of how people would use it in the real world.’
I don’t share this view. I’m not a techie but I prefer the gnu layout / interface to XG’s. Having said that, I’m not a professional / competitive player either, which no doubt involves different priorities. Also I’ve been playing GNU for a long time & it probably took a few goes to get it set up the way I wanted (not sure that’s how big a deal that is, though).
In Bray’s opinion, ‘for gnubg to reemerge as a viable alternative to XG it will need an enhanced user interface and well-integrated app version for apple & android tablets / phones, as well as upgraded neural nets.’
In my opinion, a phone version would broaden appeal / access but the most critical issue is the neural nets. XG feels like a very different animal as an opponent - noticeably more opportunistic & aggressive, so some degree of congruence asap seems critical for gnubg to hold ground.
A final note from Chris: 
‘FYI on my ToDo list is to talk to DeepMind about whether they intend to create an AlphaZero Backgammon.’
No update from him on this as yet. How about contacting them yourself to propose a gnubg / DeepMind collaboration?
Alternatively (or also):
In a subsequent email Chris mentioned ‘seeing an emerging group of highly-talented younger players, many of them Japanese.’
How about a Japanese collaboration / appeal for new coders? Language is clearly a barrier to this, but would it be possible to find some bilingual volunteers via the main gnu project to act as go-between? I don’t know how global the main project is, but backgammon is huge in parts of the Middle East so appeals here / Africa / India (where English is also more widely spoken) could also prove fruitful.
Finally, whilst I haven’t had any specific responses from UKBGF or USBGF to my general inquiries, it should be possible to put out a broadcast with either or both of these organisations (& via them to the local club networks) with specific announcements or requests for feedback eg for or from younger players / coders interested in collaboration. You may want to consider this. If it’s general player feedback you want, you probably need to frame some very specific questions. Oystein said: ‘First we need an idea, then we have to verify that idea, and then we have to set it into life, which might trigger a bigger VM or a cluster.’ As a non-techie I get the gist of this (and as a description of the work process find it intriguing!) but I’m in the dark as to what type of ideas you mean, or what kind of information you want.
There we are, then: my progress to date. Still happy to help, for what it’s worth, and I would love to be kept updated on any progress your end.
All best with that, & a Happy New Year to you all – thanks once again for all the fine work to date.
Sent from Mail for Windows 10

From: Øystein Schønning-Johansen <address@hidden>
Sent: Wednesday, December 11, 2019 5:48:59 PM
To: Sarah Payne <address@hidden>
Subject: Re: current development
Yes. Just give it some thought. If you post to the mailing list or to just me, you can decide yourself. I'm not the most active developer at the time, so maybe posting to the mailing list is a good idea.


On Wed, Dec 11, 2019 at 6:10 PM Sarah Payne <address@hidden> wrote:

Ok I’ll give this some thought. Do I reply to you or reply all?


Sent from Mail for Windows 10


From: Øystein Schønning-Johansen <address@hidden>
Sent: Monday, December 9, 2019 10:08:34 PM
To: Sarah Payne <address@hidden>
Subject: Re: current development
Hi, Sarah!

Thanks for your gratitude. I think all involved developers are busy with other day time jobs, and GNU Backgammon is hence just a spare time project for us. We hence have what we need for a living through our daily jobs. However, as you see development has slowed down the later years.

I think what we need in this project is:
- Motivation
- Enthusiasm 
- Cheering

so, I think that you email started some sparks. You saw that! Maybe if we just continue to post some messages to this mailing list, maybe something will even happen. Keep posting suggestions and question and be positive and cheer up the life of the readers. That will probably be the best contribution back to the project.

There might be occasion where some hard money can contribute and that might be when/if we start training something on big virtual servers, and these virtual servers can have some cost attached. But that is only guesswork from me. First we need an idea, then we have to verify that idea, and then we have to set it into life, which might trigger a bigger VM or a cluster. In that case we can discuss how to fund that. Sponsors or we chip in or we get voluntary gifts from backgammon enthusiast? Anyway -- It's far ahead.

Maybe fresh blood among the developers might help? Do you go to a local backgammon club?
Are there any computer geeks and nerds in you club? (apart from you?). The developers in this projects are old nerds with gray hair or no hair at all (like me), who learned to develop software last millennium, and maybe some of the code can be improved if fresh blood was added. If you are playing in a club, your contribution could be to go over to the young (younger than me and the other guys in the development team) geek in the corner and ask him/her if he she has seen GNU Backgammon, or knows neural network, or like programming, and knows the C programming language... so on....  if you get a developer interested that can also be your contribution back. :-)

That's how it goes. And thank you so much for the spark you started. It means a lot.

Best rolls and regards

On Sun, Dec 8, 2019 at 6:36 PM Sarah Payne <address@hidden> wrote:

Hi Oystein


Unfortunately I’m not a coder but I’ve used this software so often for no charge I was wondering if it’s possible to give money sometimes to the backgammon project to help keep it going / up to date. I’m sure I can’t be the only person to feel this way. You should all be very proud of this programme which has remained competitive for so long even without much recent development. In my opinion, it’s still the most user friendly set-up as well - v flexible & intuitive.


What are the biggest hurdles to keeping it competitive? Man hours, computer hours? Does the neural networking approach used up until this point need to evolve / become more resource hungry to keep up with something like Extreme Gammon for example?


Forgive my ignorance in this area – I’m very interested but understand very little of this area. And if this is not appropriate conversation for these lists, no problem, just let me know.


Thanks – and please, yes, feel encouraged to get going again! :)





Sent from Mail for Windows 10


From: Øystein Schønning-Johansen <address@hidden>
Sent: Tuesday, December 3, 2019 5:01:50 PM
To: Sarah Payne <address@hidden>
Cc: address@hidden <address@hidden>
Subject: Re: current development
Hi Sarah!

Thanks for taking contact. Good to hear that you like GNU Backgammon.
Is it still under development? Hmmm... debatable. There has not been many major improvements the last few years.

Take a look at the projects ChangeLog.

As you see there isn't much happening.

Of course you can contribute if you want. After all this project is Open Source an anyone can do whatever changes they want.
Just post comments here on the mailing list, and it can shear up some of the sleeping developers.

If you are a developer and want to contribute with code, we can of course provide you write access to the cvs repository. (Yes, it is as old that it's using cvs to do code revision).
Since everyone is more or less "sleeping", there is no real TODO list. Maybe some code janitor work? Refactoring? Maybe c99-ify some of the code. Maybe you can suggest a feature? Or report a bug?

Even though I'm not doing much on GNU Backgammon (I've not done much the last 10 years) these days, I guess if we just chat about some details, it might be the spark that starts up a new motivation among us. There are some discussions still on this mailing list, last week there was a new Match Equity Table presented (Thanks Ian). If we just chat more, maybe something can start flowing again. I'm getting more time as my kids grow older. So, who knows what happens.

Best regards,

On Tue, Dec 3, 2019 at 3:58 PM Sarah Payne <address@hidden> wrote:

Hello there. Been a huge fan for many years of gnu backgammon, many thanks to everyone involved. Is the software still under development with new versions coming? Is it possible to contribute directly to this project?






Sent from Mail for Windows 10

reply via email to

[Prev in Thread] Current Thread [Next in Thread]