A new server is being developed: Kaya.gs

[Kaya.gs]

I find this interesting, I might even put money behind it but:

• Show me some working code, please.
  I'm sorry, but I work with Change Management for a living and I see so many "almost working" projects getting stuck on something very fundamental (e.g. handling synchronous(!)/asynchronous communication). You having some fundamental working code (I don't need to see it) to show us will inspire confidence in you. For me, this is a must before I give you any money what so ever.
• What are your non-functional requirements?
  You have an impressive list of functional requirements. What about the non-functional requirements? How do you plan to meet them?
  For a project such as this I expect that you provide an architecture for the service.

I don't want to be negative, and I do hope you can do this. But you do ask for money - our money. How do we know you put them to good use? Do you even have the necessary skills? You need to give us more than "because I say so"

Hey pel. Great suggestion about demoing, i will talk to Patricio about demoing the board and the chats. However within a month we will most likely be able to play already so we might put off demo-ing the server.

The server's core language is Ruby, Tt will have a node.js server to handle Websocket/Comet/XHRpolling..etc communication with clients. The main database to be used is Redis, which is extremely fast. The server will be hosted in the Amazon cloud. Our libraries and architecture is thought off with scalability first in line.
We are considering (but no promises yet) to make some parts of the server open source. Most likely client code like the board and JS widgets.

If you want to talk more technical details out of curiosity you can pm me, but i dont think its interesting for most people.

Concerning skills, i can tell you that both my partner and i have experience with web-development, and we have a technical consultant that is a world recognized Ruby-Guru. In the last year i developed sites like http://www.poketypoke.com, http://www.phonetag.com, http://www.unsubscribe.com.

Let me add one thing. Kaya.gs is not in "idea phase", its in development. We have already played games in our private development environment. We have chatted away, created rooms, and other fancy things.

[Kirby]

Not to threadjack but since I was mentioned by name....

...Oh well, I think actually Bantari has long since nailed it on the head in that it's an emotional issue and all people really want is catharsis, which luckily can be provided by ranting on forums (=

Not to threadjack further , but my view is that your comment here is an attempt to portray "the ranters" as not having valid points, and just releasing emotion. I do not think that this is fair.

My view is that a valid alternative escaper policy has been pointed out numerous times, and has yet to be solidly refuted. Typically, by the time the discussion moves to a viable alternative policy, the discussion is shutoff as getting too emotional, or redundant, perhaps.

I do not think it is "catharsis" to simply observe that a system that works as follows:
* Have a timeout after 10 minutes, which causes one to lose the game if they don't return.
* Have an option to mutually agree upon resuming.

would not only totally eliminate escapers, but would also allow for people to postpone games if they want to.

I don't think it's an emotional reaction at all to observe that this system would be superior to KGS's.

Perhaps the reason that this is discussed time and time again is simply because a rational and better alternative exists, but some people do not want to admit this.

[Kaya.gs]

This server seems like an interesting project (as many have said very ambitious). I must admit I see a lot of challenges, and it seems like a lot of the ideas will require a large infrastructure in place (ex: sending problem sets out to people with beginning accounts, managing the bandwidth of every user potentially having streaming video). I definitely think that if they are going to try and do this full time they will need to find some sort of sponsor, and will certainly need support from East Asian countries (you say you have a business plan so hopefully you have this aspect covered or at least a solution in the works). I think one thing that seems promising about this project is that it is just two people who believe they have recognized specific things they feel are lacking in other available go outlets. It seems like the projects that actually end up working are the ones that have one or two dedicated people with a simple goal, as opposed to a large group with vague lofty dreams. One thing I might suggest is taking a look at the currently "launching" EuroGoServer and see if you can glean any lessons learned (what are they doing right, what are they doing wrong?).

Hey Mef. We have looked at EuroGoServer. It is a very similar attempt but its unfortunately not very well done. As i mentioned before in this thread, we will resemble desktop clients in overall look&feel and usability, whereas that server is organized like a turn based website with a clocked board.
We liked the concept of a 3D board though, we are definitely going to look into that (but a prettier one ).

The infrastructure is the thing i fear the least at this point of the project. I spent my time back in April talking with Software Architects and gurus to avoid pitfalls and the result is a very easy to understand and scalable result.

As you said, this is an ambitious project and to achieve it we are going to be full time. Hence our call for community funding to help us out in the beginning. We have the road set for the next steps regarding support. However any aid right now will speed things up by letting us develop quicker and faster and worrying about resources a little later.


I hope to see you in the BoF , Mef .

[iazzi]

You can see thunderbird from mozilla as an example of a java client that lost badly to online email clients.

Since Mozilla Thunderbird is not a java client (nor are most IGS clients, nor WBaduk or Tygem as far as I can tell) I wonder what you are calling "java"...

Sorry i meant desktop client for thunderbird. Thunder is made on C++ which is certainly something we wont use for Kaya.gs.

Im surprised about wbaduk if its not made with Java. The homepage does ask for Java web-start to run the client.
Regards.

I am not sure either. I installed it via the installer and saw no signs of Java. Anyway it is unimportant now that I understand your point

I still suggest to design server/client as cleanly separated as possible even if you only produce the web-client for now... If you succeed and are able to afford more programmers, or will decide to offload the client development by making it open-source you will not regret it.

Good luck!

[Bantari]

I find this interesting, I might even put money behind it but:

• Show me some working code, please.
  I'm sorry, but I work with Change Management for a living and I see so many "almost working" projects getting stuck on something very fundamental (e.g. handling synchronous(!)/asynchronous communication). You having some fundamental working code (I don't need to see it) to show us will inspire confidence in you. For me, this is a must before I give you any money what so ever.
• What are your non-functional requirements?
  You have an impressive list of functional requirements. What about the non-functional requirements? How do you plan to meet them?
  For a project such as this I expect that you provide an architecture for the service.

I don't want to be negative, and I do hope you can do this. But you do ask for money - our money. How do we know you put them to good use? Do you even have the necessary skills? You need to give us more than "because I say so"

Hey pel. Great suggestion about demoing, i will talk to Patricio about demoing the board and the chats. However within a month we will most likely be able to play already so we might put off demo-ing the server.

The server's core language is Ruby, Tt will have a node.js server to handle Websocket/Comet/XHRpolling..etc communication with clients. The main database to be used is Redis, which is extremely fast. The server will be hosted in the Amazon cloud. Our libraries and architecture is thought off with scalability first in line.
We are considering (but no promises yet) to make some parts of the server open source. Most likely client code like the board and JS widgets.

If you want to talk more technical details out of curiosity you can pm me, but i dont think its interesting for most people.

Concerning skills, i can tell you that both my partner and i have experience with web-development, and we have a technical consultant that is a world recognized Ruby-Guru. In the last year i developed sites like http://www.poketypoke.com, http://www.phonetag.com, http://www.unsubscribe.com.

Let me add one thing. Kaya.gs is not in "idea phase", its in development. We have already played games in our private development environment. We have chatted away, created rooms, and other fancy things.

Hmm... The mention of Ruby sparked my interest again.
From my personal experience, I question the wisdom of using Ruby as the basis of any large long-term-support project. I know, I have three of those running, and I wish it was something else but Ruby... even PHP.

Still, Ruby is a blast to program... fun to write, nightmare to maintain.
Why did you chose Ruby?

[Kaya.gs]

Hmm... The mention of Ruby sparked my interest again.
From my personal experience, I question the wisdom of using Ruby as the basis of any large long-term-support project. I know, I have three of those running, and I wish it was something else but Ruby... even PHP.

Still, Ruby is a blast to program... fun to write, nightmare to maintain.
Why did you chose Ruby?

I have experience creating and supporting projects that have Ruby, and i must say i haven't had your rough experience.
The only alternative i considered was python, but i rejected it after picking the framework we are using(that is not Rails nor Sinatra).

I wonder now if i should add a section "tech geeks" on kaya.gs .

I still suggest to design server/client as cleanly separated as possible even if you only produce the web-client for now... If you succeed and are able to afford more programmers, or will decide to offload the client development by making it open-source you will not regret it.

They are very independant and we have that goal in sight.

Thanks for the cheering, now support us!

[Bantari]

I have experience creating and supporting projects that have Ruby, and i must say i haven't had your rough experience.
The only alternative i considered was python, but i rejected it after picking the framework we are using(that is not Rails nor Sinatra).

Really? I only know Rails and Sinatra. What framework are you using? This is interesting.
Also - let me know if you need help. I have a full-time ruby/rails job but can do some programming after-hours if it helps. Free of charge.

[Kaya.gs]

Check your pm Bantari, lets take the convo to private .

[danielm]

Hey, I want to know too! Please add that geek section.

For what it's worth, I think you are spot on with your analysis of existing Go servers and the potential that is still in there. So much as to that I had essentially the same plan for the mid-future, but I will be happy to see you succeed instead. If you fail, I may pick up that plan again and make sure to learn from your mistakes.

The current state of Go software and services is really depressing, and I hope that you will succeed not just in creating a popular Go server, but also in raising the standard of what we expect from Go related applications in terms of design and usability.

It will be hard though, and any single misstep might spell doom for the project. So I am only cautiously hopeful. Still debating with myself what kind of donation I can justify (of course seeing a working prototype in action would help a lot with that ).

[uPWarrior]

As a fellow programmer I really hope you know the pile of work ahead of you. In projects like this, it's always difficult to find the best "release point", too many features and you'll work forever, too few and a a bad start will probably dictate its failure. Good luck in that.

In case you succeed, please open the server interface to the public. What the go community really needs is a sturdy gaming/chatting platform that others may use to build on top on. If we were able to put this "market share" issue aside, the boost in the overall go software quality would be huge.

[Mr. Mormon]

The servers that come closest to my ideal are KGS and DGS, but currently I hate all of them. You know what makes people (well, me ) happy? Options. I'm not versed in rating or timing systems, but "Go" is a broad term that encompasses many ruleset variations and real-time and turn-based socially expected playing speeds. Give me the option of customizing my ruleset to a modification of New Zealand, the ability to find players of the same strength of me without having to wait online, and turn-based functionality, and I'll join if you ever get there.

A word of warning. The heads of ambitious projects have a habit of getting more opinionated and stubborn when they finally get successful. I don't want to think your vision will make you listen to potential players like me less in the future. Will people be talking about your server like we're talking about KGS now?

[hyperpape]

I don't think it's an emotional reaction at all to observe that this system would be superior to KGS's.

Perhaps the reason that this is discussed time and time again is simply because a rational and better alternative exists, but some people do not want to admit this.

Vive la threadjack!

It just now occurs to me that you need not only show that this system is superior to KGS, but that it is worth having an essentially homogenous set of time systems for all the major servers. KGS is the only server that does not require you to reconnect within a set time period, and your proposed system eliminates that. Right now, people who want a strict time system can play on Tygem or IGS, people who want a forgiving one can play on KGS. To say your system should be adopted on KGS means that there should be no forgiving servers. Even if a forgiving time system is inferior, that doesn't mean that three strict servers is superior to two strict and one forgiving.

[Redundant]

Could we get a separate geek thread. It's sad that all the technical talk is going to pms. I'm not competent enough to program anything like this, but it is entertaining to listen in. Maybe if one of my cs* friends spontaneously decides to do some ruby web programming, I can send them your way.

[*]My CS friends are scary good at programming. As in, top 10% at a top 3 school in the world.

[tapir]

There is a business plan, but im not going to go public with it now as it can change a lot over the course of the first few months. It will be free to play, that's certain.

Of course, my partner and i are owners of the server. If you want to invest to get ownership you can talk to me in private,we are currently dealing with investors.

This crowdfunding opportunity is limited. In less than 60 days it will be closed off and we wont hand out founder accounts , the documentary on how we are building it, the first games games of the server will have passed by then. This support helps us build a better demo to show to future investors and to get a good head-start on its development.

Thank you for providing at least some answers. I am still curious how it will generate the income necessary to maintain a server on the envisaged level. (Premium accounts, side bets, advertisement?)

[Kaya.gs]

Could we get a separate geek thread. It's sad that all the technical talk is going to pms. I'm not competent enough to program anything like this, but it is entertaining to listen in. Maybe if one of my cs* friends spontaneously decides to do some ruby web programming, I can send them your way.

I will talk to Polly(Patricio) today about adding a geek section on the site, and see if there are developers out there that want to get started on doing some things.

Also we will add a meter , we have raised 460 dollars so far, not bad



We already added a Feedback tab, so feel free to add a suggestion for the server or page or anything. You can also vote some of the things people have already posted on it.