Wednesday, November 23, 2011

Facebook Login: do we want it?

Hey guys. A recent comment on Facebook reminded me that i haven't made up my mind around the subject of including Facebook into Kaya.

There are a few things Facebook does that it gives some conveniences so any site that integrates part of its funcitonality.

The first and foremost is the log-in. Traditionally any site where you are required to make an account, has some email validation, password and so on. It is a current trend that many sites allow you to log-in with your Facebook(or Google, MSN or even Yahoo ID). Normally you can still choose a nickname or such.

It has a big advantage: people with facebook accounts can log in very quickly, without email-validation.

From that, there are a few more things that can be done. Potentially, many people would like to share things that happen on the server on Facebook, like a cool game, or a tournament result, etc. So while logged in with your facebook account, we could potentially give some buttons so people can do a quick post on their own wall .

I don't have much interest in the next step, which is kaya directly posting on people's walls. I think its tacky so thats out of the question :).

There are 2 reasons for not wanting to have facebook integrated.

1- Many people loathe facebook. Since FB is so easy to integrate, even really crappy sites have it. Personally i feel more friction when i read a FB login or buttons or whatever than making a new account with e-mail validation. But on the other hand, i have used FB login in places where supplying my name and picture was appropiate. For example, own uservoice page, where we handle community feedback.

2- Go Players love being anonymous. Many people dont play on KGS due to the strain of people watching and commenting their games. Many people have secret-hidden accounts. I believe that is not actually natural of go players, but it is caused by how KGS is done. I don't think Asians go through the same pressure , except maybe pros.
It would be better for Kaya if users were identified through FB. Each account would have more identity and we would have more information. This means people wont make new accounts all the time, and they would feel more connected to the server, as people know who they are.

You can vote/comment on it there!

Monday, November 21, 2011

Game Listings, GoLibrary and my return

Hey guys!

I havent updated in a while, i really owe some news :). Had a crazy time last week between tournaments, but i made a great new feature on . I was so psyched that i worked until 7am on friday giving it the finishing touches.

Its not styled so it looks a little messy but you can get a glimpse on how it will look.

All go servers i know of show the on-going/open games in a text-like list. Although its obviously practical (and supports it, you can see the "switch view button") we believe something better can be achieved. Hence we have a chart!

Games will be categorized by number of observers and the rating of the strongest player of the match, plus colors to distinguish different types of games (Replays, Sandboxes, Matches, etc), You can click on the dot and get a dialog with some information to watch the game, or start it if its an open game. In the future, the very tooltip or popup will show a preview of the game, so you know if you want to look at it before joining.

Otherwise i dont have much more to show for in the past week for myself: i've been working on the problem solving software. Its looking better and better. I am currently making it possible to create tests by hand: the idea behind tests is to check how much you know of a given subject, making it better to test for fuseki/joseki knowledge. Here at BIBA (Blackie's Baduk Academy) we have to study a whole fuseki each day, and i think they are missing tests on it.

With this feature they will probably create a daily test, which means we will get some nice non-tsumego problems soon. Then we will add Cho Chikuns Tsumego library, and make a widget to put on :). We really want to show the power of internalizing software outside the server, and that is our goal by making a first GoLibrary version right now.

However, part of me suffers for this delays. Had i stay in Argentina in October, we would have a playable server already. After so many events i attended i could get into the rythm of working, but because i have opposite times to Argentina, polly and I agreed on working things that dont require collaboration: GoLibrary, some GoSpeed(the board) details, and things here and there.

What we are missing specifically to get the first server version is score-agreement. Its kind of done but has to be polished and tested.

After that, we have to stylize some remaining parts of the server and we can launch, which we expect to happen before new year.

Monday, November 14, 2011

Correspondence Go (A.K.A "Turn Based") and GoLibrary

So the video working frenzy is over now,and we can focus on the next tasks at hand.

I did some clean-up and implemented some features people suggested at the Feedback section of the site. Now im implementing another suggesstion which is allowing "Turn Based " Games, like DGS/OGS and other servers like them.

First of all, i hate the phrase "Turn Based". Its always turn based. Or with a clock, u can actually play twice? . Truly silly name :)

In the Feedback section some suggestions were made and i think the one that is more appropriate is Correspondence Go. The word is a little uncomfortable but its the most accurate.

For those that dont know, Correspondence Go is basically a game with a very large clock setting (as in, main time 30 days, byo-yomi 1 day). This leads to a very relaxed kind of game, with no real time pressure. You can devote very little time in a day to play a move, which is as easy as checking your email.
Also , games with this settings are more studied: its more common to check at pattern variations, study the joseki, which is a good study and makes the players be able to reach the best decision they can make. Due to time pressure, we often fear or skip variations, or are forced to make rush decisions.

Usually Correspondence Go is much easier to do than real time go, because there are no strong time constrains on the software system. Thats why "Turn Based" servers are usually smaller in features , as they are a much easier problem to solve.

In our case, we have solved Real time games already, so allowing for Correspondence games is trivial: After all, they are a subset of real-time games. (just large clock settings).
They are a little different in other aspects. Because there are games where you play 1 move a day, usually you get more games going at the same time.

So we are going hybrid and allowing both types of games. The core solution is already done, but this brought to me the challenge on how to present the different type of games for the users. So i've been working on the UI.

Novelties? you can see which games are players playing and join them through their profile, just like its done on CGoban. Correspondence games are now supported and i will polish them a little before closing this feature. It will not be as rich as DGS/OGS, with vacation time and other minorities, certainly not now, but this might be a very good way to get initial games going on on the server.

Ill be back soon with updates!

Thursday, November 10, 2011

The supporter blog is open. We published the video! and muzzling...

Exciting news everyone! We finished the demo video we planned long ago. Its now in the home page of our site . Due to the fact that we are over 130 supporters, i decided to open up the blog for anyone to read.

I really wanted to keep it private but blogger restrictions made it impossible to do so, so right now everyone will be able to read it.

So everyone check out the video and share it with your friends :). It shows about 60% of what we have in the server (we are not showing the match creation and challenging, the rooms/channels, private messagings and profile sections).

Right now i am also finishing to implement the first idea brought to me by the community: muzzling. Normally admin's have to take care of obnoxious users, spammers and other type. Personally i was struck with this inconvienience on kgs, trying to find and admin, and by the time he comes in the damage is done.

Muzzling is about the users of the server "voting" to silence someone. They cant kick him out of the server, if a given number of users choose to muzzle another one, he will not be able to publish messages for some time.

I thought this was a great idea for the community to self regulate itself. For this exact purpose, it will work faster and more on-demand than calling an admin.

Monday, November 7, 2011

Glicko Rating system is done.

Today i finished merging yoyoma's implementation of Glicko into the server. It was a good learning experience for me to work on Open source code (OpenKaya).

I had to make several tweaks on his code which i should propagate to OpenKaya soon. But besides that, it seems to work very well, and we have now a rating system on Kaya, which includes ranks.

Glicko is an improvement on ELO taking into account mainly the confidence of the players rank, as if he doesnt play for a long time, his rating will move faster for the next games. It is not , however, like KGS's decayed history algorithm that takes past games into account.

This is much closer to my initial idea of how it should be done but after much discussion Whole history rating seems interesting to try. Definitely if someone makes it for OpenKaya we will have a showdown of rating systems.

For now its not our biggest concern because we need a good player base and regular playing for a rating system to have any meaning and building that has higher priority.
Things are still going good. It is unfortunate that we will not be able to make the beta this november. Its not impossible ,but because now i am in korea, some things that require simultaneous work are inefficient and uncomfortable.

However we are moving along well and working on manythings, so even though the minimum work required for the Beta is taking longer than expected, we will have other cool features sooner, like problem solving on the site.

Hope everyone is as excited as we are :)

Sunday, November 6, 2011

Overdue Posting.Blog will go open and Tygem knows.

Hey guys! I haven't posted in a while. I've been so packed with things to do here in Korea and talking to people and working on GoLibrary and other details.

Polly is wrapping up the video and we should have it anytime soon. It's taking a lot more than expected, due to Polly's appreciation for details. But it will come soon and this blog will go public. Blogger unfortunately only allows 100 private readers, and as you all know we have way more than 100 supporters. Most people that arent included havent accepted their invitations, which is why we didn't hit this problem before.

But now, newcomers are asking for the blog and i have no solution but to open up the blog. Other options that would allow me such a thing involve work and maintenance we are not willing to give it. So as soon as we get the video, this blog will go open.

On another news, i was contacted by both WBaduk and Tygem while in Samsung Cup. A guy from Wbaduk just recognized me in person, and someone in Tygem casually looked at the list of students at BIBA and asked Kim Seung-Jun if i was the Kaya guy. (they recognized me by name). They asked for a meeting also.

Last year, i went to the Tygem main offices in Seoul. I gave them some recommendations on how to open up to the westerners, and they were very good listeners i think. They gave me a set of books as a present for my advice. However, they did not succeed in accomplishing a western community nor any of the suggestions i made for them. This happened 3 months before i started to toy around with the idea of building a server.

Now the positions are different and interesting. Tygem is part of a bigger company network of games. In a way, they race Wbaduk to get a stronghold in the west, so i think both sides are interested in Kaya as a means to do that, somehow.

On the other hand Wbaduk is a different company. It is also big, and profitable, but it doesnt have as good a technical team as Tygem has. And more, they get some government sponsorship, which they must spend. So they have money burning their pockets.

Last but not least, as i was discussing with Polly, if there is a way to help them get into the west, but we get into the East, which is a bigger and more profitable market, then we might also find something. Due to our architecture, we have an expanded flexibility to do things that dont compromise our vision, yet allow for inter-server relationships.

WARNING: All this is pure speculation. I have no way of knowing what will happen, what they want, or anything. We are at least very happy on how Kaya spread over the world, and that people can see a glimpse of our vision.