Good job guys!

[Mef]

That's quite a nice way of putting it, although I think you may also be overstating it a bit Kaya is not THAT focused on doodads. However, I do agree with you that the focus should really be on the absolute basics - rock solid game and scoring system, robust/reliable time system, competent and trustworthy ranking system, solid handicap system, and most importantly, a decent player base. I don't think the kaya guys realize the extent that they actually already have a killer feature, namely an HTML-based go server that runs over HTTP. The other stuff like tournaments, videos, karma and sensei's integration, while undoubtedly cool, and important in the long run, are far less important right now.

I honestly don't mean to be argumentative...but if anything I feel I might have been understating it.

edit: This was a list with a variety of examples, but after reviewing my post I think it will be more effective to focus on one that I think best demonstrates my point:

-I was trying the SGF editor kaya provides and loaded a recent commented game from the AGA E-Journal, and it would appear the SGF editor does not support any SGF markup (Black should play A...but there is no A?!).


So this means Kaya has:

-Hourglass time
-A button for translation
-Sensei's Library integration
-A fully functioning karma system
-Ways of filtering go vs. non-go comments
-"Malkovich" Comments by game players
-An expandable widgets menu I'll admit I don't fully understand (This is for integrated video?)
-ASR integration
-Active previews of ongoing games in the games list
-Automatic replays of professional games


But lacks:

-Basic SGF markup support (Labelled variations, marked stones, etc)--in spite of the ability to load an SGF file.

I think this is the perfect example of a feature many would expect, but never think to request or vote on. I also think that if you explicitly presented this option to the userbase, many would consider the latter much more important to a go server than many of the former options.

[badukJr]

The biggest reason that Kaya doesn't retain any users from the account release is that its a chatterbox server not a social go server. The backbone of a go server population are people who just come and play games without ever saying anything. Kaya doesn't offer these people anything, because it only offers things to people who are loud about their opinions. The bloat load that Kaya puts on computers with tons of fluff really discourage people who have older computers. So, people who just want to play go don't play at Kaya because its laggy as hell and their CPU is getting slammed.

The other thing is that Kaya is not really a social go server. Not many people want to chat in the main room of a server, they want to make their own room so that their clique can enjoy each other's company without others butting in. You can't do this in Kaya, which is why right now its a chatterbox server. Nobody wants to move to a server where their group can't move to. This is why the server is so ASR heavy, a place has been given to them so they have come.

There's still room in beta.

I just wanted to point out that you responded to Charlie's post with karma begging in your signature. Maybe you were aiming for that though.

[shapenaji]

I actually think the problem with Kaya is simpler than that it's a chatty server, or it lacks some review support.

The lack of a negotiation step when it comes to games.

It's something I take for granted everywhere else. Basically, if I put up a game request on there, I have no control over who I play. As a result, I feel like most games played are a result of negotiation in the chat room.

It's devolved into a direct challenge system. A direct challenge system can work for a server like IGS, which is so old that people are willing to accept some quirks, but it just won't work for a new server that's trying to get people to hang out and spend some time there.

[Charlie]

The reason that I mentioned ASR integration was *not* because I have anything against the ASR - in fact, I consider it to be a good idea in essence. However, it definitely polarizes the user-base, splitting the minuscule set of active users into those who are in and those who are out.

The bottom line is this, though: at any given time of day, what are your chances of getting a game on Kaya?

Every feature developed now should aim at improving that chance. Everything else can wait.

[quantumf]

Every feature developed now should aim at improving that chance. Everything else can wait.

Fair point, but much harder to clarify. Presumably this is a function primarily of the size of the userbase, and to a lesser extent, the degree to which people are happy to play on the server, which involves things that I outlined above, and others have also mentioned (e.g. good ranking system, good negotiation system). The size of the userbase is currently being explicitly limited, quite sensibly I believe.

At the moment you're a beta tester, not a regular player, and as such should really be commenting on bugs, not the ease of getting a game (although bugs or usability issues that make this hard should be raised, such as the negotiation system). I'm sure the early days of KGS or IGS were similarly ghost-town like.

[daal]

I think the biggest problem at the moment is stability. The server has gone through periods where it is unreachable, and this has severely cramped its style. Apparently this has been caused by some third-party add-on, which supports the general sentiment that it would be good to focus on basics before bells and whistles. On the other hand, it's clear to anyone who frequents Kaya, that Gabriel is hard at work to fix whatever problems arise, and to date has consistently managed to do so relatively quickly.

As to the basics - the quibbles seem minor to me. You can play and review on Kaya, and there is a functioning rating system. Sure, negotiating games and better review tools are necessary, but they don't seem like much of a hurdle - neither to people testing out the server nor to the developer. You are probably right that putting the cool stuff on ahead of making everything solid is putting the cart in front of the horse, but I think that most people who are watching Kaya's development think that its got a pretty nice cart, and the horse is growing stronger every day. The server is updated every single week. Also, some of these features such as built in translation and malkovitch comments are powerful tools for enhancing the online go experience. Imagine a pro making malkovitch comments in Korean and everyone can read them.

Don't forget also that what people are missing atm is not a place to play - KGS is still out there - but rather a place where somebody is interested in implementing interesting features that other servers apparently aren't interested in. By putting the cart before the horse, Gabriel has demonstrated his willingness and ability to do so.

There's still room in beta.

I just wanted to point out that you responded to Charlie's post with karma begging in your signature. Maybe you were aiming for that though.

Yeah, that is a bit lame.

[quantumf]

[quote="daal"]Don't forget also that what people are missing atm is not a place to play - KGS is still out there - but rather a place where somebody is interested in implementing interesting features that other servers apparently aren't interested in. By putting the cart before the horse, Gabriel has demonstrated his willingness and ability to do so.[quote="daal"]

Yeah, good point - except for the group of people who literally cannot play on KGS because of firewall or java problems, there's no urgent need for for kaya to focus on the playing basics if they believe its more important for their success to get some of the more advanced or unique features working.

[shapenaji]

I think the biggest problem at the moment is stability. The server has gone through periods where it is unreachable, and this has severely cramped its style. Apparently this has been caused by some third-party add-on, which supports the general sentiment that it would be good to focus on basics before bells and whistles. On the other hand, it's clear to anyone who frequents Kaya, that Gabriel is hard at work to fix whatever problems arise, and to date has consistently managed to do so relatively quickly.

As to the basics - the quibbles seem minor to me. You can play and review on Kaya, and there is a functioning rating system. Sure, negotiating games and better review tools are necessary, but they don't seem like much of a hurdle - neither to people testing out the server nor to the developer. You are probably right that putting the cool stuff on ahead of making everything solid is putting the cart in front of the horse, but I think that most people who are watching Kaya's development think that its got a pretty nice cart, and the horse is growing stronger every day. The server is updated every single week. Also, some of these features such as built in translation and malkovitch comments are powerful tools for enhancing the online go experience. Imagine a pro making malkovitch comments in Korean and everyone can read them.

Don't forget also that what people are missing atm is not a place to play - KGS is still out there - but rather a place where somebody is interested in implementing interesting features that other servers apparently aren't interested in. By putting the cart before the horse, Gabriel has demonstrated his willingness and ability to do so.

Don't get me wrong, I think the bells and whistles that have been added are amazing (and bloody useful).

The problem is that there was a surge of excitement when Kaya went Beta and that surge could have been channeled better. The server was getting pretty close to critical mass for a while, and had some of these basics been implemented, I think we would see people playing numerous games on it today.

Instead, I feel Kaya has wasted a big opportunity, and while it may get another one, I feel that folks' excitement will be tempered now by some less-than-stellar impressions the first go-round, which may slow its growth.

[Kaya.gs]

I actually think the problem with Kaya is simpler than that it's a chatty server, or it lacks some review support.

The lack of a negotiation step when it comes to games.

It's something I take for granted everywhere else. Basically, if I put up a game request on there, I have no control over who I play. As a result, I feel like most games played are a result of negotiation in the chat room.

It's devolved into a direct challenge system. A direct challenge system can work for a server like IGS, which is so old that people are willing to accept some quirks, but it just won't work for a new server that's trying to get people to hang out and spend some time there.

You have a very decent level of control over who you play. You can pick the specific time settings and handicap of the game that is going to be played.

We thought a lot on a way to avoid the negotiation system that has a few issues:

- It makes both players wait (the offerer and the challenger) meaning that it takes longer to start a game.
- User lockup: If the game offerer is afk, the challenger sends an offer and the game doesnt start, and he waits for acceptance. So now he is wasting time. Then, a third user could take that very same challenge, and wait. Hence 2 able players are waiting for someone afk.
- Extra user required actions: there are many user actions required to start the game.
- Multiple offering: as you wait and you are qualified to take several games, asking for a game on all of those is a common practice. This is cluttering, making many simultaneous dialog windows. Also tricky, the offerer can see people coming and going because of this efect.


The elimination of negotiation means that start games is faster and easier than ever which is evident on account releases, where people enter the server and start games within seconds.

Covering all the aspects that the negotiation system on KGS does

In the future, with blacklists you can avoid specific players you dont want to play, and also we plan to put an option to play an absolute even game(with whoever accepts).

That said, the Kaya system is not strictly better than the negotiation system: there is a specific consequence. A game offerer does not get to pick among a varied group of qualified challengers. The offerer doesnt get to see 2 players of his own rank and gets to pick who he wants to play.
I dont regard that benefit as a positive one. Being able to pick means that 2 players that could play each other are waiting the game offerer.


What I do regard as the biggest con is that regular online players are not used to it. Particularly what i see happen is that people make a game offer, and they dont expect the game to start, so they go make coffee, go afk. Then when they come back, they expected to pick challengers from a list, but the game had started, and the clock got running, so they get upset.
Maybe what we should add is info in some sort of way.


There is a first version of game reviewing, which is also high on the priority list for improvement. Adding markers, and saving the comments properly on the sgf, plus better navigation, are all listed as TODO's in the next few versions.

I think you havent visited the server in a while to see the new things, shapenaji.

Lastly i agree with daal. The switch to https was rocky, and a few issues from third party software have arised frequently lately. It is my top priority for Kaya to run super smooth, as it was running 2 months ago and even better. We are wrapping up ways to dissipate the effects of the bugs on the third party libraries.

Let me just say that we have a long way to go to increase the user base. Everyone that finds something they don't like naturally think that the lack of this or that feature is the root cause.
If you compare the basic rate of accounts registered vs the number of people online in any server you will see that less than 2% of registered accounts are online at any given moment.

All feedback is appreciated and i had a great time reading all these posts. Bear in mind that Kaya improves a lot each week, its a very live project and the donations and contributions of people supporting this project has helped enormously in the development of it.

[averell]

You have a very decent level of control over who you play. You can pick the specific time settings and handicap of the game that is going to be played.

That makes no sense at all. It sounds like you have 0 control whatsoever over *who* you play.

Quick match and such is all fine, but no way to confirm an opponent is just a terrible idea.

[quantumf]

Regarding averell's question, do you intend to offer the ability to do direct challenges, that, say, Tygem allows? Because there are definitely situations where I want to play a certain person, e.g. a friend, or a tournament opponent. The KGS system is pretty annoying, but admittedly, the Tygem situation is also annoying (getting spammed by endless strangers). I'm not clear that there is a perfect solution...

[averell]

Direct challenges already exist. Still, the current system is equivalent to Automatch on KGS (for the one making the game). Selling the lack of custom game offers as a feature is just silly.

[quantumf]

Direct challenges already exist. Still, the current system is equivalent to Automatch on KGS (for the one making the game). Selling the lack of custom game offers as a feature is just silly.

If direct offers are available, as are automatch offers, and blacklisting/censor lists, what other situations are covered by KGS-style custom games? i.e. what problems does the lengthy negotiation system of KGS solve?

[averell]

If direct offers are available, as are automatch offers, and blacklisting/censor lists, what other situations are covered by KGS-style custom games? i.e. what problems does the lengthy negotiation system of KGS solve?

You have no individual control over your opponent. For example I like to discriminate users who's name starts with vowels. Other people only want to play people with pics. FYI censoring on KGS does block those people from being automatch opponents. There is no confirmation when the game starts. If you've ever used KGS, it's just like automatch (for the one offering the game). What is being sold as a new feature, is the lack of opponent/game start confirmation.
Of course then the game starts immediately but let me just say: what is this, i don't even...

[quantumf]

You have no individual control over your opponent. For example I like to discriminate users who's name starts with vowels. Other people only want to play people with pics.

Hmm. At first I thought you were joking, but on reflection, I think you've just been a bit opaque, perhaps as a test to me I presume what you really mean is that you want to be able to avoid playing people who have (to you) offensive names or images. OK, fair enough. To me this would not be a high priority item, but I suppose for some it would be.

FYI censoring on KGS does block those people from being automatch opponents. There is no confirmation when the game starts. If you've ever used KGS, it's just like automatch (for the one offering the game). What is being sold as a new feature, is the lack of opponent/game start confirmation.

Agreed, the blacklist would need to be apply in all situations. Censoring and blacklists are not identical things, tho.