An html5 client would fix all java related issues (including the sound bug) as well as enable playing on iOS devices. I would say that these things have most definitely been asked for. Also people seem to be pushing for browser based clients (even if I don't agree with them). The fact that he keeps a blog of development progress seems to make it harder to claim development is dead.

I'd argue for sound as a "MUST" (No, I would not ever downgrade Java to fix an issue like this). I can appreciate it's just a bit of fluff for some though. I'm appreciative that it's free. I'm annoyed somewhat by it staying one man's project when he doesn't seem to have the time to fix things in a timely fashion but that's his call not anyone else's to make and it's certainly not something I'd give him crap over.

I think if people accepted that they'll never have everything they want in a go server at the same time then there'd be far less heat and a lot more light in the discussions.

I think that one thing that gets glossed over in these conversations is that KGS faced all of these same issues when it was launched. Users didn't just show up and play, they were accumulated over a period of longer than a decade (and even now, it's still not the size of the server it was aiming to improve upon)...One of the big things a new server needs is a single person or a small core of people willing to pump in a large amount of energy for a long time in order to achieve a critical mass. I think a lot of the go servers that spring up have people who have a lot of energy for a short time, then get bored and move on to other things. I'd guess it takes about 5 years or so for a server to really start to get settled in....though perhaps it's even longer...

Yes. Two other factors are funding and culture:
does anyone have the stats on how much funding and time were needed for the recent
Korean/Chinese Go server(s) to get to critical mass ?

I agree with Mef.

WMS can totally do what he pleases. Maybe if he completely broke KGS some KGS+ refunds would be in order, but it's still running and there are players on it playing in real time, as opposed to the 1 live game currently on Nova/OGS/whatever.

You can't even buy SmartGo for Windows right now, but no one is saying SmartGo is dead because it's obvious that Anders is spending his time on other things like SmartGo Kifu and SmartGo books. It works for him, he likes it, and it's popular. If you don't have an iPad or iPhone, you might be tempted to say that no one asked for that, but he has no obligation to make everyone happy.

You might just as well say Java itself is dead. They're just adding junk no one asked for, after all. Lambdas, for heaven's sake. As if anyone writes functional programs to do anything but write compilers for their own pet functional programming languages... (Actually, I like them myself, but that's likely a defect of my brain. "Functional programming combines the flexibility and power of abstract mathematics with the intuitive clarity of abstract mathematics." -xkcd)

I enjoy these discussions about the state of KGS development quite a bit. They usually follow a pattern.

1. Someone says that a missing bug fix or feature means that KGS is dead or dying (e.g., sound bug, one-colour go, speaking count-down byo-omi clocks).

2. People who have never written or maintained a go server agree that the bug fix or feature could be coded almost instantaneously.

3. Different people who use KGS regularly disagree with the original poster, saying that KGS isn't perfect, but it's fit for purpose

4. The original poster or those who joined in step 2 proclaim their love for KGS and/or go in general as an explanation of why they're so critical of KGS and/or its developer.

5. The discussion becomes technical as various contributors display their knowledge of some aspect of IT

6. The discussion gradually grinds to a halt.

7. People continue to log into KGS and play go.

It's more that it's not particularly relevant. Start 5 random go servers 5 years ago and one will become king most likely or more rarely a duopoly will emerge. Just because something is the most popular tells you little about whether its the best. Network effects (play where your friend plays) and simple inertia (I can't be bothered to start a new account and get another stable rank again) will ensure than even a sub-par server can remain top.

None of this is a criticism of KGS, I don't consider it sub-par. It's just being the biggest, or having been small once, aren't very telling other than a) the server is pretty stable and b) some factor gave it an edge in the beginning and this factor could be luck or a feature or simple good word of mouth.

KGS had:
- some initial notoriety due the prior existence of wms's CGoban version 1 (SGF editor + IGS client)
- A much friendlier user interface than IGS that still required typing commands for many operations
- Unique "killer" features like the ability to review a game
- An emphasis on social interactions and chat. KGS is a go community not just a way to play games anonymously.
- Dedication from wms and the volunteers that joined him that worked really hard to make it work.

Even with such clear advantages it took years to convince most people that KGS was worth a try. I remember not so long ago that people at the local go club did not take KGS seriously when I recommended it to them.

In the case of KGS though IGS existed (and had existed for 7 years at the time), was popular, and was stable. In fact IGS is still larger than KGS. The point was the KGS aimed to implement features it felt IGS was lacking and continued to do so for many years regardless of who was playing on it (because at many times the numbers were quite few) until it finally built up a player base. This is where many of the new fall flat (and perhaps even some of the older servers as well). They have the drive and start of implementing things they like to see in a go server (whether it be Fischer Time, WHR rating systems, malkovich comments, etc) however the second part is continuing to do this for an extended period while you build a community. This second part takes years and years to occur, barring some strange catastrophic event (e.g. godiscussions dying and spawning L19). Basically in order for a community to grow, you must have people pumping energy into it. Otherwise it will die, or if you have reached a critical mass, perhaps sustain (a la goshrine).

I disagree with neither of you about KGS's origins. I don't speak specifically about them because, well, I wasn't around for the transition from IGS to KGS by the Western community and have only read about it here and there. My point was that it doesn't matter now. KGS is no longer the young, game-changing server that it was back in the day. It could be quite a bit behind other servers and still hold onto members in the short term just due to the inertia effect you both mention regarding the early days of KGS taking people from IGS. Pointing to server populations as a measure of how good a server is generally is incorrect, just as pointing to IGS in the early days of KGS and saying it was a better server because it had the big population would be equally odd.

I always used think of OGS (RIP) and DGS. I found DGS horrible in comparison to the feature set of OGS, I could never understand why DGS was so, so much larger until I remembered the inertia effects in post-monopoly markets.

You forgot the last one:

8. People continue to get banned by BD.

(Sorry, couldn't resist that. It seems to be a common sentiment on KGS.)

_________________
Still officially AGA 5d but I play so irregularly these days that I am probably only 3d or 4d over the board (but hopefully still 5d in terms of knowledge, theory and the ability to contribute).

really? I know a number of people who turn off the annoying stone-click sounds.

its not really a "must fix" issue. It doesn't prevent playing.

Just chiming in about the sound issue. The moment I don't hear stone placement sound anymore, I stop playing (and observing) games online. For me it is definitely one of the most fundamental things that *must* be there . Without sound it's all too easy to miss when a move has been played on another part of the board. And it's a nice background noise, getting a feel on how a game is progressing if I'm doing something else and have a game open in the background.

/off-topic

Thus the "I'd" and the bit about some people considering it "fluff."

I would really like sound more optimized towards my individual needs for third reasons: first, I need to have some sound on to hear PMs and to realize that I'm in byo-yomi; second, if I forgot to turn the sound off before using KGS and later on I really need to listen to something, the only way to get the sound back on is to reboot my computer; third, I am extremely bad at using byo-yomi time, and a less harsh noise and/or a descending scale and/or actual numbers would make it much easier for me to play a sensible endgame, and maybe to move into playing blitz instead of 9x9 when I'm short of time.

These are why I see the sound as an issue. More broadly, whenever things don't work exactly as they're supposed to, it's an issue. One of WMS's big themes in KGS design has been that he wants to have as few unnecessary choices as possible, so that everything just works. I approve, and I think having sound options that work as you would expect fit into that nicely.

That said, I realize that it's an extremely low priority, and KGS is a one-man show, which is why I'm not too bothered by it. When CGoban4 comes out, I doubt I'll say "Finally! and it' about time" - more likely, I'll think, "CGoban4 already? Where has all the time gone?"



This is a very important point. A lot of energy, over a long period of time, in a project that has a very broad (as opposed to deep) appeal. The main feature of a go server is the other players, which means you need a very broad focus so as not to push anyone away, and a very long time horizon to gather them all in. It seems that a lot of people see the design philosophy of a server as one issue, and its popularity as a completely different issue. So they find a server whose design philosophy appeals to their peculiar tastes ("We've got Fischer Time and Youtube integration! Inconceivable"), and then try to drum up the popular by saying really, really nice things about the server. But one of the things that has made KGS work so well is that WMS has run it for a decade without cluttering it with every new feature that someone got excited about, or without taking on a level of additional work that made his involvement unsustainable in the long run.

maybe I've misunderstood the sound issue -- I thought it was only the stone click sounds that quit working with Java 7. All the other sounds (PM, byo-yomi, etc) still work, no?

(and your second issue is a Windows OS issue, not necessarily a Java/KGS issue)

8. KGS admins chime in with unhelpful smartass remarks, disregarding all criticism and typically saying that "people continue to log in to KGS" is good enough for them.

This is why monopolies in general are a bad thing Even benign, non-profit, accidental monopolies...

Ryuuuun, you wrote
I am tempted to ask, if you regard "dated" as a criticism, why are you interested in Go?

I accept that KGS has its defects. But while it's better than any of its rivals, I don't complain about it.

News to me, I haven't used a Windows machine in years.

There are several different versions of "the" sound problem. Mine is that if CGoban3 uses audio while no other processes are using audio, no other program can get access to the audio afterwards. And yes, to the extent that this doesn't happen to all users, sure, it's Java issue, it's an Ubuntu issue. But this isn't a problem that affects all Java programs, so I imagine that if WMS ever changes sound, the new sound will be designed for the latest version of Java and will fix everyone's problems.

There are quite a few other servers out there. KGS is neither the largest nor the oldest. http://senseis.xmp.net/?GoServers has a list of quite a few, and I'm sure there are others in Asia that aren't mentioned there.

My personal impression is that KGS gets targeted the most because it has enough stable features that aren't implemented well on other established servers. You don't hear too many people complaining about IGS's poor review capabilities, for example.