Life In 19x19

They're not breaking any kind of agreement. They're deleting support for a fairly ubiquitous codec, and replacing it with some that are, at the present time, completely marginal. To the extent that any freely available video formats are going to rise up to challenge Flash, they are going to have to have wide support. Google, unfortunately, is moving in just the opposite direction, for mostly self-interested reasons. We should not be surprised when large corporations act in their own interest.... it is to be expected. But we can safely conclude that Big G doesn't want H.264 to become a Flash challenger. Too bad.... so far it looks like the best shot.

Things are still shaking out, so who knows that the ultimate outcome will be. I think Google is really betting on Flash. There is that little site called YouTube that they have as a side interest....

h.264 is royalty-free for playback, so it's not really a patent or royalty issue. It looks more like Google is just taking a battle and trying to prove they can win, even though h.264 has massive support.

amnal wrote:Your desired result is for h.264 to challenge (and defeat) flash. This seems to be just a subset of the more general goal 'something better than flash, and as good as possible in and of itself, should challenge and defeat flash'. It isn't obvious to me that the second goal has been compromised.

Of course it has been compromised. Before, everyone encodes in .h264, if websites want to, they serve h.264 to viewers in a <video> tag, with Flash as fallback. Opera and Firefox get Flash, (and note that <video> adoption is heaviest for mobile, since Flash is rare and especially slow in that area). But now you've just increased the number of people who need Flash to serve h.264. And you can't serve WebM video to an even wider range of people. So you're stuck dual-encoding to get close to a good number of people.

You can argue about blame. There's an open source or anti-patent perspective that says Apple and Microsoft should budge and support WebM. There's an open standards or ubiquity based argument that Chrome should budge (I think Firefox just can't because of their approach towards patents--is that right?). But you don't have to answer that question to know that <video> just suffered a serious setback. Whether it'll stagnate or just slow down is an open question. But it can't be good for the adoption.

The closest I can see to an argument that doesn't matter is that Firefox is a big browser, so it doesn't matter that Chrome is in the same boat. But Chrome really is showing remarkable growth.

Bantari wrote:Just consider: IGS opened its protocol to allow exactly what you are suggesting KGS should do. And even though IGS is much older, it still does not have a decent client, and when it does - it gets broken and not-supported within a year, and the server is still bound to the ancient telnet protocol, and the whole thing seems to be just going downhill. What makes you think KGS will fare any better if it opens its protocol?

Don't you think that Tetsuki is a pretty decent client? If not could you please specify what you call "decent" in a client?

After twenty minutes or so, Tetsuki crashes for me every time I try to watch a pro game. So my experience has not been decent. YMMV

hyperpape wrote:After twenty minutes or so, Tetsuki crashes for me every time I try to watch a pro game. So my experience has not been decent. YMMV

I use Testsuki on a regular basis with iOs 3 and got no troubles at all playing or watching

By the way... There is an alternative (unauthorised) Java client actually out there. I've seen it in action. I believe it was called "KGS Hacked". From what I saw, it's just the existing client, but then someone reverse-engineered it and added some fancy colours and extra doodles to the chat functionality. Since it is quite doable to reverse-engineer the protocol, it should be also possible for an amateur to develop a simple (unsupported) iPad client, right?

Reason why it will never happen: you'd need someone who does not only have a lot of spare time, likes go and computers, and is good with programming; but who also uses unnecessary expensive low-quality gadgets such as the iPad and is insane enough to comply to Apple-verse. Dream on

gaius wrote:By the way... There is an alternative (unauthorised) Java client actually out there. I've seen it in action. I believe it was called "KGS Hacked". From what I saw, it's just the existing client, but then someone reverse-engineered it and added some fancy colours and extra doodles to the chat functionality. Since it is quite doable to reverse-engineer the protocol, it should be also possible for an amateur to develop a simple (unsupported) iPad client, right?

Actually, no. In the hacked client pasky didn't reverse engineer anything; java binaries can easily be turned back into raw java code, although it ends up kind of garbled. Pasky turned the java binaries into (garbled) java code, then figured out how to make the modifications he wanted, then compiled it back. It would certainly be possible for somebody to reverse engineer the whole protocol from the garbled java code and thus come up with an ipad client, but that isn't what pasky did as far as I know.

Did you see the changes in the hacked client?
You can scroll the moves of the game with the mousewheel, more readable color-coded games lists, and auto-joining of high-rank games.
These are all trivial to implement but provide a huge usability bonus. What is annoying, is that merging this stuff into the main client would take me like 1 hour, and Java is not even my main programming language. Of course, the current version is so outdated, it doesn't support tagged games, but for me that is a favorable trade-off. I hope the mentioned protocol changes don't break this peaceful coexistance

I've seen them. And yes, none of them are huge things to do. The mousewheel is the one thing I'd be most likely to do, that's pretty neat. Most of the others, they just aren't what I would want myself.

wms wrote:

gaius wrote:By the way... There is an alternative (unauthorised) Java client actually out there. I've seen it in action. I believe it was called "KGS Hacked". From what I saw, it's just the existing client, but then someone reverse-engineered it and added some fancy colours and extra doodles to the chat functionality. Since it is quite doable to reverse-engineer the protocol, it should be also possible for an amateur to develop a simple (unsupported) iPad client, right?

Actually, no. In the hacked client pasky didn't reverse engineer anything; java binaries can easily be turned back into raw java code, although it ends up kind of garbled. Pasky turned the java binaries into (garbled) java code, then figured out how to make the modifications he wanted, then compiled it back. It would certainly be possible for somebody to reverse engineer the whole protocol from the garbled java code and thus come up with an ipad client, but that isn't what pasky did as far as I know.

What are the chances we could convince you to merge back some of the changes? I would love to use the official client, but I simply can't go back to some of the features such as the annoying ring, the old clock, and the lack of colored text and opponents. I would buy a couple years of KGS plus just for this to be officially supported...

The chances are very small. Sorry but that's what it is. Features that I personally would not use and would not want to see are unlikely to be implemented.

Subject: Kgs on iPad

Just use this solution & everybody are happy now.
viewtopic.php?f=18&t=3648

Thank you

hyperpape wrote:

Tofu wrote:With the price of netbooks being what they are, I'm still confused about what the point of the ipad is. What does it do thats so special besides looking cool so you can show off at the coffee shop?

Actually, it's a little known fact, but the iPad isn't even a real computer. It just displays semi-random animations in response to user input, solely for the purpose of showing off at the coffee shop. Apple collects $500 profit on a $600 device because the internals are so simple.

Apple collects revenue from coffee trade groups because iPad owners spend so much time sitting in coffee shops admiring the logos on their Apple products.

The form factor and UI of the iPad or Galaxy Tab, makes it easier to sit and browse the net, as easy as reading a magazine. Most of the time, you do not need a full keyboard in front of you, so it makes sense to eliminate it and bring it up on screen when needed. Elimination is a Kaizen principal, see?
Also, reading stuff is better on a tablet, because it looks like a book and can be held in a portrait orientation. The horizontal orientation of a netbook makes you scroll a lot.
Add to that amazing battery length, as opposed to a net book, and there you go.

Bottom line, if you have a net nook and a tablet and you are on the couch wanting to check prices at kurokigoishi, more often than not, your hand will extend towards the tablet.
That is what happened to me.

I've started working on an idea to have KGS on the iPad. Using the Amazon cloud to provide a VNC connection to the KGS client.

If anyone is interested in testing and/or helping with this please let me know.

So far we have a server running on the Amazon Cloud that allows access only to the KGS client via VNC. The KGS client starts immediately and is full screen. The next step would be to allow multiple users to access independent versions of the KGS client.

There are free "lite" iOS apps that provide a VNC client, I've tried Mocha VNC Lite on my iPad and that works well. The paid version is $5.99

I am wondering if there is enough interest to keep a server running on Amazon to provide this service.