Sim Server Press is great... but wrong.

[vrm]

IMO it's a bit opaque developpment for talks about 'standards',
wait & see

[abies]

You might also see a free SDK that let you run a single local server stack for development but not for deployment.

I think it is a must. Similar to way in which many J2EE servers are made available - they are free for development, you have to pay for deployment. Of course, most of people trying it out will never become paying customers - they will just want to play with it, their projects will get scrapped before release etc. On the other hand, you may get more people actually using your technology, because entry is free and IF it is good, they won't imagine living without it. But even people who will fail to deploy finished game will give something in return - free 'support' on boards (answers to newbie questions how to do this and that), bug testing etc.

I'm not sure about limiting it for single server. People should be able to test and debug how application works in scalable environment. IMHO, requiring license to run it in open is protection good enough - after all, it is not a MS Word or single-player game, which you can use just for yourself inside a dark cellar. J2EE/database vendors offer free development environments and have no problems with piracy after all...

[Jeff]

IMO it's a bit opaque developpment for talks about 'standards',
wait & see

Absolutely.  Today its not yet fully formed and tested so tryign a standardization effort woudl be a huge mistake.  I rather expect it will go (if everything goes as I'd hope) from birth through the fires of ral development to a defacto standard that eventually could be canonized in an API/programming model standard.

Bit thats all a ways off.  The poitn as not to announce a "new stadnard" today but rather to talk abotu what the eventual benefit of a standard could be.  Sorry if that was unclear.

Games published by our own members! Check 'em out!

[Jeff]

I'm not sure about limiting it for single server. People should be able to test and debug how application works in scalable environment. IMHO, requiring license to run it in open is protection good enough - after all, it is not a MS Word or single-player game, which you can use just for yourself inside a dark cellar. J2EE/database vendors offer free development environments and have no problems with piracy after all...

Thats a good point.  The question  there wil lbe twofold: (1) what third party technologies are invovled and how those aprtners feel about.
(2) the practical issues of buiulding a "real" test backend.  Thats not cheap.  You're talking at least a 5K blade rack, a couple of 1.5K-2K blades and at least one larger machine to do master database duty.

I'd rather see those facilities be available at the epicenter myself for simple cost reasons.

[abies]

(2) the practical issues of buiulding a "real" test backend.  Thats not cheap.  You're talking at least a 5K blade rack, a couple of 1.5K-2K blades and at least one larger machine to do master database duty.
I'd rather see those facilities be available at the epicenter myself for simple cost reasons.

For deployment - sure. But for development, having machines on site and being able to do anything you want with them is necessary IMHO. You will probably replace .gear (Game Enterprise Archive ) every few hours, possibly introducing database-hogging/corrupting mistakes, which you need to handle yourself. Calling outside support every day or two to ask for database restore/change for different branch/etc would cost more in support than these blades after one month

Cost of servers can be smaller. Remember that they CAN crash - you can use bottom-shelf, economy equipment which is fast enough and don't care if it is really bulletproof - after all, you just risk one hour of your work for reconfiguration, not one hour of angry customers not being able to log in.

I really think it would be an ideal solution. Free development environment, with full docs/forums/etc online. Cash would flow from 3 routes - paid support for companies which will decide to use technology and will like to speed up development, full license purchase from big players who want to build entire gamecenter themselves and royalty payments from smaller companies which will rent gamecenter provided by you. Of course, route 1 and 3 can (and probably should ) be partially taken by third party companies, which will in turn become revenue 2 for Sun.

[Jeff]

For deployment - sure. But for development, having machines on site and being able to do anything you want with them is necessary IMHO. You will probably replace .gear (Game Enterprise Archive ) every few hours, possibly introducing database-hogging/corrupting mistakes, which you need to handle yourself.

Hmm. Okay and why does there need tro be more then one server stack for this poitn in testing? I'm just curiious.  Multiople stacks adds to your compelxity of setup and maintainence.  I tested witha  single stack for the GDC demo all the way until i needed to start doing fail-over tests...

Cost of servers can be smaller. Remember that they CAN crash - you can use bottom-shelf, economy equipment which is fast enough and don't care if it is really bulletproof - after all, you just risk one hour of your work for reconfiguration, not one hour of angry customers not being able to log in.

True enough.

I really think it would be an ideal solution. Free development environment, with full docs/forums/etc online. Cash would flow from 3 routes - paid support for companies which will decide to use technology and will like to speed up development, full license purchase from big players who want to build entire gamecenter themselves and royalty payments from smaller companies which will rent gamecenter provided by you. Of course, route 1 and 3 can (and probably should ) be partially taken by third party companies, which will in turn become revenue 2 for Sun.

Nice sumamtion and Im not really arguing much.  Assuming the economcis of (1) coudl be worked out I see no problem with this sort of free development license.

Now we need to convince management

[blahblahblahh]

Now would someone please explain why customers won't flock to Butterfly instead? If you don't mind being ensnared by this business model, why not go with the established name with the established technology and the partner relationships already in place? Oh, and being backed by an $88 billion company? Sun needs some kind of USP here, AFAICS, and until that white-paper appears only Jeff can tell us what that is...

[Middy]

Well, perhaps developers dont care too much about standards. But when seeking for funding, it means ALOT to be able to write.

"API is SUN certified"

Investors look for highest profit lowest risk. So LWJGL being a open Source project it would help ALOT to say. Dont worry sun checked it.

Or at least... this,this and this super great game uses it....

[erikd]

Well, perhaps developers dont vare too much about standards. But when seeking for funding, it means ALOT to be able to write.

"API is SUN certified"

Investors look for highest profit lowest risk. So LWJGL being a open Source project it would help ALOT to say. Dont worry sun checked it.

Or at least... this,this and this super great game uses it....

Hmmm, Sun not being a major player in the games business (yet?), I don't think it would at all make a difference wether or not LWJGL is Sun certified or not. Like "hey, Mercedes Benz checked it: this is a great TV"   (shaky analogy, I know).
As far as risk goes, a company behind LWJGL (puppygames?) being able to sell support for it would help a lot more. I believe LWJGL being open source doesn't affect it's risks negatively in itself, I think it's even an advantage.

[aNt]

rrrrr!! - the sodding Jeff Game Server thread  for my
pointlessly yet simple questions.  
 
so when it comes to development i see it runs on the
Blaze thing'e's with the Fires pluged in? what model?
 
when can we start on an NDAed dev test with your  
Games Server?- i dont mind NDAs if i can get on with  
making cool stuff.
 
how many players (at a guess) per Fire?- so we get  
that bad typo out the way from the chat 20th. 100  
sounds pritty small. i dont think i can aford the floor  
space for that much hardware.
 
what do u mean by Object comunication? as in how  
should i think about making something today that is  
along the write path to use your Games Server so when  
you do get the dev systems sorted out i am in the write  
mindset (RMI, WAWAWA).
 
what do u dudes think are a 'big and safe' games  
companys? the only company so far that have pulled off  
a good working mmpog (no shards or regoins) is CCP
with..... *stops himself* i have chatted about this area
to many times in big conferences around the world- i cant
be bothered to go into it here.....
 
Now i did try and look into GrexEngine but i couldnt find  
to much info. Blahblahblah i think is well in the zone  
when it comes to mmpog games server stuff. and i
would love to see how all that jazz works. but my NDA
is such cant tell anyone what i am realy up to. Also what
i have read from Blablabla posts he cant say much  
about it. shame all round  
 
but i would love if we can keep this thread a happy one  
its probably one the coolest area of gameing today.  
SUN and Everyone here are fantastic! and more then
one thing is always a good thing... take the bleeting to
another thread- ied love to join in the banter  

Games published by our own members! Check 'em out!

[blahblahblahh]

"good working mmog" IMHO has little to do nowadays with shards and regions (although "scalability" still has a lot to do with those features). Jeff is of completely the opposite viewpoint, IIRC, believing that these are almost defining symptoms of a good MMOG (c.f. the previous discussion on how Sun Sim Server press claimed that getting rid of shards etc was the solution to all MMOG problems that everyone had been waiting for).

Aside from that, could we perhaps keep discussion on this to one thread ? There's one over in the networking forum already...This one here is already 5 pages long and I think it would be easier for new readers to have a shorter one that is more current.

I understand that at the moment, with no noises coming out of Jeff, and the chat getting cancelled/curtailed, it seems sensible to hit every thread in the hope that one of the posts will elicit a response, but if we move everything to one thread, then we won't need to keep doing that .

http://www.java-gaming.org/cgi-bin/JGNetForums/YaBB.cgi?board=Networking;action=display;num=1088524121

[blahblahblahh]

i dont mind NDAs if i can get on with  
making cool stuff.
...  
Now i did try and look into GrexEngine but i couldnt find  
to much info... and i
would love to see how all that jazz works. but my NDA
is such cant tell anyone what i am realy up to.

AFAICS if you can't say *anything*, you won't be able to get any support, and you won't be able to get any pre-sales advice, in which case the project can't be particularly serious (because you're bound to end up using the wrong technology, or abuse it, unless you happen to be a specialist in high-end servers - in which case, you probably don't want to use 3rd party servers anyway?). Or perhaps you've been given $X million by SOE for a hush-hush skunkworks project where the mere mention of the name would cause problems.

If you have a specific project and genuinely believe that middleware could help, then you should contact sales @ grexengine.com. Although they will probably need a 2 page game-concept doc from you, in our experience people who can't produce a concept doc aren't ready to consider writing a game yet .

i have read from Blablabla posts he cant say much  
about it. shame all round  

I'm only limited by three things:
 1. practicality: I'm not the one responding to sales queries etc, we have a process for that which is done by other people. If I try being too verbose in e.g. a forum like this one then I might easily confuse things compared to the sales guys who have it all pat.
 2. I can't explain any algorithms and processes that are covered by as-yet-ungranted patents.
 3. Time

[aNt]

only a 2 page game-concept doc? woo i think ours
needs a small out building to store it in now

SOE shit i've been rumbled!

[blahblahblahh]

only a 2 page game-concept doc? woo i think ours
needs a small out building to store it in now

Nah, that's your game *design* document . Or "bible" as it's sometimes know (also sometimes "big blunt instrument for beating people up").

The concept is to a design what an executive-summary is to a business-plan

[aNt]

yer i no dude - ours is about 10pages because we like
pictures... the big fella 'design document' well catch up
with me later in life without to much of a problem.

[aNt]

jeff u out there?

[Mr_Light]

"aNt: When can we start using the Sun Games Server?

Doug Twilleager: The technology is still in its prototype phase. We are still working on product schedules, but sometime late this year or early next could be possible for early access. "

"abies: Do you plan to have a free-for-development edition? Something you can play with and pay only when the game goes into production?

Doug Twilleager: That is the goal. We have even talked about having a public playground of sorts, where large betas could be tested without incurring large costs."

from http://java.sun.com/developer/community/chat/JavaLive/2004/jl0720.html

I take it that still stands? (the time frame late this year or early next could be possible for early access.)

and that early access, covers that 'playgound' (if you get approval)

for completion, the white paper you referered to in your opening post is this one: https://games.dev.java.net/docs/simserverwp052604.pdf (right?)

so it goes by the name sim server. Is there a portal/homepage where all information is orderd on or so? if so whats the url?

to clarify; with the use of the sim server tech you can avoid a split between the contients or regions (NA, europe, Azia) or where those regions as refererd to 'virtual'/ingame regions?

wenn will the whitepaper be updated/expanded?

@jeff just out of curiousity: do you type on a laptop?  

[bbqplatypus]

GIMME GIMME GIMME!

I NEED! I NEED!