Space, mainly… I’ve got to share it with a lot of other stuff. I don’t want to have to put any more effort into changing my installer and having multiple versions of my code out there at the same time. And most of all I’d like to be able to distribute it from my own site, kosher.
Cas 
I don’t think we have EVER refused to acknowledge the demand and opportunity here. Ever.
We KNOW it’s a huge makert, but we are dealing with an equally huge problem. Would it be better to have people DL’ing a new custom VM with each game they play or have a large onetime DL with MUCH smaller incremental updates
and, thus, each game they dl is smaller? What happens to the developers who don’t use the LWJGL solution? Are they stuck in the same boat as we are now or do they create their own solution around Java, further fragmenting the market and brand?
It is not just as simple as handing over a $500 license to do what you want with it. Unfortunately, the license for embedded Java is much more.
With regard to distribution, we have been working VERY hard with pc manufacturers to have Java ship on their systems. In fact, Cas, you recently told me about the cheap Dell system your company purchased that came with Java 1.4 preinstalled. So the distribution efforts are not ENTIRELY broken 
-ChrisM
[quote]Space, mainly… I’ve got to share it with a lot of other stuff. I don’t want to have to put any more effort into changing my installer and having multiple versions of my code out there at the same time. And most of all I’d like to be able to distribute it from my own site, kosher.
Cas
[/quote]
Cool! So the issue is that the JRE size impedes what you want to do from a business sense. I respect that, no problem. So why not have the JRE included on the CD and have the game DL from your site after it is installed?
-ChrisM
Doesn’t really seem to work that way with the punters. Honestly, you have no idea just how simple things have to be for 99% of the market. (Hurrah! Finally I can actually back this up with some acumen!)
We’re not talking about fragmenting the Java brand at all. We’re talking about getting developers to use the technology and enabling them to use it until your eggs have hatched. The new Dell was a nice surprise but there are very few new Dells in the world relative to the existing installed userbase and even if every Dell from here till Kingdom Come ships with a JRE on it it’s still a massively reduced market to target. As opposed to when .net becomes the standard on Windows.
Look at it in terms of these percentages then:
95% of all PCs have Windows on. Ergo, eventually 95% of all PCs will have .net on them eventually. My market, if I target .net, is therefore 95% of the PC market.
50% of all new PCs or so are going to ship with a new JRE on them thanks to your efforts. Ergo, eventually 50% of all PCs will have Java on them eventually. My market, if I target Java, is therefore 50% of the PC market. I will make twice the money from .net.
So I’d really rather like to have the option to buy a $500 license to distribute my cutdown VM. It’s embedded btw, it can’t be used for any other purpose other than my one game, and it won’t even run anything else particularly. C’mon! How many companies have people trying to give them $500!? to buy something that costs nothing! Help me out here.
Cas
[quote] 95% of all PCs have Windows on. Ergo, eventually 95% of all PCs will have .net on them eventually. My market, if I target .net, is therefore 95% of the PC market.
[/quote]
This is a false statement because .Net will not run on all those PCs. It wil not even make it on to %50 of THOSE, even MS has said this.
Ergo, the .Net market is not %95 until %95 of the PCs being used are built after now, which most likely will never come to pass, and if so, longggggg timmmmme.
Yes, but his logic for Java being on 50% of the PCs is flawed in the exact same way - so that sort of cancels out and you are left with the MS monopoly putting .net everywhere (that matters to Cas), Whereas Java will still be significantly less than everywhere. Which is partly why he chooses to bundle a JRE… leading to issues with the size of that bundle… etc.
[quote]Ergo, the .Net market is not %95 until %95 of the PCs being used are built after now, which most likely will never come to pass, and if so, longggggg timmmmme.
[/quote]
Java is currently shipping on approx 60% of all new PCs, a percentage that continues to rise as almost every day new Manufacturers are agreeing to OEM the JRE. By the time .Net is shipping on all 95% of Windows machines, Java will be shipping on significantly greater than 60% of all new Windows (and Linux and Mac) PCs
Independent of OEM’d JVM’s we’re currently seeing over 2M downloads of 1.4 per month from Java.Com. 1.4 includes the auto-update feature, so all these PCs will continue to have the latest version of Java installed.
This doesn’t change the problem of enabling developers of casual games to take advantage of the latest JRE today, which I think is a valid issue, but it’s not one that MSoft has solved with .Net.
Believe this news. While the numbers I hear still isn’t 60%, it is significant on new PCs. The right content mix can get it out in other places as well. For example, I had a project that used JDK1.3.x for AOL. It was a requirement and as such it had influence on some things that AOL had to make available since they wanted to use the content.
Its 6 of 1, half dozen of the other. In order to work with some of the publisher types you have to have the whole Java issue sorted such that it doesn’t become a support issue. They definitely frown on anything that can potentially cause a support issue of any kind.
While getting Java onto the client platform isn’t one of my concerns, I do know that it is for more than a couple of publisher. Especially the smaller ones. The one issue that Microsoft did solve with .Net is providing an avenue to the console market. Since Xbox2 is still not beyond the veil of NDA I won’t comment - but lets just say that it makes .net an appealing choice. Its not one that I’m particularly interested in right now, but is a path that is large enough to not be ignored as a valid option going forward. Especially for commercial projects. While managed DirectX cannot be guaranteed on PCs until Longhorn (at which time it will be 100%), its pretty much a given that it will be 100% on Xbox2.
Precisely - the Xbox clinches the argument, completely. This is a major reason why smaller studios are not considering Java in their development (I discussed this at length with Pivotal Games in an interview 2 years ago). If it has to be rewritten to run on a console they might as well have started with C++ in the first place. Now they’ll be able to move in the direction of a VM, but it’ll be .net.
And 60% of all new PCs is not a good number. It’s a terrible number compared to 100%, which is the availability of PCs that can run a C++ application, or a Blitz Basic application. The sales of a game are based on factors, not addition. 65% of PCs have OpenGL drivers, 60% of them eventually have a JVM… so that’s suddenly looking like 39% (I know there’s some overlap, but my stats seem to show that it’s surprisingly small). But there it is, in black and white: you’re competing against technology with 100% availability and penetration and it makes sound business sense to go with 100%, not 60%.
Cas
[quote]And 60% of all new PCs is not a good number. It’s a terrible number compared to 100%, which is the availability of PCs that can run a C++ application, or a Blitz Basic application.
[/quote]
60% of the new PCs is a good number for a nice 3d Java game, which won’t run on older PCs anyway.
[quote]The sales of a game are based on factors, not addition. 65% of PCs have OpenGL drivers, 60% of them eventually have a JVM…
[/quote]
Any PC which could run a nice 3d Java game will have an OpenGL driver because most of them use an ATI or Nvidia graphics card.
A Java game using modern 3d graphics addresses the same market full-price 3d games do, and they can use OpenGL; in fact IMO the best 3d PC games use OpenGL (*).
If you don’t talk about 3d Java games, or mean 3d games à la Blockout, it’s another topic, you’re right. We’ve just to decide whether we talk about apples or oranges.
By the way: do you imagine how many Windows patches, IE patches, Anti-Virus definition files, etc the usual Internet user had to download during the last three months to due viri which reached a majority of the “usual” Internet PCs…?
How many megabyte this have been and will continue to be in the future? Since ~50% of the Win-NT4/5 soure is “open” now, viri will get even more clever in future.
(Two of the most “successful” viri corrupted Windows in a way it had to be shutdown every minute, so users had to install these patches or they couldn’t use their PC anymore).
(*) Some graphic card manufactures support “just” the OpenGL stuff ID’s engine(s) need. This can be a problem if your game uses more OpenGL. For example some modern ATI graphic cards tend to show problems when it comes to full OpenGL.
With respect you have no idea what you are talking about Pop along to Dexterity.com’s forums for discussion with people wot are far cleverer and richer than me who really know wot’s wot.
I’m trying to highlight here that Java technology faces competition from other technology and that the solution I’m proposing is a seed technology to grow Java in a space where it can’t currently compete.
It would be interesting if Chris could list the spaces in which Java is currently trying to penetrate and then list the competition in each space and outline quickly what the strategy for Java to compete in each area is.
As I see it I don’t see any Java strategy for competing with C++/SDL and Blitz Basic in the blossoming garage games department. My solution is to release a very, very cutdown subset of the VM, and sell it as an actual product in exactly the same way that Jet sell their compiler, or Blitz sell their IDE. Got any better ideas that will actually address the issues?
Cas
Neither of my machines (P4 3.06GHz, Athlon XP 2500+) will run Alien Flux because the level of OpenGL is inadequate. Both have ATI based video cards as they weren’t purchased for gaming they aren’t the top range models. My wife’s machine (celeron 1.7) won’t run it either although it runs my children’s games perfectly well.
Personally, I think Cas is massively exaggerating the realistic penetration of .NET. I suspect the reason behind this is really not that he’s so concerned about strict penetration percentages, but because he’s concerned that MS has over 2000 developers on .NET, and an equally huge marketing presence, AND have the power to “force” it upon their installed user-base (…because this is what MS always do). In comparison, Sun has no reputation, skill etc in spreading client technologies (I can, in fact, only think of Sun’s failures in this area, but that’s probably my ignorance of their history rather than anything else). So, it’s really important to people making platform decisions to see some evidence from Sun that they are actually fighting this war; so far, the only headline has been a single battle (getting the JVM onto some huge name box-shifters that account for a substantial percentage of the new PC market). But there are many more to come, not least when it comes to “existing PC’s” which vastly outnumber new ones.
I agree that .net won’t ever make it to many windows PC’s, but I know most will be owned by the corporates who won’t allow it, AND who have the power to not deploy it. Usually, home users do not have the option; corporates get “sepcial” versions of windows that they can customize, and effectively remove components that home users have no GUI / option / etc to remove.
So, if we look at the home-user market only, I’m afraid we’re likely to see a much bigger penetration of .net than in general.
Great, how long has this been going on? Assuming you’re counting from around the JavaONE when the first lot were announced, that’s something like hundreds of extra manufacturers by now. Where’s the list?
…and why isn’t this information displayed anywhere prominent on JGO, given how incredibly important it is to a java games developer?..
Sadly it’s another of those “windows-only” features (which, BTW, are starting to worry me - with 1.4.x. and now 1.5.x we’re seeing less and less WORA and more and more “windows gets extra features which the other platforms don’t”. I have seen Java go from non-WORA to very good WORA and now it’s slightly sinking back again 
).
I am largely unconcerned with the linux desktop Java “market” (a term which is rather stretched in definition don’t you think?) but MacOS has continued excellent support for Java and with Sun looking after Windows, the actual desktop market is perfectly catered for in terms of auto-update.
I am not exaggerating .net figures: I am presenting the worst-case scenario for Java. The best-case scenario is that .net is abandoned as a client technology and never makes it to XBox. That is even more unlikely than the other extreme.
But having highlighted the threat from .net, what’s the strategy to deal with it?
And having highlighted the (very reall, 100% actual) threat from C++ and RAD BASIC languages, what is the strategy to deal with these?
I suspect that a clued-up exec will realise there will be a bloody battle fought for hearts and minds with four basic technologies (Java, .net, C++, 4GLs). At this crucial stage mindshare in the games development space is more important than any idealistic view that the JRE be “unfragmented”. We’re not talking about user’s Java experience here; we’re talking about developer’s Java experience. Remember that the solution I propose is entirely forward-compatible with the J2SE-on-every-desktop strategy. This is a developer solution that wins a different battle to the one that J2SE is fighting. It’s training recruits. It’s about building a foundation in a place where this is no foundation or strategy.
I don’t want to sound like I’m the dog’s bollocks or anything but what would the PR blow be if I decided to abandon Java for my next project and switch to Blitz or C++? (I see no value in .net yet) Do you understand that I’m quite well known on the net in general for Java advocacy? What would the effect be for me to say “It’s not going to work out like this.”?
Cas
[quote]With respect you have no idea what you are talking about
[/quote]
What I said has been correct but it depends on which market you and me are talking about, as I said: “if you (…) mean 3d games à la Blockout, it’s another topic, you’re right. We’ve just to decide whether we talk about apples or oranges.”
[quote]Pop along to Dexterity.com’s forums for discussion with people wot are far cleverer and richer than me who really know wot’s wot.
[/quote]
Dexterity offers a lot of games in the style of Blockout. So your concerns on the JRE size is just about casual games? (No need for 3d in most cases.) Then you should say this every time you complain about the JRE’s size.
For larger/complexer 3d Java games, or CD releases, the JRE size just doesn’t matter at all.
[quote]As I see it I don’t see any Java strategy for competing with C++/SDL and Blitz Basic in the blossoming garage games department.[/i]
[/quote]
How could a a high level OO platform like Java compete with Blitzbasic? Why should it? You want Java to be cut down so that it can compete with Blitzbasic? Hm…
I think the suggestions from Chris and others are OK. If you want Java as platform you’ve to stay with its size and lib selection. If you want an Embedded Java, use such one.
I think the JRE size isn’t a problem: the JRE is a platform, not just a game API. Some pointed out that the size increase has to do a lot with international stuff and so on. Which is a good thing - look at the increasing Eastern-European, Asian, etc. market, and so on.
I don’t know what OpenGL level Alienflux needs. I know that ATI’s driver usually support OpenGL in a manner they run ID’s OpenGL 3d-engine OK. I’ve to add however that it looks like their driver improve every monthly version what’s conerning OpenGL.
If your PCs run OpenGL 3d games (one of those many ID-engine based one), JOGL/Xith applications should run too (it depends on which OpenGL version they require but that’s not the problem of the JRE, SUN, or Mr. Bean)
Have you updated your ATI drivers on your PCs? Which ATI model do you use?
Finally I’ve to add that the 3d binding is important but currently unfortunately not part of any JRE, so strictly speaking not SUN’s problem, but ours. Is this off-topic?
Nearly off topic I have decided that it would be far better now that Sun do not try and include a binding in J2SE. Because it will be the wrong binding It will also be doing the opposite of what I want to see happening to the JRE - I want a smaller, modular JRE, not an ever-larger monolithic Swiss army knife + kitchen sink.
Back on topic: I am exactly saying that Java is competing for mindshare in companies like Dexterity to produce “casual” games. But I am not talking about just casual cames - I’m talking about indie games, which range from the very simplistic Dweep Gold and Fitznik right the way up to the fantastically presented Hamsterball or Starscape Moonpod. There is nothing “casual” about Hamsterball or Moonpod. The only thing all the various indie games have in common currently is that they are of downloadable size and that they are extremely simple to install and run.
Download size kills this market. Complexity kills this market. Even reliance on OpenGL drivers is detrimental although not nearly as bad as some indies think (a lot of them are still deliberately working to DX5 level to capture more customers).
I know full well that Sun has sights only on the big studios like EAGames and Take2 and Sony. Otherwise I’d see a product roadmap outlining how Java was going to help the indie market sector. Currently there is no place for Java in this massively expanding industry sector except if you buy Jet. Sun are simply ignoring a marketing opportunity. I can understand Sun being blind to it because they have no idea about this market but I am expecting 100% support and commitment from the GTG…
Cas
Cas…if a lot are working to DX5 level then why not just use Java 1.1? It ships with EVERY PC that had a browser! I mean, come on. Dweep could not have been done in 1.1? For goodness sake, check out Command and Conquer: Attack Copter from EA http://www.eagames.com/free/home.jsp?ncc=1. If this could be done in 1.1, surely Dweep Gold or BlastORama could have. It was purely a technology comfort decision on the part of the developer to not use Java.
In fact, if the developers HAD chosen to develop these games using Java 1.1 they would have had MORE market penetration (Mac and Linux).
Not true. Yes we are focusing time there but if that’s all we cared about, why would we spend time here? Why would we work to build Java.Net? I think that is a bit unfair
[quote]Otherwise I’d see a product roadmap outlining how Java was going to help the indie market sector. Currently there is no place for Java in this massively expanding industry sector except if you buy Jet. Sun are simply ignoring a marketing opportunity. I can understand Sun being blind to it because they have no idea about this market but I am expecting 100% support and commitment from the GTG…
[/quote]
Cas, not everything is available for public consumption. Why would I put up information that kills possible market advantages for us and let’s competitors know EVERYTHING we are up to? Do you really think that everyone that hits these very public forums are Sun’s friends? There is a big difference between a /. type of forum where everyone is anon. and has no accountability. Should I spill everything about GDC right here? No suprises, no “event”? Sorry. Can’t do that. I am accountable to a publically traded company.
Companies like Sun disclose deep roadmaps all of the time. It’s done under an NDA and it’s not given to everyone, true. I don’t think that we have been anything less than 100% supportive of this community since the beginning. Infact, this site came into being with us doing what WE (the GTG) thought was right, not what Sun thought we should be doing.
Also, it’s a bit insulting to say that we are ignoring a market opportunity when we in the GTG have tied our livelyhood and personal responsibilities to driving these technologies into the games marketplace. Sun had given us the mandate to go after the market and we are doing everything we can to hit all fronts.
Again Cas, the real issue I sense here is your frustration as you believe that the way Java is required to be packaged is not suited entirely to your needs. (Look at the title of your thread). I don’t know what else to tell you buddy! We are working to make Java technology better for game development, indies and the big guys alike, and while the solutions may not meet everyone’s expectations 100% of the time across the board, you have to admit that Java game developers are a lot better off now then they were 3 years ago due, in no small part, to the efforts of the GTG.
-ChrisM