Yeh, you’re wrong 
J2GE involves a couple of specific tuning tweaks to the server VM (actually, to make it a hybrid of the client and server VM to prevent that awful chugging on game load), and cut down the packages to a tiny few classes that do the job.
Cas
So where am I wrong exactly?
Not you! Gergis.
Cas
Cool. Thanks for the correction. Now, how does one go about doing that? evil, evil grin
Cas -
I’m not really sure that the market share question really belongs in the ‘Performance Tuning’ section of the forum, but I do have one comment on your logic.
You base a basic a general assumption on making your product available globally as opposed to a North American and European crowd. This leads to more assumptions on system requirements, download size requirements, payment methods (a side note - the phone bill is another great medium as seen in Korea). This should also open questions about the effectiveness of advertising and issues about what is reasonable pricing in a ‘local’ sense.
So how does your equation change by going after a higher end market? Of course there are no restrictions exactly, anyone with enough bandwidth to adequately download the game and the CPU horsepower can play the game. But what if it is targeted at the mid point of the typical North American and European user?
Working against those same statistics provided on internet population (http://www.nua.ie/surveys/how_many_online/), 62% of Internet users are in North America and Europe. Lets work on the assumption that a mere 60% of the consumers meet the hardware requirement.
On all other points - these percentages might fluctuate UPWARDS based on the demographics (better ability to pay, more likely to download big packages, etc), but lets leave these the same for the sake of discussion.
Using your same calculations with the appropriate modifications the net results show that you’ve more than doubled the number of sales! Of course that doesn’t even account for the handful of sales outside of the market area.
Of course all of this is based on broad assumptions and working aganst an equation that no one in the world can make accurate. Then of course I always like that lovely quote from Mark Twain – “There are three kinds of lies: lies, damned lies, and statistics.”
We should move it to general business discussions really.
Although the whole equation is truly a simplified one and doesn’t take into account all the millions of interactions, it’s one that you can actually work with to get results.
But if you restrict your market to the 62% of the Internet who are European or North American in order to use higher spec machines, what you will find is that the percentage of users who can use your software does not increase correspondingly. It’s easier to explain in words than maths: if you split the market into two, the 62% who can use your software and the 38% who can’t, the total number of people who buy your software cannot increase, because there are not suddenly more people in Europe who want to buy the game: all the other factors apply to both markets.
The math looks like this: (62% x 60% x original value) + (38% x 0% x original value). In other words, you are certain to lose sales unless 100% of European users can use your software, which will maintain your sales at the original level. There is one exception which is that the extra features you have been able to provide as a result of increasing the base specification make the software more desirable, that is, you increase one of the other factors. Maybe bumpmapping gets me 10% more sales from the eye-candy factor - not enough to counteract the 38% I’ve lost by cutting out the rest of the world.
Cas
Holy cow! As soon as I saw Herkules’ reply about there being legal issues, I realized he was 100% right and walked away from the topic. I never realized my little question would balloon into this kind of debate!
Well, before this topic gets moved off to the general business forum, perhaps I can try and yank the reigns back to the original topic. So whether or not it’s morally correct to just swallow the JRE size bloat, I do think it’s a problem worth addressing. The obvious answer is to switch to c++ and make native binary builds to keep the distribution size down. But I’m not willing to give up on Java just yet, as I honestly believe it really does cut down development time.
So, given that (A) I want to develop in Java and (B) I want my distributable to be as small as possible, I’m curious to know if anyone has tried any java to native compilers? I know they exist. I believe CodeWarrior can compile my java to a windows native exe. (And possibly a mac binary as well?) What about Linux? Is there a java to native compiler there?
Yes, me. Behold Alien Flux, which is compiled by Excelsior JET 3.0 Professional. Note that it is under 5MB. The .JAR of the same is over 4MB. Food for thought.
Not only is it as fast as or faster than the Java Hotspot Server VM, it starts up in about 1/10th the time, and runs at full speed. And all for about a half meg overhead on the jar. That’s why I use it, no other reason really.
Cost me £500, and worth every penny.
There’s a Linux86 version of Jet in the works I believe.
Cas
Thanks Cas, that’s an excellent resource. I downloaded the “Personal” version and was quite happy with the result. I’m also watching the thread on GCJ now to see what becomes of that, but I’d hate to have to switch all my graphics rendering from awt-based to swt-based.
I don’t know why there’s such fear of SWT. Only some kind of bizarre luddite would’t take one look at it and realise it was technically and aesthetically superior to AWT and Swing combined any day…
…I don’t even use SWT meself, of course…
Cas
Oh, I don’t fear SWT. I just fear porting a crap load of code already written using AWT/Graphics2D. I’m definitely willing to spend a couple hundred bucks on a product that will save me the hassle. And honestly, using the Jet gui was a hell of a lot easier than trudging through websites and docs about gcj.
That said, I might consider gcj and SWT for future products. I’d like to see some more examples of people going this route first though. (Read: I’d rather someone else do the hard work for me.)
Check out the size of the Windows 1.4.2beta JRE: 1.3MB
[quote]Well, we live in a world where Windows went from 98 to 2000, Solaris 2.8 is marketed as “Solaris 8” and reports 5.8 on a uname, SimCity went from 3000 to 4, and MPEG just missed out version 3 entirely…
I’m expecting Sun to start using complex numbers next…
Oh, hang on… Oracle 8.i… :o ;D
[/quote]
ROFL…
[quote]A.F. works on the majority of systems OEM drivers, but that only means 50% or more
Deploying Games On The Internet Essay
This essay has been written before on the 'net by shareware professionals (which I’m not, yet), and probably better than I’ve done it, but the gist of it should remain the same.
[/quote]
Ah, but Cas, you miss one critically important point…
I’m relying on everyone getting a JRE to play someone else’s game, and then I can distribute smaller downloads than even you can with JET (you quote the JAR as being as much as 4Mb compared to JET’s 5MB - I choose to see that as “wow; a 1Mb saving” ;)), and rest assured that most people already have the JRE.
Which is why I’ll keep telling people that, since AF was written in Java, it “works better and you can get higher scores - but don’t tell anyone” if you play it on a JRE ;). LOL…
One of the advantages of the glacially slow pace of java releases (1.1, 1.2, 1.3 and 1.4) is that client penetration (oo-er!) is cumulative (OO-ER!), perhaps 75% so. So, we just need a few great java games to get those JRE’s deployed, and we’re away. When asking friends and family to alpha test my game, I find a lot of them are running windows and already have a JRE that can run quite a lot of stuff (assuming they grabbed the VM updates before MS was forced to stop distributing them to people who didn’t have the latest VM already). As long as my game runs at all, they happily go and get a real jre.
I suspect this is because of the aforementioned video card drivers situation. If you are an NV user, then you gave up complaining about large and frequent and ESSENTIAL downloads years ago - every game-playing NV owner knows that you get those detonator updates sooner than ASAP, ESPECIALLY if you have any game problems (and you keep the old downloads, because you soon get used to having to downgrade to get half your apps to start working again ;)).
If you sell SW on your website, then make another option for the buyer to get the SW with JVM on CD, but ask $10 more for it. Everybody will be happy that they can save the $10 and giggle the whole 30 minutes they need to download the JVM with their modem.
For those interested, I will be offering the .JAR version alongside the .EXE version, and I’ll be featuring a Get Java button on the website. The JAR version is of course the only one that runs on Linux, too 
Same goes for the full game.
Cas
[quote]Check out the size of the Windows 1.4.2beta JRE: 1.3MB
[/quote]
Ya, I was checking that out !! and couldn’t believe my eyes !!!
How does that work?? Did they cut out anything ??
On a side note, it might help their cause if SUN
made the JRE download process a little easier…
I mean, for a amature game designer like myself,
with not enough resources for real web hosting.
I can’t serve the JRE along with my Game on
free web folder.
And to make matters worse,
I can’t even put up a like that will directly download
a JRE from sun’s site for people to click on…
Some people will just get turned away by the.
Fill out your info -> Read all this legal stuff ->
Agree on something you didn’t read.
Why can’t sun make a 1 click download for the JRE ?
Anyways, just my 2cents.
They didn’t cut too much. Just the whole JRE. 
That download is the “online installer” download. When you run the installer, it contacts Sun’s site and downloads the rest.
HAHA Okay… That explains a whole lot…
Silly me. .
:o :o :o
1.4.2 is now 14.1 MBs …
So much for cutting down the size …
At least 1.4.2 was a lot quicker to download and install than the Windows 2000 updates I installed this week (which also required accepting licence agreements).