JRE Download Size

[kevglass]

I also read up a bit in this thread. Cas made a good point above that webstart is written in Java. When I typed the above I thought someone most have already brought the point up. Why isn't there a webstart plugin for each platform written in their native tongue. That'd make life much easier for the end user wouldn't it?

I'd have lost 90% of my sales if I delivered via Webstart at this point in time (conversion rate between Win32 and Webstart versions is similar).

that doesn't actually follow does it? If you hadn't offered a Windows version nothing to say that the people who used the Win32 version wouldn't have used the webstart one instead.

Kev

[princec]

I've got a statistic for everything, and while I am now unable to measure what happens if I remove my Win32 demo (as my affiliates are spreading that about too) I can tell you that of the 35% of users who were asked to download OpenGL drivers to make the game work, only 1% actually bother.

Besides, guess what happens when you click on a JNLP link when you don't have Webstart installed? Yes, that's right, it downloads an XML file which does nothing. So then you start having to stick platform specific code on web pages to try and get the Java plug-in. Oh no! That's precisely the opposite of the WORA dream!

Lest we all forget, Flash is the de facto format for interactive web content now, and the reason is that Flash is distributed along with the operating system. (The fact that it's better at doing what it does helps of course but no-one would have cared if it weren't ubiquitous)

So: faced with the cold hard stats, could anyone possibly say with a straight face that Webstart is currently a good deployment mechanism? And given that you'd lose a lot of credibility by trying to argue otherwise, what would the collective's easy quick-fix interim solution consist of?

Cas

[kevglass]

I've got a statistic for everything, and while I am now unable to measure what happens if I remove my Win32 demo (as my affiliates are spreading that about too) I can tell you that of the 35% of users who were asked to download OpenGL drivers to make the game work, only 1% actually bother.

Seems thats a more scary problem, that people can't be bother to download GL drivers.

So: faced with the cold hard stats, could anyone possibly say with a straight face that Webstart is currently a good deployment mechanism? And given that you'd lose a lot of credibility by trying to argue otherwise, what would the collective's easy quick-fix interim solution consist of?

I still don't quite go with your stats that webstart is bad because more people downloaded your win32 distribution, but that aside your logic is sound, clicking on a JNLP link and being faced with nothing but a XML page is utterly useless.

As above, the solution would seem to be to have a native plugin which orchestrates the download itself (via JNLP or what ever other method you'd like to reinvent). Better still would to have the plugin available on every platform before you get there (e.g. ship it with the machine) but that doesn't seem possible since most machines are already shipped . I suppose it would be possible to write a deployment app in Flash if you really wanted to?

At the end of the day tho, the only thing that seems to make _your_ end customers happy is to have a tiny download. If thats the case then a binary distribution of just the bits you need seems to be the only thing you can do.

Cas, out of interest, why did you consider webstart knowing that it would require you users to get a JRE? You already has a JET compiled small binary didn't you? EDIT: Oh, I only every played the webstart version, was there an installer with the Win32 version?

Kev

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

[princec]

We believe very strongly in Webstart's idea (but not the current implementation). And we also believe in getting the game out to Linux (and one day Mac...) users, and Webstart does all this for us.

The download doesn't have to be tiny, it just has to reflect the amount of game you're getting. If I downloaded 24MB of stuff and got Alien Flux I'd be annoyed because it doesn't seem to have 24MB of content. At the end of the day though it's simply not our responsibility to get the JRE into the hands of users - we just don't have the resources to make it happen - and we just want a temporary interim solution.

A one word answer to a question I've asked would do the trick.

Cas

[Preston]

Seems thats a more scary problem, that people can't be bother to download GL drivers.

Again it depends on the game genres we talk about.

Those customers who want to play low-end MAME type games don't need a high end graphics card, that's right...
Those customers who want to play high-end 3d games (no matter what budget it is) already have got a new OpenGL 1.4 driver installed (or 1.3 in case of Ogl-unfriendly ATI).

Coming from the full price games market I still fail to see why the low-end PC market would be the main target for independent developers? The 1++ GHz / GF2++ PC market is full with game customers who look for good new games (and in contrast to casual gamers they're ready to pay for). Also I know so many of them who are really sick of those mega complex games which take you two weeks to get just started. Give' em high-end low-budget games. :-)

IMHO independent developers don't need to restrict themselves to low-end games.

[swpalmer]

Besides, guess what happens when you click on a JNLP link when you don't have Webstart installed? Yes, that's right, it downloads an XML file which does nothing. So then you start having to stick platform specific code on web pages to try and get the Java plug-in. Oh no! That's precisely the opposite of the WORA dream!

The platform specific code you are taling about IS irriating - only because broswer standards are so bad... but if it works... If your stats say that your market is 98% Windows users I think this code  is worth it.  It's not like using it will lock out everyone else.  So long as Linux and Mac users can still get at it.
Linux = more experienced in general (my assumption) So if you tell them on the page, "If it doesn't work go here and install the JRE" they are morelikely to deal with it.
Mac = less experienced in general (another assumption) But of course they have Web Start already, so it's no big deal.  Except if they only have Java 1.3.1.. I'm not sure that Apple provides a proper URL and mechansim for Web Start to do it's thing and download the upgraded JRE.

[Jeff]

ISo: faced with the cold hard stats, could anyone possibly say with a straight face that Webstart is currently a good deployment mechanism? And given that you'd lose a lot of credibility by trying to argue otherwise, what would the collective's easy quick-fix interim solution consist of?

Cas

Sorry Cas.  You have some numbers. What you DON'T have in an incontrovertible conclusion.  There is a world of distance from statistics to their interpretation and interpretation is *very* hard and *very* error prone.

As i posted elsewhere I can see at *least* 2 confounding factors in your study.  The first is that your Win32 download is different from the Web Start one in that its an exe.  We have no way of measuring how much "exe prejudice" is effecting your numbers.

Secondly, as you have already commented, you have the Win32 download directly linked from other sites but I assume you don't have the webstart linked in the same way from the same sites?  if so, the entire difference could simply be a popularity difference between your own site and one of your affiliates.

Thats just two of many things that could be influencing what is fundamentally a highly uncontrolled study.  In the end you must be *very* careful when interpreting the results of any study to avoid the fallacy of confusing correlation with causality (as you seem to be here.)  Its a mistake too many people make too often.

[kevglass]

This is why I prefer logic to stats, however, logically speaking, Cas still makes a very good point. JNLP or rather webstart being in Java means you have to have java before it can go and get Java for you.

Are there any plans for a native web start implementation?

I actually started considering a Java 1.1.x version of webstart that could run as an applet (since everyone gets an old VM built in anyway). You'd have to do some signing but I think it works on the later 1.1.6 stuff. This way you go to the applet, it runs webstart and you're off...

Kev

[Jeff]

This is why I prefer logic to stats, however, logically speaking, Cas still makes a very good point. JNLP or rather webstart being in Java means you have to have java before it can go and get Java for you.

Are there any plans for a native web start implementation?

Not from Sun.  You could start a project though to do an OpenSource native JNLP client.  All the info is available.

JWS, good or bad, serves the clientele it was created for well (enterprise IT groups.)  Java is the easiest platform because it handles all the security and HTML stuff that otherwise you are going to have to write yourself :/

The issue with doing a 1.1 (MSVM) version is only that by the time you have it done MSVM may well be history.

[kevglass]

It might be history legally, but it'll still be resident on loads of users home machines won't it?

Kev

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

[princec]

Re: Linux users being more experienced - turns out that the failure rate on Linux is about 50%. More than that I can't deduce though.

Re: exe prejudice - well, then there's exe prejudice. It can't be ignored but it can be factored in. The truth is, if the consumer was aware of what Webstart and Java offered they'd never touch another exe download (we're going to start highlighting this on our site soon to encourage Webstart usage).

Re: that bit of code that does the detection - we did actually use that for a month on our site but it kept ballsing up or misbehaving at a time when we were getting 20,000 visitors a day. In the end it created so many support calls we got rid of it. Nowadays I dare say we could use it again because our traffic's a lot more reasonable!

Re: affiliate stats skewing - fortunately I can filter out all my affiliates from the results and look purely at the ones from my own download page. And the results look like: 28% of installs are Webstarted from Puppy Games, and 72% are Win32 exes from Puppy Games. Still 3 times as likely.

Cas