[MUD-Dev] java clients

ceo ceo at grexengine.com
Wed Nov 12 10:24:14 CET 2003

Christopher Kohnert wrote:
> On Nov 8, 2003, at 6:55 AM, ceo wrote:

>> If you want to develop and deliver java clients, the story is
>> simple - tell your uses to get the current version of java!
>> Frankly, I find the widespread obsession with 1.1 compliance
>> rather strange - it doesn't even have basic data structures like
>> List!

>> I'm impressed at people's masochism in choosing to code in such a
>> barren landscape :P...

> Hey, we work with what we've got. Java client developers would
> love nothing more than to have a stable, working, featureful
> version of Java on every user's machine. The reality is that most
> users don't, and we've got to code to the lowest common
> denominator in order to keep the barrier of entry low (in a medium
> that is already pretty high). Even WebStart, which is what the
> initial java release should have been, is annoyingly difficult to
> get users to use.

Right, lowest common denominator == no java installed. I'm not
kidding, I'm serious - in percentage terms, it's true. MS's JVM is
soon to disappear ENTIRELY (unless there's a change of heart in the
legal wranglings), and it was *never* compatible with 1.1 java
anyway, at least not until February/March this year. Anyone who
didn't download it before the end of Spring is now not ALLOWED to
download that update (due to court decisions).

OTOH, if you look at only people who already have a 1.1-compliant
version I would be surprised if 1.2 isn't the most widely-installed
minimum version. It is now 5 years old, which is a long time for it
to have penetrated via upgrades. Note also that nearly all java apps
require 1.2, so the chances of someone having already been forced to
upgrade by now are reasonably high.

As a separate issue, JWS is often difficult to get people to use
because it often DOES NOT WORK under windows! This is a quite a
major problem. I don't know yet what the conditions are that cause
it not to work, but this affects a large number of windows users.

>> Mostly, though, developers tell themselves it's impossible when
>> it's merely slightly undesirable. Downloads of technologies over
>> the past decade have proven time and again that if your content
>> is compelling, people will download. e.g. dialup users still
>> download 30-50Mb regularly if they have an nVidia grahpics
>> card...).

> If users wanted new and working Java, this wouldn't be an
> issue. ;) The fact is, they don't care enough about it to download
> and install it (for the most part), and convincing them otherwise
> is pretty damn difficult.

How many people actually *want* new nVidia drivers? Certainly the
hardcore gamers do, but most people couldn't care less except that
their games aren't working - or in some cases their business apps /
desktop keep crashing, or are going very slow.

AFAICS, it's all about expectations. Today, in 2003, Joe Bloggs
home-user is accustomed to having to download a bevy of upgrades to
his system every now and then to use an app / play a game / surf a
website. Flash flourished even in the days when this expectation did
not exist, and has by now set a precedent for how we surf the web -
we expect periodically to have to upgrade aspects of our
system. Those of us using Windows are accustomed to > 10Mb downloads
every few months - and many of us get the huge multi-meg-monster
that is a new point-release of DirectX every year or so too. All
this is in addition to the graphics drivers etc. Windows Update just
makes the process even less exceptional and more "standard".

If you have some evidence to the contrary of this:

   >> Downloads of technologies over the past decade have proven
   >> time and again that if your content is compelling, people will
   >> download.

then please give it to me. I'll happily raise hell with Sun over the
issue via their official Game Technology Group. Some game developers
have been trying to persuade Sun to move the JRE into a
fully-modular system, where you download just the JVM, then
libraries are added "on demand". This would make the basic Java
install very small, probably less than a megabyte, but at the cost
of weakening the "all pervasiveness" of the stanardard
libraries. Data to show the download size is seriously harming
uptake would be a fantastic weapon to make their case with, yet so
far no-one has been able to produce any. Which makes an (albeit
weak) statement in itself.

Adam M
MUD-Dev mailing list
MUD-Dev at kanga.nu

More information about the mud-dev-archive mailing list