“Sun’s Java is locked on the future in this new release
With this significant upgrade of the Java standard, Sun directly takes on Microsoft’s XP and .NET framework. Whether you’re a developer or just a Java enthusiast, you’ll need to know what Sun has in store for 2002.” Read the interesting preview of the new Java version at ExtremeTech.
Well I think it’s more appropriate to call this Extreme Tech article a preview of SWING in JDK 1.4 than a preview of “the new Java version”.
Wow. Great. Like anyone cares about new widgets. And the Java Webstart is much needed for the bloat of a half dozen different JVM’s sitting on your client PC.
Nevermind the “Another feature that has been added in the JDK 1.4 release is the ability to cut-and-paste between most Swing controls using a clipboard.” Oy Vey! (rolls eyes)
And this is supposed to compete with .NET *how* ?
They are locked on the twisting road with no breaks, is my take.
Note I am not a .NET fan.
Anonymous:
Java 2 Standard Edition 1.4, Part I
Kon, you start talking about Swing, and then compare it with Microsplodge’s DOTNOT — and one has absolutely, categorically nothing to do with the other (not even in related fields). You might not be a DOTNOT fan, but at least get your facts -remotely- straight before posting a load of codswallop.
Back when I was even greener behind the ears than I am now, I did the Java thing (got Sun-certified too). The novelty has since wore off. I found out that Sun didn’t/doesn’t really care about the client side (yes, I know, “big surprise”).
–Where is JavaOS?
–Where are JavaChips?
–Ok, how about Sun-backed ahead-of-time Java compilers? (Yeah, I know about gcc, but where’s Sun’s support there? Where’s Swing there?)
–Why is Sun using Gnome for Solaris 9 instead of some sort of “OpenJavaDesktop” of sorts? “All metal, all the time” would rock.
Early on, I’d hoped that Java was the smaller, tighter language struggling to emerge out of c++ (that Bjarne has referred to). I think maybe that the language I was looking for is just more disciplined use of a well-chosen subset of c++.
Anonymous douchebag – maybe you should read the news item text.
Thanks for repeating *exactly* what I said. Thank you drive through.
> I found out that Sun didn’t/doesn’t really care about the client
> side (yes, I know, “big surprise”).
Not true. Java is already great on the server side and Sun is now
focusing more and more on the client side.
> –Where is JavaOS?
Scrapped. There is no reason to build an OS using Java, you would
still need lots of native code in the lower layers. So, it’s easier to
take a Linux kernel and put Java on top of that.
> –Where are JavaChips?
There are many Java chips for embedded devices. For normal desktop or
server computers there is no reason to create special Java chips. They
will not be faster than modern JIT’s.
> –Ok, how about Sun-backed ahead-of-time Java compilers? (Yeah, I
> know about gcc, but where’s Sun’s support there? Where’s Swing there?)
Sun doesn’t support ahead-of-time compilation. There’s no reason to do
it. The Java applications will not run faster. Actually the IBM JVM
generates much faster code than GCJ, see
http://www-106.ibm.com/developerworks/java/library/j-native.html?lo…
for a recent comparison.
> Sun doesn’t support ahead-of-time compilation. There’s no reason to do
> it. The Java applications will not run faster.
Java programs run plenty fast for me. The problem is, they take up gobs of memory (maybe around 10 MB or so,.. I forget, it’s been a while since I had a jvm around) and they take way too long to start up.
Not to mention the fact that if you start up 2 java apps, you’ve got 2 full jvm’s running.
There’s also the fact that you need a comparatively large and complex jvm to run Java programs — something that many small OS projects can’t manage. With Java’s relatively simple syntax, it shouldn’t be that hard for Sun to create a “reference Java (x86 or sparc assembly) compiler”.
I really don’t know why they are clinging to WORA. WOCA would be perfectly acceptable (especially if Sun provided a simple “click here to make this app ready to run on this platform” to handle the compilation for inexperienced users).
Anyway, thanks Phazer — I’m not surprised if IBM’s jvm is faster than what GCJ produces; I’m sure IBM can afford to make it very fast.
A follow up to the benchmark I linked in my previous post; the
benchmark program used “long” instead of “int” and GCJ isn’t very well
optimized for this. If you use “int” instead GCJ is as fast as the IBM
JVM. More info here:
http://gcc.gnu.org/ml/java/2002-01/msg00268.html
> Java programs run plenty fast for me. The problem is, they take up
> gobs of memory (maybe around 10 MB or so,.. I forget, it’s been a
> while since I had a jvm around) and they take way too long to start
> up.
>
> Not to mention the fact that if you start up 2 java apps, you’ve got
> 2 full jvm’s running.
The start up time and memory consumption is a problem. One solution,
which Sun is working on for JDK 1.5, is to load only one JVM in the
system and that all Java programs share classes and other
resources. This would probably solve both problems.