Has Sun Given Up on the Desktop?

[Ovidiu Platon]

"MarkN" <m@n.com> wrote in message news:3d8871fa$1@10.1.10.29...
>
> The offical release is that old(young). But .Net really is much older.
It
> had Java to learn from. How long has it really been in development? We
> won't learn the truth but I suspect right about the time MS was trying to
> fork Java.
>
> And it doesn't do all that Java does. It does have some features Java
doesn't.
> Not anything Java isn't going to have or doesn't need. .Net may do
everything
> you 'need' to do and so seems to be able to everything Java does. For at
> least one - it doesn't run on z/OS.
>
>
Well, then, keep praising Java (actually, I like it too, it's just that
there's too much unused potential in there...) and let others live too...
After all, if .NET isn't a Java killer, it isn't that good and everything,
you have nothing to worry about, right? And, to get back to the initial
issue: I don't like Swing and I think they'll need another re-work in order
to make it work decently...

[MarkN]

>Well, then, keep praising Java
.Net is a good tool too.

> (actually, I like it too, it's just that
>there's too much unused potential in there...)
Learn to use it.

>and let others live too...
I do.

>After all, if .NET isn't a Java killer, it isn't that good and everything,
>you have nothing to worry about, right?
Doesn't matter. I have nothing to worry about. I can (and do) .Net. It
is simple anyway - right? That is how MS tells it. Being good or best or
right has nothing to do with what survives.

>And, to get back to the initial
>issue: I don't like Swing and I think they'll need another re-work in order
>to make it work decently...

You have the right to think that. Doesn't make it reality. There is some
truth to it though. But everything can be improved.

[Brad O'Hearne]

Hello Ovidiu, BradO here. A couple things -- I think we need to keep proper
perspective here....comments below:

"Ovidiu Platon" <ovidiupl@microsoft-lab.pub.ro> wrote in message
news:3d86f95d@10.1.10.29...
> I tried in 2000 the Forte 4 Java tool produced by Sun and it was crappy at
> most... Garbage collections that halted the app every 10 minutes on a
system
> with 256 Megs of RAM (that's 2 and a half years ago),

Windows itself doesn't perform all that terrific on 2 year old machines
(which would have put top cpu's @ AMD Athlon 750 or Intel P3 7xx) machines
with 256MB RAM. XP itself has lags in the simple start menu display if you
aren't running with the latest hardware and large amounts of memory.

>slow response times
> (does anyone remember the rule of 1/10 of a second?).

No...Windows itself doesn't meet this standard, and I haven't seen too many
web sites lately that load in that amount of time (of course, due to network
latency, but it's all moot to an end-user).

>Last autumn I tried
> Forte v3 and I was amazed: even scrolling was much improved; the secret?
> Some of the code (the performance critical parts, I guess) seem to have
been
> converted from Java to C (I'm talking about some internal classes that
have
> vanished from the source code of Swing from one version to the next one;
the
> same thing seems to have happened with Forte). Where's the performance of
> the JVM? Where's the portability? Can Sun handle the task of converting
(and
> maintaining) the Java code to platform specific C code for every
bottleneck
> in it's APIs?

Keep in mind that the entire JDK is built on native code. This is why we
have to download a platform-specific version of it and the JVM. Converting
portions of Swing code better served as native code doesn't change
portability at all -- and portability doesn't somehow come for free -- when
Sun puts it in the jdk, it becomes portable, because it has to be
retrofitted to each platform's version of the JDK. I don't see what there
is to complain about here, as the is Sun simply improving the JDK for the
community. This is a good thing.

> I wish we had a truly component-oriented GUI framework that was well
> designed, performed well on any platform and let me focus on my
application,
> not on common problems faced by every GUI developer. I think it's not
going
> to be Swing, because the JVM as it is today, has yet to be improved when
it
> comes to performance issues

Let's talk specifics. I would argue that most true Swing performance issues
are design problems.

> (Swing it's reasonably well designed and Java
> does work on many platforms); it might be .NET, because it has reasonably
> good performance (the CLR performs way better than the JVM from many
points
> of view),

Again, let's talk specifics.

> .NET Framework is quite well designed although there's room for
> improvement and it has the potential of becoming platform independent

Your complaints about a vendor handling the "task of converting (and
maintaining)" code for every platform would apply then to .NET. The CLR
isn't magic, there's native code required there too.

> (although MS will never support this).
> As a developer, I try to keep pace with both of these worlds and do
whatever
> my customers want me to do. At this moment, when it comes to desktop apps,
I
> do reccomend .NET Framework since it outperforms Swing

Again, specifics.

> and, let's face it,
> most of the people have Windows on their desktop.

Java runs on Windows....

Cheers,

BradO

--
Brad O'Hearne
brado@neurofire.com
DevX Section Leader

[MarkN]

>The system does experience some
>disk thrashing when I load 3 or 4 large apps and switch among them, but
other
>than that, I find the performance quite acceptable.

See, that wouldn't work for me. I usually have at least that many open.


The real point is that new software usually requires more resources. With
your setup Java apps will run just fine.

[Phil Weber]

> Windows itself doesn't perform all that terrific on 2 year old machines
> (which would have put top cpu's @ AMD Athlon 750 or Intel P3 7xx)
> machines with 256MB RAM. XP itself has lags in the simple start menu
> display if you aren't running with the latest hardware and large amounts
> of memory.

Brad: That's not been my experience: I'm running XP Pro on a 4-year-old 333 MHz
Celeron (overclocked to 416 MHz) with 256MB RAM. The system does experience some
disk thrashing when I load 3 or 4 large apps and switch among them, but other
than that, I find the performance quite acceptable.
--
Phil Weber

[Brad O'Hearne]

Hey Phil,

"Phil Weber" <pweber@nospam.fawcette.com> wrote in message
news:3d8896ea$1@10.1.10.29...
> Brad: That's not been my experience: I'm running XP Pro on a 4-year-old
333 MHz
> Celeron (overclocked to 416 MHz) with 256MB RAM. The system does
experience some
> disk thrashing when I load 3 or 4 large apps and switch among them, but
other
> than that, I find the performance quite acceptable.

I am running on nearly the same specs as you (and same age of computer -- my
notebook that is). I didn't say the performance wasn't acceptable. I am
more or less fine (most of the time) with it. But what I am saying is that
response time is not the .10 second standard our friend in this post was
suggesting. And that underscores one of the main ponts I was making in the
article -- the real gauge for performance is not nanoseconds or
milliseconds -- it is usability. And as discussed, we even tolerate waits
with Windows. The key is the ability to adequately use a system without
intolerable or productivity-defeating delays. On that scale, I see Swing
measuring up fine in the performance category. "Good" GUI performance is
more related to human sensory perception and speed of human input device
usage than time, or cup cycles consumed.

Cheers,

BradO

--
Brad O'Hearne
DevX Section Leader

[Ovidiu Platon]

See my comments inline
"Brad O'Hearne" <brado@neurofire.com> wrote in message
news:3d888b50$1@10.1.10.29...

> >slow response times
> > (does anyone remember the rule of 1/10 of a second?).
>
> No...Windows itself doesn't meet this standard, and I haven't seen too
many
> web sites lately that load in that amount of time (of course, due to
network
> latency, but it's all moot to an end-user).

True, Windows doesn't meet the "standard" - for Windows compliance, MS
reccomends 400ms response times (and response time usually means some clue
that the app has received the command and does something about it). And the
"rule" is meant for desktop applications, not for web sites - it's
impossible to offer estimates of the response time given the structure of
the net.

> Keep in mind that the entire JDK is built on native code. This is why we
> have to download a platform-specific version of it and the JVM.
Converting
> portions of Swing code better served as native code doesn't change
> portability at all -- and portability doesn't somehow come for free --
when
> Sun puts it in the jdk, it becomes portable, because it has to be
> retrofitted to each platform's version of the JDK. I don't see what there
> is to complain about here, as the is Sun simply improving the JDK for the
> community. This is a good thing.

As a developer, am I supposed to provide the long-praised bytecodes that run
the same way on any platform or should I provide some code that works and
later convert it to platform-specific code for performance? The HotSpot (or
JIT, or however) technology from Sun hasn't convinced me so far... The JIT
technology from MS did.

>
> Let's talk specifics. I would argue that most true Swing performance
issues
> are design problems.
>

You're right, but I think there are also problems with the concepts behind
the JVM that make this worse.

> > (Swing it's reasonably well designed and Java
> > does work on many platforms); it might be .NET, because it has
reasonably
> > good performance (the CLR performs way better than the JVM from many
> points
> > of view),
>
> Again, let's talk specifics.

Try writing a simple MDI Notepad app or something simple like that and
compare the memory footprint and the execution times. Good luck! Now think
how a large app would look like on your system... And something else: I do
prefer delegates and events instead of inner/anonymous classes; when
compiled, the actual code is pretty much the same, but the source looks much
better.

>
> > and, let's face it,
> > most of the people have Windows on their desktop.
>
> Java runs on Windows....

How many people actually run Swing apps on their desktop? True, you could
ask the same thing about .NET, but, then again, how old is Swing compared to
..NET?

Best regards,
Ovidiu.

[Ovidiu Platon]

> Doesn't matter. I have nothing to worry about. I can (and do) .Net. It
> is simple anyway - right? That is how MS tells it. Being good or best or
> right has nothing to do with what survives.
>
Oh, it does. The survival of the fittest, remember?

[MarkN]

"Ovidiu Platon" <ovidiupl@microsoft-lab.pub.ro> wrote:
>> Doesn't matter. I have nothing to worry about. I can (and do) .Net.
It
>> is simple anyway - right? That is how MS tells it. Being good or best
or
>> right has nothing to do with what survives.
>>

>Oh, it does.
Don't see how. I can do either one. With well architected code it would
be easy to convert to .Net or to Java. I won't have much of problem.

>The survival of the fittest, remember?
No I don't. That is not true (If you believe in Evolution then it is for
Evolution - but this isn't). How many examples do you want?

[Rob Abbe]

.Net has been available long before it was released. Even books were available.
You should be disappointed as should all .Netters that after being available
for so long, that it has not taken off.

Because .Net is just an off brand version of Java, simply by learning .Net
you're already taking great strides in learning Java.

Congratulations and welcome to Java!

"Ovidiu Platon" <ovidiupl@microsoft-lab.pub.ro> wrote:
> You're right... or not. Think about it this way: .NET is 7 months old (I'm
>not talking about the marketing logo, I'm talking about the framework, the
>development tools and so on). Where was Java 7 months after it's first
>release version? Cleaning up the AWT mess? Swing wasn't even at the design
>stage at that moment, to get back to the original point... .NET does all
>Java does and it does it somewhat better (imo, surely you'll say a different
>story).
> If I'm right, you'll be learning .NET programming in a couple of years.
If
>not, then I'll switch entirely to Java (I already know Java so it won't
be a
>severe shock for me). So what?
> Ovidiu.
>
>

[Markn]

I studying C# (beta) a year or so ago and was laughing (sort of) as I read.
I kept saying - this is Java.

But I agree. In many ways it isn't. .Net is so limiting.


"Brad O'Hearne" <brado@neurofire.com> wrote:
>Hey Rob, BradO here...
>
>"Rob Abbe" <rabbe@mn.rr.com> wrote in message news:3d8b1d0c$1@10.1.10.29...
>> Because .Net is just an off brand version of Java, simply by learning
.Net
>> you're already taking great strides in learning Java.
>
>I'm curious, can you clarify this statement? Strictly speaking, .NET isn't
>Java in any way. But I am sure you were just drawing a parallel here.
>Anyway, I am interested in what you meant.
>
>BradO
>--
>Brad O'Hearne
>brado@neurofire.com
>DevX Section Leader
>
>

[Brad O'Hearne]

Hey Rob, BradO here...

"Rob Abbe" <rabbe@mn.rr.com> wrote in message news:3d8b1d0c$1@10.1.10.29...
> Because .Net is just an off brand version of Java, simply by learning .Net
> you're already taking great strides in learning Java.

I'm curious, can you clarify this statement? Strictly speaking, .NET isn't
Java in any way. But I am sure you were just drawing a parallel here.
Anyway, I am interested in what you meant.

BradO
--
Brad O'Hearne
brado@neurofire.com
DevX Section Leader

[Rob Abbe]

BradO,

Thanks for giving me the chance to back pedal, but I never pull a punch on
this issue. I feel that .Net seems to be what J++ would have been had Sun
not stepped in. A version of Java limited (currently) to the Windows platform.

What do you see as the major differences? C# is an unsafe spin off of the
Java language (i.e. the unsafe keyword) Also, it compiles to byte code(MISL),
the plumbing is very similar... garbage collection and finalization not to
mention it is JIT compiled. The model for JIT is different, but JIT by any
other name...

I see them as being fundamentally the same.

Rob

(Also, I don't work for Sun as has been suggested by others in the past)

"Brad O'Hearne" <brado@neurofire.com> wrote:
>Hey Rob, BradO here...
>
>"Rob Abbe" <rabbe@mn.rr.com> wrote in message news:3d8b1d0c$1@10.1.10.29...
>> Because .Net is just an off brand version of Java, simply by learning
.Net
>> you're already taking great strides in learning Java.
>
>I'm curious, can you clarify this statement? Strictly speaking, .NET isn't
>Java in any way. But I am sure you were just drawing a parallel here.
>Anyway, I am interested in what you meant.
>
>BradO
>--
>Brad O'Hearne
>brado@neurofire.com
>DevX Section Leader
>
>

[Amused Reader]

>you will see that there are just about 2 Java jobs for every 1 VB job.
>
>It seems that .Net has not created much in the way of NEW opportunities
for
>Microsoft developers. That is not to say that existing shops are not using
>it, but it has not been the Java killer MS hoped it would be... not even
>close.
>
>Rob

Funny interpretation you have.... twice the number of java job opportunities
going unfilled for lack of interested developers for every VB opportunity.
Normally, in a job market place, when demand rises, over time, supply follows
to match demand and eventually exceed it. When demand for a particular skill,
over the long run (NOT short run) fails to be met with a rising supply, that
means the interest in providing that skill is falling, not rising, as is
the case with Java. Recent studies have even now started to show this drop
in demand, rather than the expectation of supply exceeding demand, as in
the case for lawyers, MBAs, and any other "hot" profession/skill in the past
20 years.

Here's an analogy. I have some money to invest (aka a technical problem to
solve). My brother-in-law convinces me to invest the money in a particular
stock A(aka, solve the problem with a particular technology, hence 'invest"
in it). I invest more money in the stock than I have, also known as borrowing
on margin. (aka the problem was bigger than I can solve myself). I must sell
some stock to meet a margin call (aka hire a java developer who has also
'invested' in the technology to help me meet my need). I place an ad on monster.com
for a java job (aka I place a sell order for some of the stock to share the
investment). Now, we find that there are far more jobs available for Java
skills, than another skill (aka more sell orders for stock A than stock B).


Based on your twisted logic, I should believe that stock A is far more in
demand than stock B, because there are more sell orders than can be met in
A, than B. ROTFLMAO.....

[Rob Abbe]

Amused,

Show me your studies. Evans and Gartner dispute what you are saying. It
is simple supply and demand as you suggest. When there is demand, a company
will go in search of developers by listing ads in news papers or websites.
The demand for VB has dropped way off. I don't think I can make it any
simpler for you. Java developers are in high demand, that is why they are
paid more than VB developers as well. Check the salary survey in Java Pro
for further evidence of this.

Rob

"Amused Reader" <anoymous@hotmail.com> wrote:
>
>>you will see that there are just about 2 Java jobs for every 1 VB job.
>>
>>It seems that .Net has not created much in the way of NEW opportunities
>for
>>Microsoft developers. That is not to say that existing shops are not using
>>it, but it has not been the Java killer MS hoped it would be... not even
>>close.
>>
>>Rob
>
>Funny interpretation you have.... twice the number of java job opportunities
>going unfilled for lack of interested developers for every VB opportunity.
> Normally, in a job market place, when demand rises, over time, supply follows
>to match demand and eventually exceed it. When demand for a particular skill,
>over the long run (NOT short run) fails to be met with a rising supply,
that
>means the interest in providing that skill is falling, not rising, as is
>the case with Java. Recent studies have even now started to show this drop
>in demand, rather than the expectation of supply exceeding demand, as in
>the case for lawyers, MBAs, and any other "hot" profession/skill in the
past
>20 years.
>
>Here's an analogy. I have some money to invest (aka a technical problem
to
>solve). My brother-in-law convinces me to invest the money in a particular
>stock A(aka, solve the problem with a particular technology, hence 'invest"
>in it). I invest more money in the stock than I have, also known as borrowing
>on margin. (aka the problem was bigger than I can solve myself). I must
sell
>some stock to meet a margin call (aka hire a java developer who has also
>'invested' in the technology to help me meet my need). I place an ad on
monster.com
>for a java job (aka I place a sell order for some of the stock to share
the
>investment). Now, we find that there are far more jobs available for Java
>skills, than another skill (aka more sell orders for stock A than stock
B).
>
>
>Based on your twisted logic, I should believe that stock A is far more in
>demand than stock B, because there are more sell orders than can be met
in
>A, than B. ROTFLMAO.....
>
>