Java is Going to Be the Death of Java
Jack Richards
open) is touted as the great solution, when it is actually corrupting the
world of real technologies.
Take Jini. When you look at it, it has a bunch of cute marketing names for
such things as introspection (discovery; compare IUnknown in COM),
repository (JavaSpaces), etc.
Jini has basically taken the concept of shippable places, for distributed
objects, and pretended it's some great Sun invention when done on a network.
Curiously, Sun talks about object oriented in terms of components, as in the
separation of interface and implementation, when the OO concepts were
encapsulation and inheritence. See
http://www.sunworld.com/swol-06-1999/swol-06-jiniology.html.
"Jini attempts to raise the level of abstraction for distributed systems
programming, from the network protocol level to the object interface level."
Gee, a stroke of genius. Except it had already been raised. Again, why
borrow the concept of interface-based or componentized systems, while
limiting what is basically a non-language standard to a single language?
This is technological corruption.
Do it but don't call it open. Call it a proprietary language platform
disguished as openness. It goes like this: we are heavily invested in a
particular area, in this case, a language. Therefore we will "dumb down" the
basic technology we are using, ie binary-based, not language-based,
components.
Then wait for the applause from the suckers. Who then, because they have so
much invested in this abberation, go to extreme lengths mentally to defend
it, resulting in the dimunition of the professional judgements of so many.
Why not a little honesty so an accurate assessment can be made of
technologies, instead of burying them in emotional agendas?
For devices, a universal VM would be appropriate to an intellectually honest
"open" standard, particularly one that included languages and other
technologies that have already been used successfully with devices.
Other perpherials, such as printers, already use VMs, so why not have a
universal VM that builds on the work already done?
For componentized systems, why bastardize the concept in the interest of
controlling the intellectual property? Fundamentally, all the work done in
the industry on components was to overcome the very problem of the limits of
high level languages in achieving interoperability.
In that sense, Java is simply a throwback. Java was developed in the era of,
and strongly influenced by, components. It has strong characteristics of
interface-based systems, within truly being a component. Since the component
functionality is handled by the componentized system, it is likely that more
flexible languages that do not have to carry this burden will be more
successful in the long run.
It's very similar to the problem CORBA is having adjusting to TP
infrastructure. CORBA has so much of this baggage, that simpler systems like
COM/DCOM work, and particularly scale better, when the TP and other
infrastructure handle all the service baggage, for security, etc.
This is the whole lesson learned with OO languages in general. If you build
in systems for controlling modules, so that it is absolutely enforced with
interfaces, then code breaks free of monolithic structures and become
reuseable. The very success of OO languages in polymorphism had created its
own problems.
Then once interface based systems are utilized, componentized systems break
free of the language constraint.
Hence, components had a binary, not a language standard, so multiple
languages could be supported. Creating a componentized system with a
particular language as its standard, like JavaBeans, is a second rate vision
of interoperability disguished as some extraordinary idealism.
Calling a componentized system a cute marketing like Beans does not excuse
corrupting the basic technology of components.
And it will be the ruin of the Beans system,as ardor for Java cools off, as
it always does with these language fads end. Components, because they are a
binary standard, have the advanage of supporting the next language fad, and,
it is obvious that more will follow because we are in the front end of this
communications web era.
Now we have Jini which Sun will pump up, in order to replace CORBA, into
some version of DCOM (after trashing it for so long) by having EJB and other
infrastructure handle all the heavy lifting, as DCOM does with MTS, etc.
(which Sun the great "innovator" is furiously copying, while continuing to
harp competitor-can't-innovate party line).
Jack
Yann
> Now we have Jini which Sun will pump up, in order to replace CORBA, into
> some version of DCOM (after trashing it for so long) by having EJB and other
> infrastructure handle all the heavy lifting, as DCOM does with MTS, etc.
> (which Sun the great "innovator" is furiously copying, while continuing to
> harp competitor-can't-innovate party line).
>
> Jack
Huh?
How to you look up for a DCOM component again?
CoCreateInstance with the server host name as an argument?
I'm actualy glad they didn't copy that.
Yann.
Osvaldo Pinali Doederlein
> Take Jini. When you look at it, it has a bunch of cute marketing names for
> such things as introspection (discovery; compare IUnknown in COM),
> repository (JavaSpaces), etc.
It's funny how you always use COM as an example of the right/original way to
do things. IUnknown and COM's dynamic features are a big joke compared to
what you have in CORBA (Interface Repository, DSI/DII, Implementation
Repository, all sorts of hooks like interceptors, etc.)...
> "Jini attempts to raise the level of abstraction for distributed systems
> programming, from the network protocol level to the object interface
level."
> Gee, a stroke of genius. Except it had already been raised. Again, why
> borrow the concept of interface-based or componentized systems, while
> limiting what is basically a non-language standard to a single language?
> This is technological corruption.
Do you understand the advantages of not having hardwired network protocols??
Of course you are trading it for the obligation of using Java (at least the
bytecode, and at least in some points of the network) and of course the
tradeoff is good for Sun, but it has real benefits anyway.
> For devices, a universal VM would be appropriate to an intellectually
honest
> "open" standard, particularly one that included languages and other
> technologies that have already been used successfully with devices.
Granted. Now, there is a difference between people who talk about
technology and people who do it. I hear lots of whining from academics and
from Smalltalk vendors -- "the JVM is too Java-specific!! Aaargh!! I cannot
use it efficiently for Scheme / Smalltalk / whatever!! The bytecode
sucks!!" -- then tell me, why didn't the industry develop and promote an
Universal VM technology that would make everybody happy? You know, VMs are
not anything new. But what happens is that only Sun succeeded in having a
VM installed on millions of sites and becoming ubiquitous. After Sun did
the hard part, it's easy to cry and ask them to make the JVM more
"universal".
And the hard truth is, I have yet to see an UVM. Some fundamental choices
will have big impact on different languages: for example, you can either
support primitive types as raw machine values or as boxed types; I don't
think you can reasonably mix both approaches in a single VM, and either
option will suck for half of the existing OOPLs. And the security
advantages of Java require certain constraints on bytecode that ain't good
for other languages. There's no such a things as a true UVM -- AFAIK IBM's
UVM is about customization: a framework from which they can instantiate
specialized VMs; but it's not the same thing.
> Hence, components had a binary, not a language standard, so multiple
> languages could be supported. Creating a componentized system with a
> particular language as its standard, like JavaBeans, is a second rate
vision
> of interoperability disguished as some extraordinary idealism.
> Calling a componentized system a cute marketing like Beans does not excuse
> corrupting the basic technology of components.
The Java language is not the standard and it is not important. Java
bytecode is the standard. It is still much less interoperable than e.g.
CORBA, but it's a tradeoff you're free to choose -- you can program to your
favorite interoperable middleware with Java.
> Now we have Jini which Sun will pump up, in order to replace CORBA, into
> some version of DCOM (after trashing it for so long) by having EJB and
other
> infrastructure handle all the heavy lifting, as DCOM does with MTS, etc.
> (which Sun the great "innovator" is furiously copying, while continuing to
> harp competitor-can't-innovate party line).
You surely forget that all this new stuff as COM+, DNA, etc. are ripoff of
OMG's Object Management Architecture, and EJB is copying CORBA more than
COM. Gee, when I see that now COM+ offers a publish/subscribe event
service, security, distributed naming and blah blah... and Java too... I
used to give courses on these things using CORBA three years ago.
A+
Osvaldo
Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.
George Reese
: Nothing is more sad than the pretended idealism where Java (even if it were
: open) is touted as the great solution, when it is actually corrupting the
: world of real technologies.
: Take Jini. When you look at it, it has a bunch of cute marketing names for
: such things as introspection (discovery; compare IUnknown in COM),
: repository (JavaSpaces), etc.
Those are not cute names. They are descriptive names of certain bits
of functionality.
: Jini has basically taken the concept of shippable places, for distributed
: objects, and pretended it's some great Sun invention when done on a network.
Sun did it and has shown it working. That is an acccomplishment.
: Curiously, Sun talks about object oriented in terms of components, as in the
: separation of interface and implementation, when the OO concepts were
: encapsulation and inheritence. See
: http://www.sunworld.com/swol-06-1999/swol-06-jiniology.html.
Do you know anything about OO, really? There are FOUR pillars to the
OO structure, encapsulation, abstraction, polymorphism, and
inheritance. Java manages all four quite well.
: "Jini attempts to raise the level of abstraction for distributed systems
: programming, from the network protocol level to the object interface level."
: Gee, a stroke of genius. Except it had already been raised. Again, why
: borrow the concept of interface-based or componentized systems, while
: limiting what is basically a non-language standard to a single language?
: This is technological corruption.
Really? Most languages out there simply are not up to Java's standards
in supporting OO concepts. You should take a look at Jim Waldo's
discussions on why it is beneficial to assume a particular language.
: Do it but don't call it open.
It is open. Anyone can implement it. Even in other languages. Give me
something you have in Java for the Java platform and I can provide the
exact same functionality and interopability in Python.
: Call it a proprietary language platform
: disguished as openness.
You can call it that. You would be incorrect.
: It goes like this: we are heavily invested in a
: particular area, in this case, a language. Therefore we will "dumb down" the
: basic technology we are using, ie binary-based, not language-based,
: components.
: Then wait for the applause from the suckers. Who then, because they have so
: much invested in this abberation, go to extreme lengths mentally to defend
: it, resulting in the dimunition of the professional judgements of so many.
: Why not a little honesty so an accurate assessment can be made of
: technologies, instead of burying them in emotional agendas?
: For devices, a universal VM would be appropriate to an intellectually honest
: "open" standard, particularly one that included languages and other
: technologies that have already been used successfully with devices.
Why is the Java VM not a universal VM?
: Other perpherials, such as printers, already use VMs, so why not have a
: universal VM that builds on the work already done?
Why is the Java VM not a universal VM?
[ a bunch of false history deleted ]
: This is the whole lesson learned with OO languages in general. If you build
: in systems for controlling modules, so that it is absolutely enforced with
: interfaces, then code breaks free of monolithic structures and become
: reuseable. The very success of OO languages in polymorphism had created its
: own problems.
: Then once interface based systems are utilized, componentized systems break
: free of the language constraint.
: Hence, components had a binary, not a language standard, so multiple
: languages could be supported. Creating a componentized system with a
: particular language as its standard, like JavaBeans, is a second rate vision
: of interoperability disguished as some extraordinary idealism.
You are on drugs. Like I said, any component you can write in Java, I
can write in python and have it interoperate with the rest of your
application. Thus, your statement above is patently false.
--
George Reese (bo...@imaginary.com)
http://www.imaginary.com/~borg
"The dead know only one thing: it's better to be alive"
-Joker in Full Metal Jacket
D'Arcy Smith
> Nothing is more sad than the pretended idealism where Java (even if it were
> open) is touted as the great solution, when it is actually corrupting the
> world of real technologies.
> Take Jini. When you look at it, it has a bunch of cute marketing names for
> such things as introspection (discovery; compare IUnknown in COM),
> repository (JavaSpaces), etc.
COM runs well on my Linux box... especialy all those COM components
that are compiled to Win32 (like all of them).
COM is not a distribited service framework... Jini is.
> Jini has basically taken the concept of shippable places, for distributed
> objects, and pretended it's some great Sun invention when done on a network.
What does COM do for me in this area? What other technologies do the
same thing?
> "Jini attempts to raise the level of abstraction for distributed systems
> programming, from the network protocol level to the object interface level."
> Gee, a stroke of genius. Except it had already been raised. Again, why
> borrow the concept of interface-based or componentized systems, while
> limiting what is basically a non-language standard to a single language?
> This is technological corruption.
So that the code is _PORTABLE_ to any client system.
Read the article on JavaSoft about Bill Joy... it'll help
clear some things up gor you.
> Then wait for the applause from the suckers. Who then, because they have so
> much invested in this abberation, go to extreme lengths mentally to defend
> it, resulting in the dimunition of the professional judgements of so many.
Please name another technology that lets me do a similar thing
as Jini (there are some - I just want to see what you think
they are).
> Other perpherials, such as printers, already use VMs, so why not have a
> universal VM that builds on the work already done?
And where is this universal VM? JavaBytecodes are not specific to
the Java language either BTW. You can compile Scheme, ada and Python
into Java bytecodes (IIRC).
> For componentized systems, why bastardize the concept in the interest of
> controlling the intellectual property? Fundamentally, all the work done in
> the industry on components was to overcome the very problem of the limits of
> high level languages in achieving interoperability.
You totaly do no understand what Jini is doing by the looks of it.
Please summarise what your think Jini is for.
Please sumarise how you would do it differently with todays technology
(that means unless you have a Universal VM you can't make use of
one in your answer).
> It's very similar to the problem CORBA is having adjusting to TP
> infrastructure. CORBA has so much of this baggage, that simpler systems like
> COM/DCOM work, and particularly scale better, when the TP and other
> infrastructure handle all the service baggage, for security, etc.
Java doesn't compete with COM/CORBA... what does this have to do with
anything?
> Hence, components had a binary, not a language standard, so multiple
> languages could be supported. Creating a componentized system with a
> particular language as its standard, like JavaBeans, is a second rate vision
> of interoperability disguished as some extraordinary idealism.
Too bad your COM components are totoaly useless in a non Win32
environment. Too bad your CORBA components are not useable on
the client - the code cannot be downloaded from the server and
run on the client.
Looks like you don't understand Java at all...
> Calling a componentized system a cute marketing like Beans does not excuse
> corrupting the basic technology of components.
Because they can "Run Anywhere" it is a corruption?
> And it will be the ruin of the Beans system,as ardor for Java cools off, as
> it always does with these language fads end. Components, because they are a
> binary standard, have the advanage of supporting the next language fad, and,
> it is obvious that more will follow because we are in the front end of this
> communications web era.
Componenets, because they are a binary standard, are useless in a
heterogeneous computing environment.
> Now we have Jini which Sun will pump up, in order to replace CORBA, into
> some version of DCOM (after trashing it for so long)
Since you don't understand Java or Jini I don't see how you can
make this argument. ALso since eveything else you posted was wrong
I'm safe in saying this is wrong too.
> by having EJB and other infrastructure handle all the heavy lifting,
> as DCOM does with MTS, etc. (which Sun the great "innovator" is furiously
> copying, while continuing to harp competitor-can't-innovate party line).
Who cares if Sun is copying MS (don't agree with you but...) - MS hasn't
invented one thing. Ever. Period.
..darcy
Yann
Programmatic ID.
Package names are so cool...
Yann.
anonc...@my-deja.com
"Jack Richards" <jrxxx...@worldnet.com> wrote:
> Take Jini. When you look at it.
Which you obviously haven't, let alone used it.
What is all this crap about COM,COM+,MTS... this is related to Jini how?
Sounds like you need to get that chip off your shoulder and go work for
a different company before you go insane!
> Jack knows (ooh, its tempting ...)
Reality is a point of view
| Do you know anything about OO, really? There are FOUR pillars to the
| OO structure, encapsulation, abstraction, polymorphism, and
| inheritance. Java manages all four quite well.
Mr. Kay seems to disagree.
--
Gary Johnson gjoh...@season.com
Privacy on the net is still illegal.
Jack Richards
Yann wrote in message <37BBD14E...@ims.ltd.uk>...
>Jack Richards wrote:
>
>> Now we have Jini which Sun will pump up, in order to replace CORBA, into
other
>> infrastructure handle all the heavy lifting, as DCOM does with MTS, etc.
>> (which Sun the great "innovator" is furiously copying, while continuing
to
>> harp competitor-can't-innovate party line).
>>
>
>Huh?
>
>How to you look up for a DCOM component again?
>CoCreateInstance with the server host name as an argument?
>
Yann,
I'm not touting DCOM. I'm just pointing out the shift of services to the
infrastructure. Here's an article from 1998 on the importance of the
management issue:
http://www.dbmsmag.com/9805d07.html.
I usually don't get into these "copy" games, mainly because it is difficult
enough to understand computer technologies (in the sense of understanding
the implications of things) without pretending the ability to assess their
influences.
That's just a little dig because Sun, as well as everyone else, has
obviously been borrowing all over the place. The Sun-MS copyright (contract)
suit would have examined at trail what part of Java was original with Sun.
The copyright Altai test, used most prominently in the MS-Apple GUI case,
requires the judge to go through a particular procedue in reaching the
decision. Of course, the "home court" judge issued a summary judgement
instead on the copyright issues aspect. This has virtually no chance of
making it on appeal.
Jack
>
>Yann.
>
Jack Richards
Osvaldo Pinali Doederlein wrote in message
<002b01beea36$142dec00$4505...@mde.emn.fr>...
>From: Jack Richards <jrxxx...@worldnet.com>
>> Take Jini. When you look at it, it has a bunch of cute marketing names
for
>> such things as introspection (discovery; compare IUnknown in COM),
>> repository (JavaSpaces), etc.
>
to
>do things.
Where did you read that in?
>>searching between the lines
> IUnknown and COM's dynamic features are a big joke compared to
>what you have in CORBA (Interface Repository, DSI/DII, Implementation
>Repository, all sorts of hooks like interceptors, etc.)...
>
>> "Jini attempts to raise the level of abstraction for distributed systems
>> programming, from the network protocol level to the object interface
>level."
>> Gee, a stroke of genius. Except it had already been raised. Again, why
>> borrow the concept of interface-based or componentized systems, while
>> limiting what is basically a non-language standard to a single language?
>> This is technological corruption.
>
>Do you understand the advantages of not having hardwired network
protocols??
Agreed. Again you read your own mind into my writing. That is such an
elemental point, but Sun likes to market these basics as exciting new ideas.
Sun, you know, has replaced Microsoft as the great marketeer.
>Of course you are trading it for the obligation of using Java (at least the
>bytecode, and at least in some points of the network) and of course the
>tradeoff is good for Sun, but it has real benefits anyway.
>
>> For devices, a universal VM would be appropriate to an intellectually
>honest
>> "open" standard, particularly one that included languages and other
>> technologies that have already been used successfully with devices.
>
>Granted. Now, there is a difference between people who talk about
>technology and people who do it. I hear lots of whining from academics and
>from Smalltalk vendors -- "the JVM is too Java-specific!! Aaargh!! I
cannot
>use it efficiently for Scheme / Smalltalk / whatever!! The bytecode
>sucks!!" -- then tell me, why didn't the industry develop and promote an
>Universal VM technology that would make everybody happy? You know, VMs are
>not anything new. But what happens is that only Sun succeeded in having a
>VM installed on millions of sites and becoming ubiquitous. After Sun did
>the hard part, it's easy to cry and ask them to make the JVM more
>"universal".
You raise some good points. Innovation is indeed getting some tried and true
"concept" to actually work.
But don't keep feeding the "openness" mantra, when you basically do
something that favors you in the market that is not really open in important
respects.
Of course, when you become language dependent, and a new language fad pops
on the scene, don't be complaining how someone else's marketing or whatever
caused your "vision" to fail.
>And the hard truth is, I have yet to see an UVM. Some fundamental choices
>will have big impact on different languages: for example, you can either
>support primitive types as raw machine values or as boxed types; I don't
>think you can reasonably mix both approaches in a single VM, and either
>option will suck for half of the existing OOPLs. And the security
>advantages of Java require certain constraints on bytecode that ain't good
>for other languages. There's no such a things as a true UVM -- AFAIK IBM's
>UVM is about customization: a framework from which they can instantiate
>specialized VMs; but it's not the same thing.
Yes, there are always tradeoffs, aren't there? It's certainly beyond my
ability to assess, especially as to the implications of these technologies
in use over time in many scenarios.
>> Hence, components had a binary, not a language standard, so multiple
>> languages could be supported. Creating a componentized system with a
>> particular language as its standard, like JavaBeans, is a second rate
>vision
>> of interoperability disguished as some extraordinary idealism.
>> Calling a componentized system a cute marketing like Beans does not
excuse
>> corrupting the basic technology of components.
>
>The Java language is not the standard and it is not important. Java
>bytecode is the standard. It is still much less interoperable than e.g.
>CORBA, but it's a tradeoff you're free to choose -- you can program to your
>favorite interoperable middleware with Java.
You miss my point. I have no problem with what you are saying about Java
bytecodes, etc.. It is JavaBeans, Sun's componentized system that I am
talking about. Confining it to the Java language, IMO, is the offense I
mentioned.
>
>> Now we have Jini which Sun will pump up, in order to replace CORBA, into
>> some version of DCOM (after trashing it for so long) by having EJB and
>other
>> infrastructure handle all the heavy lifting, as DCOM does with MTS, etc.
>> (which Sun the great "innovator" is furiously copying, while continuing
to
>> harp competitor-can't-innovate party line).
>
>You surely forget that all this new stuff as COM+, DNA, etc. are ripoff of
>OMG's Object Management Architecture, and EJB is copying CORBA more than
>COM. Gee, when I see that now COM+ offers a publish/subscribe event
>service, security, distributed naming and blah blah... and Java too... I
>used to give courses on these things using CORBA three years ago.
The CORBA-COM rivalry began back in 1986, I believe. See my reply to Yann
about "copying."
Here is an interesting statement from Roger Sessions on the CORBA-COM
rivalry:
"JAVA HISTORY
Lets start with the history of Java. Java was introduced by Sun. Many
companies jumped on the Java bandwagon. The most notable of these early
adopters was IBM.
What fueled this odd alliance between Sun and IBM? The incredible
superiority of Java as a programming technology? Unlikely. Remember, this
alliance was formed back when Java had the most minimal of capabilities and
was hardly more than a toy.
We can get some insight into the purpose of this alliance by looking at the
last time these two companies joined together to lead a major industry
effort. This last collaboration resulted in the distributed object
technology known as CORBA, produced by an industry consortium called the
Object Management Group (OMG).
I was very involved in the OMG. I worked for IBM at the time, and I was a
lead architect for one of the CORBA services. I wrote my last book on this
topic (Object Persistence; Beyond Object-Oriented Databases).
The OMG consortium was held together by one common purpose: to stop
Microsoft. The OMG failed in this effort, and is now limping along with only
a few companies making even a minimal profit on CORBA technologies.
Many people accuse Microsoft of meeting any technological threat with an
embrace and smother strategy. This is certainly not true of Microsoft's
reaction to CORBA. Despite the millions of dollars that IBM, Sun, and others
were spending on CORBA, Microsoft never took CORBA seriously. Microsoft
joined the OMG, and occasionally sent a representative to OMG meetings, but
looked on the whole affair with an air of amusement, as if it was watching
children at play.
By 1994, it was clear to Sun and IBM (and later Netscape and the rest of the
industry) that OMG was not going to be the knight in shining armor that
would protect them from the Microsoft dragon. So they started looking for a
new protector. And they thought they found it in Java. Soon most of the
industry had jumped on the Love-Java/Stop-Microsoft bandwagon." See
http://www.objectwatch.com/issue7.htm"
Jack
>
>A+
>Osvaldo
Jack Richards
George Reese wrote in message ...
>Jack Richards <jrxxx...@worldnet.com> wrote:
>: Nothing is more sad than the pretended idealism where Java (even if it
were
>: open) is touted as the great solution, when it is actually corrupting the
>: world of real technologies.
>
for
>: such things as introspection (discovery; compare IUnknown in COM),
>: repository (JavaSpaces), etc.
>
>of functionality.
>
>: objects, and pretended it's some great Sun invention when done on a
network.
>
>
>: Curiously, Sun talks about object oriented in terms of components, as in
the
>: separation of interface and implementation, when the OO concepts were
>: encapsulation and inheritence. See
>: http://www.sunworld.com/swol-06-1999/swol-06-jiniology.html.
>
>OO structure, encapsulation, abstraction, polymorphism, and
>inheritance. Java manages all four quite well.
>
>: programming, from the network protocol level to the object interface
level."
>
>: Gee, a stroke of genius. Except it had already been raised. Again, why
>: borrow the concept of interface-based or componentized systems, while
>: limiting what is basically a non-language standard to a single language?
>: This is technological corruption.
>
>in supporting OO concepts. You should take a look at Jim Waldo's
>discussions on why it is beneficial to assume a particular language.
>
>: Do it but don't call it open.
>
>It is open. Anyone can implement it. Even in other languages. Give me
>something you have in Java for the Java platform and I can provide the
>exact same functionality and interopability in Python.
>
>: Call it a proprietary language platform
>: disguished as openness.
>
>You can call it that. You would be incorrect.
>
>: It goes like this: we are heavily invested in a
>: particular area, in this case, a language. Therefore we will "dumb down"
the
>: basic technology we are using, ie binary-based, not language-based,
>: components.
>
so
>: much invested in this abberation, go to extreme lengths mentally to
defend
>: it, resulting in the dimunition of the professional judgements of so
many.
>
>: technologies, instead of burying them in emotional agendas?
>
honest
>: "open" standard, particularly one that included languages and other
>: technologies that have already been used successfully with devices.
>
>
>: universal VM that builds on the work already done?
>
>
>[ a bunch of false history deleted ]
>
>: This is the whole lesson learned with OO languages in general. If you
build
>: in systems for controlling modules, so that it is absolutely enforced
with
>: interfaces, then code breaks free of monolithic structures and become
>: reuseable. The very success of OO languages in polymorphism had created
its
>: own problems.
>
>: Then once interface based systems are utilized, componentized systems
break
>: free of the language constraint.
>
>: languages could be supported. Creating a componentized system with a
>: particular language as its standard, like JavaBeans, is a second rate
vision
>: of interoperability disguished as some extraordinary idealism.
>
>can write in python and have it interoperate with the rest of your
>application. Thus, your statement above is patently false.
This a well known fact. See
http://www.devx.com/upload/free/features/javapro/1999/07jul99/ts0799/ts0799.
asp
(sorry about the ASP page).
"CORBA Anyone?
Sun has been a member of the Object Management Group for almost a decade
now, so it wasn't exactly surprising to find that last December's release of
Java 2, Standard Edition, contained much-improved support for CORBA. (See
"Demystifying CORBA," by Scott Grant on page 22 for more on that subject.)
And in doing research for this article, it was interesting to note how Sun's
representatives seemed to be going out of their way to avoid positioning EJB
as a replacement for CORBA. It would seem that Java is growing friendlier
with CORBA.
As a practical matter, however, the CORBA doomsayers are probably correct.
CORBA is a solution for heterogeneous environments, both in terms of
platforms and in terms of languages. In many respects, Java is a solution in
both those domains. If you are designing a brand-new distributed application
in Java, there's little reason to use CORBA instead of Enterprise JavaBeans.
CORBA will continue to be useful for connecting heterogeneous and legacy
systems, but as Java extends through the enterprise we can expect such
requirements to diminish. Indeed, it looks like CORBA 3.0's adoption of the
EJB model is an elegant acquiescence by the members of the OMG to Java's
increasingly pervasive presence in the enterprise."
Jack
Jack Richards
D'Arcy Smith wrote in message <37BC0F60...@itools.symantec.com>...
>Jack Richards wrote:
>
>> Nothing is more sad than the pretended idealism where Java (even if it
were
>> open) is touted as the great solution, when it is actually corrupting the
>> world of real technologies.
>
>> Take Jini. When you look at it, it has a bunch of cute marketing names
for
>> such things as introspection (discovery; compare IUnknown in COM),
>> repository (JavaSpaces), etc.
>
>that are compiled to Win32 (like all of them).
>
>COM is not a distribited service framework... Jini is.
Since any componentized systems offers services at an any interface, and is
therefore said to be a service extensible model.
Tivoli in 1989 started developing a distributed object framework for network
management. Devices are a lot easier that managing more complex entities.
Why are people skeptical about Jini? Basically, Sun takes these techologies
and bastardizes them to fit the Java platform. Basically, it's a square Java
peg in a round technolgical hole result. None of this stuff needs Java.
Sometimes it's overkill and sometimes it's underkill, but its the same
balony.
How does it get away with this? Because it has the Java faithful.
The Java hype is just about over. And its practicioneers are gradually
becoming the laughingstock of the industry. The key test will be whether
Java servlets stack up against all non-Java technologies that do the job
better.
It MUST succeed here, not just in doing something as well as what it
replaces. It has to deliver superior performance in at least one
technological niche.
If not, the industry will move on, and Java will be there as a marginally
useful language as well as a wannabee collection of disparate technologies
that will be running jokes for their pretension in taking over the world,
the internet, etc.
The failure of the ABM camp to block the integration of the browser in the
OS is going to mean that users forming powerful markets on the desktop,
including point and click ecommerce solutions, and low-level plug-in
programming to replace the Java-ites.
If Java doesn't offer plug in functionality for the user, then someone else
will. The result will rule the market.
It's hard to get anything about markets through to the server culture crowd.
They don't realize server software will be ubiquitous, that every user will
be a producer.
The first killer app that allows all the desktop functionality to be
incorporated in a web container, subject to being distributed, will drive
the rest of the market, up through the mid-point of the server markets.
Just as components have suddenly dominated the server software markets, such
that Sun is desperately pushing JavaBeans in order to be part of it, then a
like phenomenon will occur on the client side.
The whole of components is to stabilize the code, so plug-in software can be
built on top of it. Once this is done, then "apps" to manipulate the
components can be built that are simply point and click.
The attempt to subsume the client as a device in a basic server economy has
failed. Now we will see how much of the server market, including programmers
and administrators, will be subsumed by the client market, soon to gain
considerable weight from the functionality required to do this.
Jack
>
>
>> Jini has basically taken the concept of shippable places, for distributed
>> objects, and pretended it's some great Sun invention when done on a
network.
>
>What does COM do for me in this area? What other technologies do the
>same thing?
>
>
>> "Jini attempts to raise the level of abstraction for distributed systems
>> programming, from the network protocol level to the object interface
level."
>
>
>> Gee, a stroke of genius. Except it had already been raised. Again, why
>> borrow the concept of interface-based or componentized systems, while
>> limiting what is basically a non-language standard to a single language?
>> This is technological corruption.
>
>So that the code is _PORTABLE_ to any client system.
>Read the article on JavaSoft about Bill Joy... it'll help
>clear some things up gor you.
>
>
>> Then wait for the applause from the suckers. Who then, because they have
so
>> much invested in this abberation, go to extreme lengths mentally to
defend
>> it, resulting in the dimunition of the professional judgements of so
many.
>
>Please name another technology that lets me do a similar thing
>as Jini (there are some - I just want to see what you think
>they are).
>
>
>> Other perpherials, such as printers, already use VMs, so why not have a
>> universal VM that builds on the work already done?
>
>And where is this universal VM? JavaBytecodes are not specific to
>the Java language either BTW. You can compile Scheme, ada and Python
>into Java bytecodes (IIRC).
>
>
>> For componentized systems, why bastardize the concept in the interest of
>> controlling the intellectual property? Fundamentally, all the work done
in
>> the industry on components was to overcome the very problem of the limits
of
>> high level languages in achieving interoperability.
>
>You totaly do no understand what Jini is doing by the looks of it.
>Please summarise what your think Jini is for.
>Please sumarise how you would do it differently with todays technology
>(that means unless you have a Universal VM you can't make use of
>one in your answer).
>
>
>> It's very similar to the problem CORBA is having adjusting to TP
>> infrastructure. CORBA has so much of this baggage, that simpler systems
like
>> COM/DCOM work, and particularly scale better, when the TP and other
>> infrastructure handle all the service baggage, for security, etc.
>
>Java doesn't compete with COM/CORBA... what does this have to do with
>anything?
>
>
>> Hence, components had a binary, not a language standard, so multiple
>> languages could be supported. Creating a componentized system with a
>> particular language as its standard, like JavaBeans, is a second rate
vision
>> of interoperability disguished as some extraordinary idealism.
>
>Too bad your COM components are totoaly useless in a non Win32
>environment. Too bad your CORBA components are not useable on
>the client - the code cannot be downloaded from the server and
>run on the client.
>
>Looks like you don't understand Java at all...
>
>
>> Calling a componentized system a cute marketing like Beans does not
excuse
>> corrupting the basic technology of components.
>
>Because they can "Run Anywhere" it is a corruption?
>
>
>> And it will be the ruin of the Beans system,as ardor for Java cools off,
as
>> it always does with these language fads end. Components, because they are
a
>> binary standard, have the advanage of supporting the next language fad,
and,
>> it is obvious that more will follow because we are in the front end of
this
>> communications web era.
>
>Componenets, because they are a binary standard, are useless in a
>heterogeneous computing environment.
>
>
>> Now we have Jini which Sun will pump up, in order to replace CORBA, into
>> some version of DCOM (after trashing it for so long)
>
>Since you don't understand Java or Jini I don't see how you can
>make this argument. ALso since eveything else you posted was wrong
>I'm safe in saying this is wrong too.
>
>
>> by having EJB and other infrastructure handle all the heavy lifting,
>> as DCOM does with MTS, etc. (which Sun the great "innovator" is furiously
>> copying, while continuing to harp competitor-can't-innovate party line).
>
D'Arcy Smith
> D'Arcy Smith wrote in message <37BC0F60...@itools.symantec.com>...
> >Jack Richards wrote:
> >> Nothing is more sad than the pretended idealism where Java (even if it
> were
> >> open) is touted as the great solution, when it is actually corrupting the
> >> world of real technologies.
> >
> >> Take Jini. When you look at it, it has a bunch of cute marketing names
> for
> >> such things as introspection (discovery; compare IUnknown in COM),
> >> repository (JavaSpaces), etc.
> >
> >COM runs well on my Linux box... especialy all those COM components
> >that are compiled to Win32 (like all of them).
> >COM is not a distribited service framework... Jini is.
> Since any componentized systems offers services at an any interface, and is
> therefore said to be a service extensible model.
Can you download a COM thing from a Win32 machine to my Linux machine
and
have it work?
> Why are people skeptical about Jini? Basically, Sun takes these techologies
> and bastardizes them to fit the Java platform.
Uh... no. Sun, in part, made the Java platform to allow for Jini.
> Basically, it's a square Java
> peg in a round technolgical hole result. None of this stuff needs Java.
> Sometimes it's overkill and sometimes it's underkill, but its the same
> balony.
Please show me how I can download code from a service and run it onto a
client using existing technology. You have no knowledge of the
OS/machine type in advance.
> How does it get away with this? Because it has the Java faithful.
No - because it works. It is a good useable solution.
> The failure of the ABM camp to block the integration of the browser in the
> OS is going to mean that users forming powerful markets on the desktop,
> including point and click ecommerce solutions, and low-level plug-in
> programming to replace the Java-ites.
Have your Win32 only world. Unfortunatly for you the world is
still larget than Win32.
> The first killer app that allows all the desktop functionality to be
> incorporated in a web container, subject to being distributed, will drive
> the rest of the market, up through the mid-point of the server markets.
The day I use a word processor in a browser somwbody can come
and stick the noose around my neck. It is a useless idea.
> The whole of components is to stabilize the code, so plug-in software can be
> built on top of it. Once this is done, then "apps" to manipulate the
> components can be built that are simply point and click.
Right and COM does't give me that. It gives me that on Win32... too
bad I am not always oon Win32.
---
So you still know nothing about Java or Jini obviously.
..darcy
Jack Richards
D'Arcy Smith wrote in message <37BC92ED...@itools.symantec.com>...
>Jack Richards wrote:
>
>> D'Arcy Smith wrote in message <37BC0F60...@itools.symantec.com>...
>> >Jack Richards wrote:
>
>> >> Nothing is more sad than the pretended idealism where Java (even if it
>> were
>> >> open) is touted as the great solution, when it is actually corrupting
the
>> >> world of real technologies.
>> >
>> >> Take Jini. When you look at it, it has a bunch of cute marketing names
>> for
>> >> such things as introspection (discovery; compare IUnknown in COM),
>> >> repository (JavaSpaces), etc.
>> >
>> >COM runs well on my Linux box... especialy all those COM components
>> >that are compiled to Win32 (like all of them).
>
>> >COM is not a distribited service framework... Jini is.
>
>> Since any componentized systems offers services at an any interface, and
is
>> therefore said to be a service extensible model.
>
>and
>have it work?
>
>
>
techologies
>> and bastardizes them to fit the Java platform.
>
>
>
>
>> peg in a round technolgical hole result. None of this stuff needs Java.
>> Sometimes it's overkill and sometimes it's underkill, but its the same
>> balony.
>
>client using existing technology. You have no knowledge of the
>OS/machine type in advance.
You really think this is hot stuff. ANY componentized system that is
designed for such a scenario would work just fine.
It's odd that Sun who is the last to develop anything approaching "plug and
play" on its own computers, which require highly trained professionals to
operate them, comes up with the "concept," which presupposes a Java-only
world, whether overkill or underkill via resources, then borrows these
ideas, and effectively steals the language of components, pretends this is
wonderful new stuff, and markets it to the media eager for an "electric car"
story.
Sun is actually replacing CORBA with a Java-language-only replacement. It's
nothing new, despite your swooning.
>> How does it get away with this? Because it has the Java faithful.
>
>No - because it works. It is a good useable solution.
>
>
>> The failure of the ABM camp to block the integration of the browser in
the
>> OS is going to mean that users forming powerful markets on the desktop,
>> including point and click ecommerce solutions, and low-level plug-in
>> programming to replace the Java-ites.
>
>Have your Win32 only world. Unfortunatly for you the world is
>still larget than Win32.
The Win32 world happens to be the world of the computer user, for the most
part. Your disdain for this world will effectively mean that this computer
user world will replace half baked technological solutions with real
technologies from whatever source that increase their collective power.
Isn't that what the Java-ites are doing? Increasing their worlds at the
expense of the user?
>> The first killer app that allows all the desktop functionality to be
>> incorporated in a web container, subject to being distributed, will drive
>> the rest of the market, up through the mid-point of the server markets.
>
>The day I use a word processor in a browser somwbody can come
>and stick the noose around my neck. It is a useless idea.
You don't get it. All the functionality of generations of programmers have
created for the user will be distributed. All it takes is a first bite of
this, and the limited functionality of Java re the desktop is history. It is
simply an underpowered technology.
If Java has actually worked as promised, Netscape would have taken over the
desktop. Now it's like some additional app. The productivity apps that
people use can get their web services via the OS. Why leave the app to open
Netscape?
The point I am making and you are avoiding with your disdain of the user, is
that the app user who gather web info, massages it in their app somehow,
cannot then distribute it before of the Netscape-Sun lock on the desktop
client.
You cannot publish with an ActiveX control, or otherwise, especially with an
underpowered applet (Java's trademark) because the COM technology is blocked
for anti-competitive reasons by the ABM alliance of server vendors.
Forget the whole world. This is the situation that has been enforced on the
PC desktop by an alien culture, who, like yourself, cares about their own
selfish interests.
Jack
>> The whole of components is to stabilize the code, so plug-in software can
be
>> built on top of it. Once this is done, then "apps" to manipulate the
>> components can be built that are simply point and click.
>
D'Arcy Smith
> >Please show me how I can download code from a service and run it onto a
> >client using existing technology. You have no knowledge of the
> >OS/machine type in advance.
> You really think this is hot stuff. ANY componentized system that is
> designed for such a scenario would work just fine.
Fine. Please describe how COM or CORBA can do what I have said above.
Be explicit.
> >Have your Win32 only world. Unfortunatly for you the world is
> >still larget than Win32.
> The Win32 world happens to be the world of the computer user, for the most
> part. Your disdain
Where do you get "disdain" from the above sentence?
It is a simple observation - the world is not 100% microsoft OSs.
Solutions that only run on one platform are pretty close to
useless over the long term.
> for this world will effectively mean that this computer
> user world will replace half baked technological solutions with real
> technologies from whatever source that increase their collective power.
> Isn't that what the Java-ites are doing? Increasing their worlds at the
> expense of the user?
How do you get that anything in Java is at the "expense of the user"?
> >> The first killer app that allows all the desktop functionality to be
> >> incorporated in a web container, subject to being distributed, will drive
> >> the rest of the market, up through the mid-point of the server markets.
> >The day I use a word processor in a browser somwbody can come
> >and stick the noose around my neck. It is a useless idea.
> You don't get it. All the functionality of generations of programmers have
> created for the user will be distributed. All it takes is a first bite of
> this, and the limited functionality of Java re the desktop is history. It is
> simply an underpowered technology.
Please show how COM or CORBA will work for this. Remeber the stuff
actualy
runs on the CLIENT computer NOT on the SERVER.
> If Java has actually worked as promised, Netscape would have taken over the
> desktop.
No.
> Now it's like some additional app. The productivity apps that
> people use can get their web services via the OS. Why leave the app to open
> Netscape?
Why did MS tie it to their browser - why not allow any browser component
to do the work?
> The point I am making and you are avoiding with your disdain of the user,
Still don't know how you "know" I have "disdain" for anything.
> is
> that the app user who gather web info, massages it in their app somehow,
> cannot then distribute it before of the Netscape-Sun lock on the desktop
> client.
And COM or CORBA helps me how?
> You cannot publish with an ActiveX control, or otherwise, especially with an
> underpowered applet (Java's trademark) because the COM technology is blocked
> for anti-competitive reasons by the ABM alliance of server vendors.
No COM won't run on my Linux box. Pretty simple.
> Forget the whole world. This is the situation that has been enforced on the
> PC desktop by an alien culture, who, like yourself, cares about their own
> selfish interests.
The world != PCs.
> >> The whole of components is to stabilize the code, so plug-in software can
> be
> >> built on top of it. Once this is done, then "apps" to manipulate the
> >> components can be built that are simply point and click.
> >Right and COM does't give me that. It gives me that on Win32... too
> >bad I am not always oon Win32.
> >So you still know nothing about Java or Jini obviously.
Well at least you don't disagree with me on that.
..darcy
bill@x
>
>
>The Win32 world happens to be the world of the computer user, for the most
>part.
This is such a simplistic view of the world.
75% of the world data is held on IBM mainframes.
close to 90% of the world network and backend processing is done
on non windows PC's.
Just becuase you use MS word, and 90% of the secreteries do, does
not mean the world runs on Win32.
Not only is win32 is a closed and runs on one platform only. But the Java API
is considered by many a much more powerfull and easier to use for
enterprise computing.
Java is not perfect, but it is the best thing we have right now.
Win32 and that cryptic modified C++ and COM+ is a lousy alternative. If you
choose that, go for it. Be married to Bill gate for the rest of your
life. The rest of us want open, industry wide API and a platform that
is available everyone. My Java applications run on Linux, windows,
solaris, VMS and AS400 unchanged, and I like that, and nothing in
the world will make me go back to win32 or C or C++. They will
have to drag me screaming and shouting to get me touch that rubbish
again.
Bill
Yann
However, why such a thread title then?
Yann.
hungr...@my-deja.com
the most
> >part.
>
>
> 75% of the world data is held on IBM mainframes.
> close to 90% of the world network and backend processing is done
> on non windows PC's.
>
Whoopee. We are talking about a battle for control of a seperate
middle tier that incorporates component technology. Of course, the
data tier can also benefit from component technology so the mainframe
dinosaurs are strictly legacy.
> Just becuase you use MS word, and 90% of the secreteries do, does
> not mean the world runs on Win32.
>
> Not only is win32 is a closed and runs on one platform only. But the
Java API
> is considered by many a much more powerfull and easier to use for
> enterprise computing.
>
Zzzt wrong. The tight integration COM has with its underlying
operating environment will result in better performance, simplified
administration, and a consistent programming model. These aren't
luxuries, but critical to the success of componentized systems.
> Java is not perfect, but it is the best thing we have right now.
> Win32 and that cryptic modified C++ and COM+ is a lousy alternative.
If you
> choose that, go for it. Be married to Bill gate for the rest of your
> life. The rest of us want open, industry wide API and a platform that
> is available everyone. My Java applications run on Linux, windows,
> solaris, VMS and AS400 unchanged, and I like that, and nothing in
> the world will make me go back to win32 or C or C++. They will
> have to drag me screaming and shouting to get me touch that rubbish
> again.
>
Oh my, another Java zealot telling users to go to hell.
> Bill
Yann
> The tight integration COM has with its underlying operating environment
> will result in better performance, simplified administration, and a
> consistent programming model. These aren't luxuries, but critical to the
> success of componentized systems.
"will"?
Is that a snapshot from a marketing brochure or what?
Hungry Lion, you are advocating tight integration with the OS because of:
- performance: App Servers and Orbs are written in Java. They are used in
real life situation where performance is an issue. They are pure and not
tight to your OS.
- simplified administration: give me a decent DCOM lookup service when I
don't need to know the host name of a COM server. Hint: there isn't any. If
there was one, it surely isn't in the current MS OSes.
- consistent programming model: huh? You mean consistent component
architecture where one can script components? Like VBA in all your apps
that script COM components? What has it got to do with tight OS
integration? I can do the same thing with JPython and my remote Java
objects.
Don't get me wrong: I program COM components 5 days a week.
> Oh my, another Java zealot telling users to go to hell.
I find the users argument always laughable. Your users don't care whether
you run COM/MTS or CORBA/EJB systems. What matters is the service you
provide them with. Depending on your problem domain, use COM or Java or
CORBA or any combination which will deliver the required service.
But please don't highjack users in this argument.
Yann.
Osvaldo Pinali Doederlein
> Osvaldo Pinali Doederlein wrote in message
> >to do things.
> Where did you read that in?
In your determinate backing of everything MS does, and use of COM as a
counterpoint of almost everything. ;)
> But don't keep feeding the "openness" mantra, when you basically do
> something that favors you in the market that is not really open in
important
> respects.
I don't know about you, but some people only accept as "truly open" what is
GPL'd or an ISO standard. This makes people blind to the reality. Is Win32
more open than Java? Is C++ more open than Java only because it already got
a formal standard, even though it strongly promotes all sorts of vendor
lock-in, and developers have less influence in C++'s evolution than in
Java's?
> >The Java language is not the standard and it is not important. Java
> >bytecode is the standard. It is still much less interoperable than e.g.
> >CORBA, but it's a tradeoff you're free to choose -- you can program to
your
> >favorite interoperable middleware with Java.
> You miss my point. I have no problem with what you are saying about Java
> bytecodes, etc.. It is JavaBeans, Sun's componentized system that I am
> talking about. Confining it to the Java language, IMO, is the offense I
> mentioned.
Beans are defined on the language, but so what? If you support bytecode,
you can access any bean and you can expose your stuff as beans, regardless
of the language you use. To give you one interesting example, I use an
expert system for Java (JESS) which contains a LISP-like language to do the
rule-based code, but from that code I can use beans easily, the system will
use the bean introspection to import beans into the appropriate knowledge
representation. So there is no such a thing as Java language-dependency.
All Java standards are ultimately binary standards. The entire core APIs
are a binary standard.
> Here is an interesting statement from Roger Sessions on the CORBA-COM
> rivalry:
> "I was very involved in the OMG. I worked for IBM at the time, and I was a
> lead architect for one of the CORBA services. I wrote my last book on this
> topic (Object Persistence; Beyond Object-Oriented Databases)."
Roger always forgets to mention that he teamed with MS and began his jihad
of anti-CORBA FUD after he was responsible for the failure of the original
persistent service.
> "The OMG consortium was held together by one common purpose: to stop
> Microsoft. The OMG failed in this effort, and is now limping along with
only
> a few companies making even a minimal profit on CORBA technologies."
This is very funny to read, considering that my company (Visibroker master
in south america and official OMG representative) is doing quite well, and
our clients too, with CORBA technology.
You may interested to know that CORBA is recently beating COM in enterprise
applications -- I posted recently the numbers from Evans Marketing in US and
another from the Japanese market, but I got no replies.
> "Many people accuse Microsoft of meeting any technological threat with an
> embrace and smother strategy. This is certainly not true of Microsoft's
> reaction to CORBA. Despite the millions of dollars that IBM, Sun, and
others
> were spending on CORBA, Microsoft never took CORBA seriously. Microsoft
> joined the OMG, and occasionally sent a representative to OMG meetings,
but
> looked on the whole affair with an air of amusement, as if it was watching
> children at play."
I wonder how much $$$ one gets from MS to publish such a marvelous piece,
showing BillG as some sort of god in the top of his Olympus, and his
competitors as "children at play"? I don't think MS is very "amused" with
the success of CORBA.
> "By 1994, it was clear to Sun and IBM (and later Netscape and the rest of
the
> industry) that OMG was not going to be the knight in shining armor that
> would protect them from the Microsoft dragon. So they started looking for
a
> new protector. And they thought they found it in Java. Soon most of the
> industry had jumped on the Love-Java/Stop-Microsoft bandwagon." See
> http://www.objectwatch.com/issue7.htm"
Java is today a complement rather than replacement (the paragraph is a
subtle sophism), because CORBA doesn't yet cover GUI, didn't have a
component spec till recently, and also the portability of binary code adds
another dimension of interoperability to the whole solution. Of course the
stop-Microsoft was a strong motivation for the adoption of Java in its early
days, but this is not that simple, see my article
"The Microsoft Java Dilemma"
http://student.vub.ac.be/~opinalid/msjava.htm
Jack Richards
Yann wrote in message <37BD2DE1...@ims.ltd.uk>...
>Point taken.
>
>However, why such a thread title then?
The title should be "Java-language-only is going to be the death of Java
Beans."
But that's too dull.
Hey, it's the age of spin.
Jack
>
>Yann.
>
Jack Richards
Osvaldo Pinali Doederlein wrote in message
>From: Jack Richards <jrxxx...@worldnet.com>
>> Osvaldo Pinali Doederlein wrote in message
way
>> >to do things.
>> Where did you read that in?
>
>counterpoint of almost everything. ;)
It would be more correct to say "the backing of everything to empower
consumers, especially knowledge workers via distributed components.
Who had done anything in this area? COM and CORBA, and CORBA has almost no
profile in this area.
Actually, I have followed EJB (and JavaBeans as it evolved into a real
component system), e.g. JavaBeans will be plug-ins for the power user.
>
>> But don't keep feeding the "openness" mantra, when you basically do
>> something that favors you in the market that is not really open in
>important
>> respects.
>
>I don't know about you, but some people only accept as "truly open" what is
>GPL'd or an ISO standard. This makes people blind to the reality. Is
Win32
>more open than Java? Is C++ more open than Java only because it already
got
>a formal standard, even though it strongly promotes all sorts of vendor
>lock-in,
Sounds like Java to me. EJB is known to be plagued by vendor lock in.
The problem is that Sun is not very talented to this point as a software
co., so it sets out a spec and then vendors implement it by adding features
that limit portibility.
Don't be surprised to see COM on all major platforms. The amount of server
software business completely dwarfs the client software business.
The reality is that competition pushes each side to be more open in reality.
Actually, I have always favored the EJB competition.
But someone has to stand up for the knowledge worker on the NOS (used to be
desktop). Java has been a big disappointment there, even if it has been big
money for some people.
>and developers have less influence in C++'s evolution than in
>Java's?
>
>> >The Java language is not the standard and it is not important. Java
>> >bytecode is the standard. It is still much less interoperable than e.g.
>> >CORBA, but it's a tradeoff you're free to choose -- you can program to
>your
>> >favorite interoperable middleware with Java.
>> You miss my point. I have no problem with what you are saying about Java
>> bytecodes, etc.. It is JavaBeans, Sun's componentized system that I am
>> talking about. Confining it to the Java language, IMO, is the offense I
>> mentioned.
>
>Beans are defined on the language, but so what? If you support bytecode,
>you can access any bean and you can expose your stuff as beans, regardless
>of the language you use. To give you one interesting example, I use an
>expert system for Java (JESS) which contains a LISP-like language to do the
>rule-based code, but from that code I can use beans easily, the system will
>use the bean introspection to import beans into the appropriate knowledge
>representation. So there is no such a thing as Java language-dependency.
>All Java standards are ultimately binary standards. The entire core APIs
>are a binary standard.
Yes, and if you code OO just right, you can make components. Your view as to
the Java language dependency is not widely shared. Take a look at
Szyperski's book.
>
>> Here is an interesting statement from Roger Sessions on the CORBA-COM
>> rivalry:
>> "I was very involved in the OMG. I worked for IBM at the time, and I was
a
>> lead architect for one of the CORBA services. I wrote my last book on
this
>> topic (Object Persistence; Beyond Object-Oriented Databases)."
>
>Roger always forgets to mention that he teamed with MS and began his jihad
>of anti-CORBA FUD after he was responsible for the failure of the original
>persistent service.
Have you ever read Roger's columns in ObjectWatch? He has some amusing
stories about going back into these worlds. Actually, the top people hold
seminars and conduct themselves as civilized people. it's only the followers
who . . .
>
>> "The OMG consortium was held together by one common purpose: to stop
>> Microsoft. The OMG failed in this effort, and is now limping along with
>only
>> a few companies making even a minimal profit on CORBA technologies."
>
>This is very funny to read, considering that my company (Visibroker master
>in south america and official OMG representative) is doing quite well, and
>our clients too, with CORBA technology.
>
>You may interested to know that CORBA is recently beating COM in enterprise
>applications -- I posted recently the numbers from Evans Marketing in US
and
>another from the Japanese market, but I got no replies.
>
>> "Many people accuse Microsoft of meeting any technological threat with an
>> embrace and smother strategy. This is certainly not true of Microsoft's
>> reaction to CORBA. Despite the millions of dollars that IBM, Sun, and
>others
>> were spending on CORBA, Microsoft never took CORBA seriously. Microsoft
>> joined the OMG, and occasionally sent a representative to OMG meetings,
>but
>> looked on the whole affair with an air of amusement, as if it was
watching
>> children at play."
>
>I wonder how much $$$ one gets from MS to publish such a marvelous piece,
>showing BillG as some sort of god in the top of his Olympus, and his
>competitors as "children at play"? I don't think MS is very "amused" with
>the success of CORBA.
>
>> "By 1994, it was clear to Sun and IBM (and later Netscape and the rest of
>the
>> industry) that OMG was not going to be the knight in shining armor that
>> would protect them from the Microsoft dragon. So they started looking for
>a
>> new protector. And they thought they found it in Java. Soon most of the
>> industry had jumped on the Love-Java/Stop-Microsoft bandwagon." See
>> http://www.objectwatch.com/issue7.htm"
>
>Java is today a complement rather than replacement (the paragraph is a
>subtle sophism), because CORBA doesn't yet cover GUI, didn't have a
>component spec till recently, and also the portability of binary code adds
>another dimension of interoperability to the whole solution. Of course the
>stop-Microsoft was a strong motivation for the adoption of Java in its
early
>days, but this is not that simple, see my article
>"The Microsoft Java Dilemma"
>http://student.vub.ac.be/~opinalid/msjava.htm
An interesting article. You miss one thing.
Users, with their market power, will control what succeeds both on the
client and increasingly on the server as well, because the first system that
truly empowers the user with plug-in programming (akin to and far beyond
what was realized with Visical)--this functionality will determine what
software system succeeds in the sense of commanding the market.
Now we have the app productivity side and the web distributed side. Users
have been converted basically to consumers on the web distribuited side, not
the producers they have been on the app productivity side.
Now, I'm going to try to avoid client-server because that is mixed today.
With a link, anyone can be server side if they have server software on their
NOS, which is what the modern computer has become. Most of the early
intranets were run off PCs, and this is ubiquitous today.
Right now I can take much of this hard-won app-productivity and convert it
to the web distributed side via an ActiveX component. However, this only
works on an intranet as a practical matter. The Sun-Netscape alliance, in
particular, continues to block this technology from being generally
available, while they hope Java takes over the computing world.
The rest of it is just talk. However, it is anti-competitive and should be
the subject to an antitrust investigation itself. Someday it might if enough
consumers raise their voices. We are starting to see the media take up this
point and that will increase, as the real story of the Netscape/Java history
comes out.
Now if this functionality were to be created or, even better, exceeded by
Java, that would be great.
What's needed are some form of distributed super-apps, allowing plug-in
programming by users. Java talks a little of this talk, but doesn't walk the
walk.
The failure of Java on the client side has been glossed over by Java
advocates. But there is a big penalty to be paid. No matter what the success
on the server side, the fact that Java has basically failed, WHILE ITS
ADHERENTS BLOCK TECHNOLOGY THAT EMPOWERS USERS.
This is an untenable situation. Either produce or stop blocking the road.
My big gripe is that Java is foisted on the industry a dumbed down version
of a componentized system, being language centered to protect Sun
intellectual property. They are even worse than Microsoft because the PC
world has empowered the user, while Sun just blocks the road with the
inferior Java technology.
The industry is not going to centralize around a single language. Those who
continue to promote this idea, because of their investments and own limited
talents will eventually be exposed for what they are--opportunists hiding
behind a veneer of idealism. Does this not sound like a major theme of our
age of spin?
Jack
Terry Sikes
Jack Richards <jrxxx...@worldnet.com> wrote:
>
>D'Arcy Smith wrote in message <37BC92ED...@itools.symantec.com>...
>>Jack Richards wrote:
>>
>>> D'Arcy Smith wrote in message <37BC0F60...@itools.symantec.com>...
>>> >Jack Richards wrote:
Let me open this by saying that you, Jack, are clearly a person with a
closed mind and an obvious agenda. Like JTK, you've been corrected on
many points with verifiable facts, but continue to spount your 'party
line' in defiance of them. This is the definition of 'zealot', IMO.
Many of the folk that you, JTK and your ilk tar with that brush don't
deserve it, since they don't exhibit this behavior. (BTW, would you
mind disclosing your employer?)
[snip]
>>> Since any componentized systems offers services at an any interface, and
>>> is therefore said to be a service extensible model.
>>
>>Can you download a COM thing from a Win32 machine to my Linux machine
>>and have it work?
No comment, Jack?
>>> Why are people skeptical about Jini? Basically, Sun takes these
>>> techologies and bastardizes them to fit the Java platform.
>>
>>Uh... no. Sun, in part, made the Java platform to allow for Jini.
No comment, Jack?
>>> Basically, it's a square Java
>>> peg in a round technolgical hole result. None of this stuff needs Java.
>>> Sometimes it's overkill and sometimes it's underkill, but its the same
>>> balony.
>>
>>Please show me how I can download code from a service and run it onto a
>>client using existing technology. You have no knowledge of the
>>OS/machine type in advance.
>
>You really think this is hot stuff. ANY componentized system that is
>designed for such a scenario would work just fine.
Great. Let us know when you've invented one and gleaned 1% of the
industry support Java enjoys. Its clear COM/DCOM don't fit the bill.
>It's odd that Sun who is the last to develop anything approaching "plug and
>play" on its own computers, which require highly trained professionals to
>operate them, comes up with the "concept," which presupposes a Java-only
>world, whether overkill or underkill via resources, then borrows these
>ideas, and effectively steals the language of components, pretends this is
>wonderful new stuff, and markets it to the media eager for an "electric car"
>story.
I guess you've never installed a Sun workstation. You plug it in,
answer about two questions, and it works.
Regardless, the gist of your argument is that no company should ever
invent something that takes it into new markets...which is patently
absurd. I've come to expect this kind of reasoning from you.
>Sun is actually replacing CORBA with a Java-language-only
>replacement. It's nothing new, despite your swooning.
How can Sun 'replace' CORBA with something that isn't
language-neutral? How do you explain Java-CORBA integration, and
CORBA's ongoing success? It has considerably more mindshare than DCOM
according to recent surveys.
>>> How does it get away with this? Because it has the Java faithful.
>>
>>No - because it works. It is a good useable solution.
No comment, Jack?
>>> The failure of the ABM camp to block the integration of the browser in
>>> the OS is going to mean that users forming powerful markets on the desktop,
>>> including point and click ecommerce solutions, and low-level plug-in
>>> programming to replace the Java-ites.
>>
>>Have your Win32 only world. Unfortunatly for you the world is
>>still larget than Win32.
>
>The Win32 world happens to be the world of the computer user, for the most
>part.
If this is the case, you have no problem. Use ActiveX, IE and Win32
and conquer the world. Just leave the rest of us alone while you
pursue your windmill... :-)
Feel free to ignore Linux and the Mac, even though the userbase of
each is growing rapidly (Linux growth is outpacing NT).
>Your disdain for this world will effectively mean that this computer
>user world will replace half baked technological solutions with real
>technologies from whatever source that increase their collective power.
[staring at this sentence with some concern] Have you sought
professional help?
>Isn't that what the Java-ites are doing? Increasing their worlds at the
>expense of the user?
No. Its by no means a zero-sum game.
>>> The first killer app that allows all the desktop functionality to be
>>> incorporated in a web container, subject to being distributed, will drive
>>> the rest of the market, up through the mid-point of the server markets.
>>
>>The day I use a word processor in a browser somwbody can come
>>and stick the noose around my neck. It is a useless idea.
>
>You don't get it. All the functionality of generations of programmers have
>created for the user will be distributed. All it takes is a first bite of
>this, and the limited functionality of Java re the desktop is history. It is
>simply an underpowered technology.
Underpowered compared with what? Win32, COM and C++?!? Excuse me for
a moment... BWAHAHAHAHAHAHHAHAHAAHHAAHAHAHAHAHAHAHAAHHAHAHAAHAHAHAHA!
There, that was a good laugh. Thanks.
>If Java has actually worked as promised, Netscape would have taken over the
>desktop.
What computer technology has 'worked as promised' on first release?
Windows 1.0? OLE/COM? C++? Nope. Java is much further along _given
its age_ than any of those...sorry to burst your bubble.
Netscape should bear quite a bit of the blame also...their development
resources were(are?) so fragmented its no wonder the Java
implementation sucked rocks.
>Now it's like some additional app. The productivity apps that
>people use can get their web services via the OS. Why leave the app to open
>Netscape?
Er, you don't have to 'leave the app' to open Netscape.
>The point I am making and you are avoiding with your disdain of the user, is
>that the app user who gather web info, massages it in their app somehow,
>cannot then distribute it before of the Netscape-Sun lock on the desktop
>client.
What the heck are you talking about? The app generates HTML then
writes or FTPs it to the Web server... Of course the client is
welcome to use the full power of the Java 2 plugin if desired...with
signed applets pretty much anything is possible. I don't see your
point at all.
Further, if anyone wanted such a thing, it would be trivial to write
an ActiveX plugin for Netscape. In fact, isn't such a thing
available? A quick Google search reveals:
http://www.softseek.com/Programming/ActiveX/Internet_and_Networking/Review_22036_index.html
<sarcasm>
Wow, I can't imagine why Microsoft hasn't woken up and done the same
thing, it would have *freed all those trapped Netscape users*!!!! (Who
presumably are too stupid or obstinate to switch to the one true browser,
IE.)
</sarcasm>
>You cannot publish with an ActiveX control, or otherwise, especially
>with an underpowered applet (Java's trademark) because the COM
>technology is blocked for anti-competitive reasons by the ABM alliance
>of server vendors.
Actually, current Java technology for IE has been blocked by
Microsoft's shortsightedness and allegedly illegal activities. That
is the real problem. No one wants ActiveX on the Internet due to
large file sizes, platform dependence and security issues. Don't
forget all those iMac users out there...
>Forget the whole world. This is the situation that has been enforced on the
>PC desktop by an alien culture, who, like yourself, cares about their own
>selfish interests.
_You_ appear to be the representative of an alien culture. What
planet are you from? Mars? Uranus? Gatesanus? (Sorry, just
couldn't resist that one..feel free to deduct 10 civility points.;)
Terry
--
tsi...@netcom.com
Jack Richards
This is typical of the Java hype. It is the basic "anti" reason for its
existence.
Components provide multi-platform, multi-language functionality. And much
more.
Sun has bastardized JavaBeans to support only Java. I support JavaBeans
without the bastardization.
Java has failed on the desktop.
Users need greater functionality that is distributed as well.
The thin client needs to fatten to support app functionality in the form of
some kind of super-app.
Java on the desktop, with's Sun alliance with Netscape, simply blocks real
technologies.
If you can't do it, don't block what can.
Plug in functionality will be built on top of componentized systems so users
can easily do a great part of the programming, be it in Java or whatever.
This will occur on the client and server via web authoring.
Web authoring will be part of every desktop app. It will put place app
functionality, say financial analysis, on the "server" for distribution,
whether the server is on the
The programmers/companies that supply this functionality will be insanely
rich. The mass market provided by users is the mechanism for this. One
enabling app for users, done by a team of great progammers, is worth 10,000
ordinary programmers.
Java programmers, for the most part, are part of an old fashioned EDP
concept of programming. They are doing today what is equivalent to inserting
formulas in spreadsheets before a team of great programmers created Visical.
Except they are doing it in the web culture in this early days.
The current server world mentality has never known real competition. It is
basically a group of Luddites in a new world of industrialized software who
gamely cheer each other on with the same old, same old mantas.
Real competition is provided by the opportunity to get insanely rich
building apps, now server side apps, that allows markets of millions to do
what is hand coded today. You can count on it and no amount of Java hype
will make any difference.
The motivation to make big money with infrastructure and programs having
great functionality for millions of users is what drives technology.
Component systems are building built in every area.
We can expect plug in modules to shortly follow so knowledge workers in
every area of endeavor can plug in the programming they need without going
to IT.
A good possibility is that the open source movement to contribute greatly
with a new version of a componentized system, just as Linux does currently.
No amount of anti-Win32 mantras will substitute for these basic facts.
Your personal attacks are essentially meaningless against an argument based
on technologies.
It is a climate of intimidation that is used to block discussion in order to
cover up technological failures. You as an individual, of course, are not
that important in the scale of things, but this climate is fostered by
important companies, and others who are heavily invested, who greatly fear
competition.
Employees of these companies, as well as those heavily investing in various
sectors, such as book publishing, seminar training, etc. regularly
contribute to a current climate of intimidation, while spouting a cover
about some MS whatever of old.
They think that their behavior won't be remembered. Their efforts have
contributed to denying computer users greater functionality in multitudes of
professions. Now the pendulum is swinging and the exposes are starting about
the Java hype, one of the all time con games of technology.
The statements made privately, while publicly expousing the hype will be
good material for the legal profession, when many of the Java startups fail
and its litigation time.
Here is the start of the revisionism from the heart of the evil empire (Sun
version) at
http://www.sunworld.com/sunworldonline/swol-08-1999/swol-08-bookshelf.html
"Another revelation in this book answers a question I have always had about
Java. Sun is in some ways the ultimate Silicon Valley techie paradise; it
has a history of technology that sells well despite truly mediocre marketing
and laughable attempts to reach out past the techie community to corporate
executives and other nontechie audiences. How did Sun strike gold with Java,
vaulting it from a decent programming language to the most heavily hyped
technology since the Internet itself? Did some marketing droid within Sun
show unexpected talent and do this ... or was it just luck?
According to Rivlin, it was a combination of two things. First was the idea
of hyping Java as a technological paradigm shift. That came from the noted
techno-futurist George Gilder. Second was Java's attractiveness as an
antidote to Microsoft, which was another story that appealed to the mass
media. In other words, the high public profile and success of Java as much
more than a programming language was due to no one at Sun."
The comedy unfolds.
Jack
Terry Sikes wrote in message <7pkcvr$o...@dfw-ixnews9.ix.netcom.com>...
William Brogden
Matt Kennel
:Your response boils down to: no Win32.
:
:This is typical of the Java hype. It is the basic "anti" reason for its
:existence.
No. The real reason is a deep and consistent object-oriented API.
:Components provide multi-platform, multi-language functionality. And much
:more.
:
:Sun has bastardized JavaBeans to support only Java. I support JavaBeans
:without the bastardization.
Um? JavaBeans is about Java.
:Java has failed on the desktop.
Windows desktop.
:Users need greater functionality that is distributed as well.
:
:The thin client needs to fatten to support app functionality in the form of
:some kind of super-app.
:
:Java on the desktop, with's Sun alliance with Netscape, simply blocks real
:technologies.
:
:If you can't do it, don't block what can.
Blocks? Is this a MS propaganda book?
If somebody writes a deep cross-platform alternative to a crufty Microsoft
object system that's ''blocking real technologies''??
:Plug in functionality will be built on top of componentized systems so users
:can easily do a great part of the programming, be it in Java or whatever.
:This will occur on the client and server via web authoring.
:
:Web authoring will be part of every desktop app. It will put place app
:functionality, say financial analysis, on the "server" for distribution,
:whether the server is on the
Notice how ''web authoring'' is the heralded feature of Office 2000.
:The programmers/companies that supply this functionality will be insanely
:rich. The mass market provided by users is the mechanism for this. One
:enabling app for users, done by a team of great progammers, is worth 10,000
:ordinary programmers.
Netscape Navigator? (well maybe 1.0 was great programming)
:Java programmers, for the most part, are part of an old fashioned EDP
:concept of programming. They are doing today what is equivalent to inserting
:formulas in spreadsheets before a team of great programmers created Visical.
:Except they are doing it in the web culture in this early days.
:
:The current server world mentality has never known real competition. It is
:basically a group of Luddites in a new world of industrialized software who
:gamely cheer each other on with the same old, same old mantas.
:Real competition is provided by the opportunity to get insanely rich
:building apps, now server side apps, that allows markets of millions to do
:what is hand coded today. You can count on it and no amount of Java hype
:will make any difference.
Huh? Programming is programming: you have to be able to think and
properly organize abstractions.
:The motivation to make big money with infrastructure and programs having
:great functionality for millions of users is what drives technology.
:Component systems are building built in every area.
:
:We can expect plug in modules to shortly follow so knowledge workers in
:every area of endeavor can plug in the programming they need without going
:to IT.
Oh, what do you mean? Subroutines? A great invention, agreed.
Calling subroutine libraries 'components' doesn't change the reality
that programming is still hard and calling the pieces different names
doesn't change reality.
That takes hard work. There are sometimes some fundamental core advances,
when some very smart people figure out an automated, general solution for
difficult problems, e.g. register allocation algorithms, garbage collection.
:It is a climate of intimidation that is used to block discussion in order to
:cover up technological failures. You as an individual, of course, are not
:that important in the scale of things, but this climate is fostered by
:important companies, and others who are heavily invested, who greatly fear
:competition.
:
:Employees of these companies, as well as those heavily investing in various
:sectors, such as book publishing, seminar training, etc. regularly
:contribute to a current climate of intimidation, while spouting a cover
:about some MS whatever of old.
Intimidation? Um?
:
:Here is the start of the revisionism from the heart of the evil empire (Sun
:version) at
:http://www.sunworld.com/sunworldonline/swol-08-1999/swol-08-bookshelf.html
:
:"Another revelation in this book answers a question I have always had about
:Java. Sun is in some ways the ultimate Silicon Valley techie paradise; it
:has a history of technology that sells well despite truly mediocre marketing
:and laughable attempts to reach out past the techie community to corporate
:executives and other nontechie audiences. How did Sun strike gold with Java,
:vaulting it from a decent programming language to the most heavily hyped
:technology since the Internet itself? Did some marketing droid within Sun
:show unexpected talent and do this ... or was it just luck?
:
:According to Rivlin, it was a combination of two things. First was the idea
:of hyping Java as a technological paradigm shift. That came from the noted
:techno-futurist George Gilder. Second was Java's attractiveness as an
:antidote to Microsoft, which was another story that appealed to the mass
:media. In other words, the high public profile and success of Java as much
:more than a programming language was due to no one at Sun."
Completely wrong. The real reasons:
1) people were waking up from their hangover and realize that they were
very tired of C++.
2) Java was attached to the web because Andressen thought it was cool and
the web was getting cool.
--
* Matthew B. Kennel
* Institute For Nonlinear Science/University of California, San Diego
I would not, could not SAVE ON PHONE,
I would not, could not BUY YOUR LOAN,
I would not, could not MAKE MONEY FAST,
I would not, could not SEND NO CA$H,
I would not, could not SEE YOUR SITE,
I would not, could not EAT VEG-I-MITE,
I do *not* *like* GREEN CARDS AND SPAM! MAD-I-AM!
D'Arcy Smith
> Yann wrote in message <37BD2DE1...@ims.ltd.uk>...
> >Point taken.
> >However, why such a thread title then?
> The title should be "Java-language-only is going to be the death of Java
> Beans."
What other component models allow you to write components that
will run on different platforms without recompiling?
Serious question.
..darcy
Terry Sikes
Jack Richards <jrxxx...@worldnet.com> wrote:
>Your response boils down to: no Win32.
No, better read it again.
>This is typical of the Java hype. It is the basic "anti" reason for its
>existence.
Wrong twice in a row.
>Components provide multi-platform, multi-language functionality. And much
>more.
OK, fine. Java components are "best of breed".
>Sun has bastardized JavaBeans to support only Java. I support JavaBeans
>without the bastardization.
How can Sun "bastardize" something when it was designed that way in
the first place?
>Java has failed on the desktop.
Java has only just begun to fight!
>Users need greater functionality that is distributed as well.
Which will be provided by Java 2.
>The thin client needs to fatten to support app functionality in the form of
>some kind of super-app.
Dynamic class loading helps ensure that Java apps don't 'fatten' too
much.
>Java on the desktop, with's Sun alliance with Netscape, simply blocks real
>technologies.
I thought you said it had already failed? If so, those other vapor
technologies ought to be going like gangbusters...
>If you can't do it, don't block what can.
How can a 'failed' (your words) product block _anything_?
>Plug in functionality will be built on top of componentized systems so users
>can easily do a great part of the programming, be it in Java or whatever.
>This will occur on the client and server via web authoring.
Yes, I think this will happen with Java.
>Web authoring will be part of every desktop app. It will put place app
>functionality, say financial analysis, on the "server" for distribution,
>whether the server is on the
Not complete, so I won't answer. I addressed a fairly identical point
earlier anyhow.
>The programmers/companies that supply this functionality will be insanely
>rich. The mass market provided by users is the mechanism for this. One
>enabling app for users, done by a team of great progammers, is worth 10,000
>ordinary programmers.
Agreed, and I suspect the next such app will be written largely in
Java. If not, it will still need to be cross-platform to succeed.
>Java programmers, for the most part, are part of an old fashioned EDP
>concept of programming. They are doing today what is equivalent to inserting
>formulas in spreadsheets before a team of great programmers created Visical.
>Except they are doing it in the web culture in this early days.
This paragraph amply illustrates the depth of your incomprehension.
>The current server world mentality has never known real competition. It is
>basically a group of Luddites in a new world of industrialized software who
>gamely cheer each other on with the same old, same old mantas.
Why did Bill Gates say "Java scares the hell out of me."?
>Real competition is provided by the opportunity to get insanely rich
>building apps, now server side apps, that allows markets of millions to do
>what is hand coded today. You can count on it and no amount of Java hype
>will make any difference.
Yes it will, since these apps will likely be written _in_ Java.
>The motivation to make big money with infrastructure and programs having
>great functionality for millions of users is what drives technology.
>Component systems are building built in every area.
More empty words without substance.
>We can expect plug in modules to shortly follow so knowledge workers in
>every area of endeavor can plug in the programming they need without going
>to IT.
Perhaps, and perhaps not. Myself, I'd want experts building my
mission critical systems. Have you ever used a program produced by a
VB novice?
>A good possibility is that the open source movement to contribute greatly
>with a new version of a componentized system, just as Linux does currently.
How does Linux do this currently???
>No amount of anti-Win32 mantras will substitute for these basic facts.
Ah, but an industry moving away from Win32 will. Did you read today
that Compaq laid off over 100 Windows NT/Alpha engineers, and that
NT/Alpha is on life support?
>Your personal attacks are essentially meaningless against an argument based
>on technologies.
Where's the beef? Mostly what I hear from you are vague
generalizations rather than technological substance.
>It is a climate of intimidation that is used to block discussion in order to
>cover up technological failures. You as an individual, of course, are not
>that important in the scale of things, but this climate is fostered by
>important companies, and others who are heavily invested, who greatly fear
>competition.
I assume you're talking about Microsoft and Bill Gates? Microsoft is
clearly the computer industry's most paranoid company, as well as the
least open.
>Employees of these companies, as well as those heavily investing in various
>sectors, such as book publishing, seminar training, etc. regularly
>contribute to a current climate of intimidation, while spouting a cover
>about some MS whatever of old.
I'd prefer to discuss the MS of today - you know, the one that ships
mediocre to horrible products, languages and tools.
>They think that their behavior won't be remembered. Their efforts have
>contributed to denying computer users greater functionality in multitudes of
>professions.
Yes, once again you must be speaking of Microsoft...I'm sure of it!
>Now the pendulum is swinging and the exposes are starting about
>the Java hype, one of the all time con games of technology.
The Windows hype has been far worse for far longer, and has done
immeasurable damage to the industry. As I mentioned below, remember
Windows 1.0?
>The statements made privately, while publicly expousing the hype will be
>good material for the legal profession, when many of the Java startups fail
>and its litigation time.
I'd say that given the history of OS/2 (remember when Microsoft
evangalized it and signed up lots of ISVs then turned around and
dumped it for Windows after those ISVs had a substantial OS/2
investment?) Sun has nothing to worry about, even in the unlikely
event of Java somehow 'failing'.
>Here is the start of the revisionism from the heart of the evil empire (Sun
>version) at
>http://www.sunworld.com/sunworldonline/swol-08-1999/swol-08-bookshelf.html
>
>"Another revelation in this book answers a question I have always had about
>Java. Sun is in some ways the ultimate Silicon Valley techie paradise; it
>has a history of technology that sells well despite truly mediocre marketing
>and laughable attempts to reach out past the techie community to corporate
>executives and other nontechie audiences. How did Sun strike gold with Java,
>vaulting it from a decent programming language to the most heavily hyped
>technology since the Internet itself? Did some marketing droid within Sun
>show unexpected talent and do this ... or was it just luck?
>
>According to Rivlin, it was a combination of two things. First was the idea
>of hyping Java as a technological paradigm shift. That came from the noted
>techno-futurist George Gilder. Second was Java's attractiveness as an
>antidote to Microsoft, which was another story that appealed to the mass
>media. In other words, the high public profile and success of Java as much
>more than a programming language was due to no one at Sun."
Well, it sounds as though Sun has some smart marketing people...which
unfortunately is a necessity in the post-Microsoft computer industry.
>The comedy unfolds.
Yes, your post was quite funny once again. Espousing COM+Win32 as a
better alternative to Java shows comedic genius!!! ;-)
Most of the sane members of the computer industry recognize that its
high time a quality alternative to Microsoft's technologies emerged.
Linux+Java looks like a fine combo to me. BTW, did you note that IBM
just released a supported 1.1.6 JRE for Linux? :-)
By the way, try actually addressing a few of the points I made next
time...
Terry
D'Arcy Smith
> Your response boils down to: no Win32.
> This is typical of the Java hype. It is the basic "anti" reason for its
> existence.
No Java is not anti-anything - except anti-OS/architecture lock.
> Components provide multi-platform, multi-language functionality. And much
> more.
Please name a component system that allows componets to be run on
any architecture.
> Sun has bastardized JavaBeans to support only Java. I support JavaBeans
> without the bastardization.
Kind of hard to "bastardize" JavaBeans into something when it is still
in its first rev. It was/is/always has been a JVM technology. Note
that it is was not/is not/has not been a Java language only technology.
Any class file can be introspected/treated as a Bean/etc... regardless
of the language it was written in. It just has to have been compiled
to the JVM.
> Java on the desktop, with's Sun alliance with Netscape, simply blocks real
> technologies.
Please list the technologies that Java is blocking and how the
technologies are better.
> If you can't do it, don't block what can.
Oh obviuusly your talking about MS not Sun!
> Plug in functionality will be built on top of componentized systems so users
> can easily do a great part of the programming, be it in Java or whatever.
> This will occur on the client and server via web authoring.
And the componet has to run on several architectures. Please list
technologies that allow you to do that.
> Java programmers, for the most part, are part of an old fashioned EDP
> concept of programming. They are doing today what is equivalent to inserting
> formulas in spreadsheets before a team of great programmers created Visical.
> Except they are doing it in the web culture in this early days.
Problem with your logic - Java is not Web centered.
> A good possibility is that the open source movement to contribute greatly
> with a new version of a componentized system, just as Linux does currently.
Ah so COM/CORBA are not the answer.
> No amount of anti-Win32 mantras will substitute for these basic facts.
Where fo you get that Java == anti-Win32?
> It is a climate of intimidation that is used to block discussion in order to
> cover up technological failures.
Your saying that Java is blocking Win32 right? Serious question.
> They think that their behavior won't be remembered. Their efforts have
> contributed to denying computer users greater functionality in multitudes of
> professions.
Wow... total disagreement... but Wow.
MS has kept the computer industry back by at least 10 years... but you
think that the rest of it has been holding MS back? Wow.
> Now the pendulum is swinging and the exposes are starting about
> the Java hype, one of the all time con games of technology.
How is it a con game?
> Here is the start of the revisionism from the heart of the evil empire (Sun
> version) at
> http://www.sunworld.com/sunworldonline/swol-08-1999/swol-08-bookshelf.html
Where is the "rivisionism" in the article?
..darcy
Jack Richards
D'Arcy Smith wrote in message <37BDDF96...@itools.symantec.com>...
The platform has to support the component model. Fortunately with
components, they interop easily and can be "wrapped" with minimal overhead.
The serious question is why don't they.
You have two basic forces. The "server" forces trying to enter the desktop
and the "client" forces trying to enter the server market. Web middleware
has been the battleground.
Both are "alien" to the other. However, it is clear that the PC revolution
has done more for human progress. Empowering users, knowledge workers, home
and small business, and little companies that supply the innovation and
competition.
It is really the Unix/mainframe worlds that are subject to the wrenching
change as this PC revolution has subsumed the low to middle workstation and
server markets. Technology keeps rolling over higher end systems,
fortunately for the average person.
Despite the Gate-as-Rockefeller angst, there is little evidence I've seen of
great innovation that Microsoft has denied the computer world.
Nonetheless, competition on the desktop is a desired goal. But not a
Java/Netscape web middleware technology that thwarts the industry move
toward components, the interchangeable software part, upon which plug-in
user programming can be built.
This is an alien server world culture that is a "Trogon Horse" which
defeats an existing component technology that allows desktop app
functionality to be distributed NOW.
Top it but don't thwart it, without being labeled anti-user. Compete with it
and produce more distributed functionality, but don't give the user
"cocktail party" Java hype.
"Yes, it will be happen in 2002, when Java is mature."
Don't dumb down the desktop to the level of a device, and then preen and
grin like an imp, when the real objective is to avoid PC industry
competition.
Middle tier development opens a server world of functionality, on whatever
platform, simply through web authoring. Interestingly, thwarting the user on
the client has speeding up the penetration there, so the Unix server world
is even more at risk of being displaced by user-friendly technologies.
Just because Microsoft is hated does not excuse palming off second-rate
technologies that WILL NEVER WORK. Java as high level language and VM is
never going to work as a system of interchangeable software parts.
JavaBeans that bastardizes the component concept for the benefit of Sun
delusions of being the next Microsoft also will never work, because, dream
as you will, the computer industry will never organize around a single
language.
When the hype ends, then the proponents look like fools. That is happening
now.
Very few Java products are totally Java, even without the 100% Pure bull.
Let Sun chase windmills in the embedded world, but don't dumb down the
user's computer to the level of a device, and pretend you're saving the
world.
Because of the Sun-Netscape alliance, every user cannot distribute basic
desktop functionality. Java on the desktop basically stands for blocking
progress.
Jack
>
>..darcy
Jack Richards
You are completely right. I'm going to have to go cold turkey.
You know, I just hate hype.
But thanks for reminding me. I look at how many messages I've posted and it
actually shocks me.
Usually the only way is to stop completely.
I did want to get some feedback on these topics though. To test weaknesses
in concepts, etc. The great range of thought gives me pause to think.
I may write an article on "Dumbing the Desktop Down to a Device."
Jack
William Brogden wrote in message <37BDD88E...@bga.com>...
D'Arcy Smith
> D'Arcy Smith wrote in message <37BDDF96...@itools.symantec.com>...
> >Jack Richards wrote:
> >> Yann wrote in message <37BD2DE1...@ims.ltd.uk>...
> >> >Point taken.
> >> >However, why such a thread title then?
> >> The title should be "Java-language-only is going to be the death of Java
> >> Beans."
> >What other component models allow you to write components that
> >will run on different platforms without recompiling?
> >Serious question.
> The platform has to support the component model. Fortunately with
> components, they interop easily and can be "wrapped" with minimal overhead.
That did not answer my question. Try again.
Wrapping is _not_ a solution. You compile your component on Linux.
What can it do on Win32? Not much (actualy nothing).
> The serious question is why don't they.
Because that doesn't solve the problem for downloadable components.
> You have two basic forces. The "server" forces trying to enter the desktop
> and the "client" forces trying to enter the server market. Web middleware
> has been the battleground.
Where did you see me mention middleware, client, or server?
I have component X I want to use it on QNX. It is written in C++ and
compiled for a Mac. What component model helps me with that?
> Nonetheless, competition on the desktop is a desired goal. But not a
> Java/Netscape
WHere the hell did Netscape come into this? Java doesn't need Netscape
(or any browser) to run. Stick to the JVM please. Don't intruduce
other things to muddle the _technical_ discussion.
> This is an alien server world culture that is a "Trogon Horse" which
> defeats an existing component technology that allows desktop app
> functionality to be distributed NOW.
What do you mean when you say "distributed"?
> Don't dumb down the desktop to the level of a device, and then preen and
> grin like an imp, when the real objective is to avoid PC industry
> competition.
That isn't what Java is doing.
> Just because Microsoft is hated does not excuse palming off second-rate
> technologies that WILL NEVER WORK. Java as high level language and VM is
> never going to work as a system of interchangeable software parts.
It already does.
> JavaBeans that bastardizes the component concept for the benefit of Sun
> delusions of being the next Microsoft also will never work, because, dream
> as you will, the computer industry will never organize around a single
> language.
It already is working. And it works in ways that COM and other
MS technologies _CANNOT_ work.
> Because of the Sun-Netscape alliance, every user cannot distribute basic
> desktop functionality. Java on the desktop basically stands for blocking
> progress.
What does a web browser have to do with my desktop?
..darcy
James A. Robertson
> Jack Richards wrote:
>
> > Yann wrote in message <37BD2DE1...@ims.ltd.uk>...
> > >Point taken.
>
> > >However, why such a thread title then?
>
> > The title should be "Java-language-only is going to be the death of Java
> > Beans."
>
> What other component models allow you to write components that
> will run on different platforms without recompiling?
>
Using Distributed Smalltalk components with VisualWorks, for one. Using
IBM's VisualAge distribution model with their Smalltalk, for another. Using
OpenTalk with VisualWorks 5i if you don't care about interop with other
languages, for a third.
Using the Java component model, can I modify components on the fly as they
are running ?
>
> Serious question.
>
> ..darcy
James A. Robertson
> In article <7pkku4$6rb$1...@nntp1.atl.mindspring.net>,
> Jack Richards <jrxxx...@worldnet.com> wrote:
>
>
> >Components provide multi-platform, multi-language functionality. And much
> >more.
>
> OK, fine. Java components are "best of breed".
>
Compared to what ?
>
James A. Robertson
> Jack Richards wrote:
>
> > Your response boils down to: no Win32.
>
> > This is typical of the Java hype. It is the basic "anti" reason for its
> > existence.
>
> No Java is not anti-anything - except anti-OS/architecture lock.
>
> > Components provide multi-platform, multi-language functionality. And much
> > more.
>
> Please name a component system that allows componets to be run on
> any architecture.
>
Use VisualWorks parcels and they can be loaded dynamically (over a wire or from a
file system). VW supports 12 platforms - not 'everything', but just about
everything that's in serious usage.
>
> And the componet has to run on several architectures. Please list
> technologies that allow you to do that.
>
Any VM based system that has a portable design. VisualWorks for one, Squeak for
another. If you want open source, Squeak is the way to go. Unlike Java, no one
has iron fisted control over it.
> ..darcy
D'Arcy Smith
> D'Arcy Smith wrote:
> > Jack Richards wrote:
> > > >Point taken.
> > > >However, why such a thread title then?
> > > The title should be "Java-language-only is going to be the death of Java
> > > Beans."
> > What other component models allow you to write components that
> > will run on different platforms without recompiling?
> Using Distributed Smalltalk components with VisualWorks, for one. Using
> IBM's VisualAge distribution model with their Smalltalk, for another. Using
> OpenTalk with VisualWorks 5i if you don't care about interop with other
> languages, for a third.
Cool.
> Using the Java component model, can I modify components on the fly as they
> are running ?
What type of modifications do you want to make?
..darcy
James A. Robertson
> Cool.
>
> > Using the Java component model, can I modify components on the fly as they
> > are running ?
>
> What type of modifications do you want to make?
>
Add behaviors to a running system. Add (or delete) attributes to take account of
changed conditions - all while the server is still up and running.
This works in Smalltalk
>
> ..darcy
D'Arcy Smith
> D'Arcy Smith wrote:
> > Cool.
> > > Using the Java component model, can I modify components on the fly as they
> > > are running ?
> > What type of modifications do you want to make?
> Add behaviors to a running system. Add (or delete) attributes to take account of
> changed conditions - all while the server is still up and running.
Yes ... but it is far from as simple as Smalltalk. Basicaly you
use a ClassLoader to load new .class files into the VM. There is
no (simple) way to convert existing instances into the new version
of the class though.
It takes some planning but it can be done. Smallkalk easily wins
in that area... and that is one of the things I would love to see
change in Java.
Note that the Symantec VM in VCafe is tweaked to allow true class
re-loading with conversion of existing instances into the new class.
I can't say how it is done... but it is possible in a modified VM
to get behaviour similar to Smalltalk without so much work... just
not the standard behaviour unfortunatly.
..darcy
John Hughes
D'Arcy Smith <nos...@itools.symantec.com> wrote in message
news:37BC92ED...@itools.symantec.com...
> Jack Richards wrote:
>
> > D'Arcy Smith wrote in message <37BC0F60...@itools.symantec.com>...
> > >Jack Richards wrote:
>
it
> > were
> > >> open) is touted as the great solution, when it is actually corrupting
the
> > >> world of real technologies.
> > >
> > >> Take Jini. When you look at it, it has a bunch of cute marketing
names
> > for
> > >> such things as introspection (discovery; compare IUnknown in COM),
> > >> repository (JavaSpaces), etc.
> > >
> > >COM runs well on my Linux box... especialy all those COM components
> > >that are compiled to Win32 (like all of them).
>
> > >COM is not a distribited service framework... Jini is.
>
is
> > therefore said to be a service extensible model.
>
> Can you download a COM thing from a Win32 machine to my Linux machine
> and
> have it work?
>
Java = working......hmmmmmm
Some reading for you. http://www.mainsoft.com/
>
>
> > Why are people skeptical about Jini? Basically, Sun takes these
techologies
> > and bastardizes them to fit the Java platform.
>
> Uh... no. Sun, in part, made the Java platform to allow for Jini.
>
>
>
> > Basically, it's a square Java
> > peg in a round technolgical hole result. None of this stuff needs Java.
> > Sometimes it's overkill and sometimes it's underkill, but its the same
> > balony.
>
> Please show me how I can download code from a service and run it onto a
> client using existing technology. You have no knowledge of the
> OS/machine type in advance.
>
>
> > How does it get away with this? Because it has the Java faithful.
>
> No - because it works. It is a good useable solution.
>
>
> > The failure of the ABM camp to block the integration of the browser in
the
> > OS is going to mean that users forming powerful markets on the desktop,
> > including point and click ecommerce solutions, and low-level plug-in
> > programming to replace the Java-ites.
>
> Have your Win32 only world. Unfortunatly for you the world is
> still larget than Win32.
>
>
> > The first killer app that allows all the desktop functionality to be
> > incorporated in a web container, subject to being distributed, will
drive
> > the rest of the market, up through the mid-point of the server markets.
>
> The day I use a word processor in a browser somwbody can come
> and stick the noose around my neck. It is a useless idea.
>
>
> > The whole of components is to stabilize the code, so plug-in software
can be
> > built on top of it. Once this is done, then "apps" to manipulate the
> > components can be built that are simply point and click.
>
> Right and COM does't give me that. It gives me that on Win32... too
> bad I am not always oon Win32.
>
> ---
>
> So you still know nothing about Java or Jini obviously.
>
> ..darcy
John Hughes
<bill@x> wrote in message news:7pj3ev$1v...@drn.newsguy.com...
> In article <7piiv9$irc$1...@nntp5.atl.mindspring.net>, "Jack says...
> >
>
> >
> >The Win32 world happens to be the world of the computer user, for the
most
> >part.
>
>
> 75% of the world data is held on IBM mainframes.
> close to 90% of the world network and backend processing is done
> on non windows PC's.
>
> not mean the world runs on Win32.
>
> Not only is win32 is a closed and runs on one platform only. But the Java
API
> is considered by many a much more powerfull and easier to use for
> enterprise computing.
>
The Java API is more powerful? How?
> Java is not perfect, but it is the best thing we have right now.
> Win32 and that cryptic modified C++ and COM+ is a lousy alternative. If
you
> choose that, go for it. Be married to Bill gate for the rest of your
> life.
Your jealous arent you!
The rest of us want open, industry wide API and a platform that
> is available everyone. My Java applications run on Linux, windows,
> solaris, VMS and AS400 unchanged, and I like that, and nothing in
> the world will make me go back to win32 or C or C++. They will
> have to drag me screaming and shouting to get me touch that rubbish
> again.
>
> Bill
>
>
John Hughes
Yann <ylel...@ims.ltd.uk> wrote in message
news:37BD4B46...@ims.ltd.uk...
> hungr...@my-deja.com wrote:
>
> > The tight integration COM has with its underlying operating environment
> > will result in better performance, simplified administration, and a
> > consistent programming model. These aren't luxuries, but critical to
the
> > success of componentized systems.
>
> "will"?
> Is that a snapshot from a marketing brochure or what?
>
> Hungry Lion, you are advocating tight integration with the OS because of:
> - performance: App Servers and Orbs are written in Java. They are used in
> real life situation where performance is an issue. They are pure and not
Thats what Sun says. You have any performance benchmarks or Case studies
detailing the solution. And I dont mean Suns usual 'so and so uses Java!'
case studies.
> - simplified administration: give me a decent DCOM lookup service when I
> don't need to know the host name of a COM server. Hint: there isn't any.
If
> there was one, it surely isn't in the current MS OSes.
What problems does this cause?
> - consistent programming model: huh? You mean consistent component
> architecture where one can script components? Like VBA in all your apps
> that script COM components? What has it got to do with tight OS
> integration? I can do the same thing with JPython and my remote Java
> objects.
>
> Don't get me wrong: I program COM components 5 days a week.
>
> > Oh my, another Java zealot telling users to go to hell.
>
> I find the users argument always laughable. Your users don't care whether
> you run COM/MTS or CORBA/EJB systems. What matters is the service you
> provide them with. Depending on your problem domain, use COM or Java or
> CORBA or any combination which will deliver the required service.
> But please don't highjack users in this argument.
>
> Yann.
>
Have you delivered a distributed component based Java system?
John Hughes
Terry Sikes <tsi...@netcom.com> wrote in message
news:7pkcvr$o...@dfw-ixnews9.ix.netcom.com...
> In article <7piiv9$irc$1...@nntp5.atl.mindspring.net>,
> Jack Richards <jrxxx...@worldnet.com> wrote:
> >
> >D'Arcy Smith wrote in message <37BC92ED...@itools.symantec.com>...
> >>Jack Richards wrote:
> >>
> >>> D'Arcy Smith wrote in message
<37BC0F60...@itools.symantec.com>...
> >>> >Jack Richards wrote:
>
> Let me open this by saying that you, Jack, are clearly a person with a
> closed mind and an obvious agenda. Like JTK, you've been corrected on
> many points with verifiable facts, but continue to spount your 'party
> line' in defiance of them. This is the definition of 'zealot', IMO.
> Many of the folk that you, JTK and your ilk tar with that brush don't
> deserve it, since they don't exhibit this behavior. (BTW, would you
> mind disclosing your employer?)
>
close mind because he has a different point of view from you? I for one
think he has some very interesting points. Im sure Jack can prove his
beliefs. So far Ive yet to see anything in Java thats any good.
> [snip]
>
> >>> Since any componentized systems offers services at an any interface,
and
> >>> is therefore said to be a service extensible model.
> >>
> >>Can you download a COM thing from a Win32 machine to my Linux machine
> >>and have it work?
>
> No comment, Jack?
Not for Linux though. The said they were having problems with it.
>
> >>> Why are people skeptical about Jini? Basically, Sun takes these
> >>> techologies and bastardizes them to fit the Java platform.
> >>
> >>Uh... no. Sun, in part, made the Java platform to allow for Jini.
>
> No comment, Jack?
>
Jini sounds like another Sun NC or JavaOS. Where is the implemention going
to come from? Suns hype just doesnt work. I like the concept of Jini but why
are Sun trying to make everything language dependant? Not good.
John Hughes
William Brogden <wbro...@bga.com> wrote in message
news:37BDD88E...@bga.com...
> You must get paid by the word.
KO'd in the first round!
What dont you agree with William? Please explain....
Shame IBM had to get involved with Sun.
John Hughes
James A. Robertson <jar...@home.com> wrote in message
news:37BE1836...@home.com...
> Terry Sikes wrote:
>
> > In article <7pkku4$6rb$1...@nntp1.atl.mindspring.net>,
> > Jack Richards <jrxxx...@worldnet.com> wrote:
> >
> >
> > >Components provide multi-platform, multi-language functionality. And
much
> > >more.
> >
> > OK, fine. Java components are "best of breed".
> >
>
>
> >
>
He read it on Suns site.
Osvaldo Pinali Doederlein
> Jack Richards <jrxxx...@worldnet.com> wrote:
> >The Win32 world happens to be the world of the computer user, for the
most
> >part.
> each is growing rapidly (Linux growth is outpacing NT).
Look at this extremely interesting poll:
http://www.borland.com/linux/survey/
The really fun part is that most of the respondents (24 thousand) are
clearly Windows developers -- more: Delphi developers -- not the kind of
people you associate with Linux hacking. It shows in the questions about
current practices, and also in some clueless answers about Linux
technologies (e.g. the inconsistency of KDE/GNOME and GTK+/Qt numbers). And
this people want to go to Linux; they have customers demanding Linux
support; they think that source code compatibility Win32/Linux is very
important (hint, hint); they are willing to pay for commercial development
tools on Linux and to put up with closed-source compilers although they
otherwise want support for OSS and plan to do some of it -- among other
findings.
Seems like a lot of people develop Windows apps because they have no option,
and not because they think Windows is wonderful or all they customers do.
And tools are a big part of it. There are hordes of developers who don't
feel like moving from their comfortable Windows RAD tools like Delphi or VB
to Linux (gcc + emacs + POSIX libs) or who don't have resources to do Linux
versions if the port is a major rewrite.
Osvaldo Pinali Doederlein
> Java programmers, for the most part, are part of an old fashioned EDP
> concept of programming. They are doing today what is equivalent to
inserting
> formulas in spreadsheets before a team of great programmers created
Visical.
> Except they are doing it in the web culture in this early days.
This paragraph is clear idiocy, of course. I see too much innovation in the
Javaland to make any sense of your "old fashioned", other than the product
of some Reality Distortion Field.
> We can expect plug in modules to shortly follow so knowledge workers in
> every area of endeavor can plug in the programming they need without going
> to IT.
> A good possibility is that the open source movement to contribute greatly
> with a new version of a componentized system, just as Linux does
currently.
Blah blah plug-in modules blah blah Linux components blah blah whatever.
You are basically saying "my vaporware will shortly follow and everything
else is a dead horse". People have been talking about software-IC heavens
where final users easily build applications and are free from the evil IT
dept., if this is what you mean, since Brad Cox and other pioneers, and the
whole idea is certainly good but it's been kinda hard to implement and it's
not going to happen any time soon. And it is certainly not because "Java is
blocking it", what a joke.
> It is a climate of intimidation that is used to block discussion in order
to
> cover up technological failures. You as an individual, of course, are not
> that important in the scale of things, but this climate is fostered by
> important companies, and others who are heavily invested, who greatly fear
> competition.
The company who greatly fears competition more than anybody else is well
known and its HQ is in Redmond. If this is not true, please explain why
this company continually (and ilegally) seizes its existing monopolies to
establish new ones. Case in point: if MS doesn't fear competition, why the
hell would they give away MSIE while their competitor was a commercial
product with a price tag, AND distribute it in the OEM channel, AND bully
OEM vendors to not preinstall Netscape, AND later purposedly plan to corrupt
Java's portability (point proven by their own internal documents) so Win32
would not have a competitor... not to mention that by your own logic, MS
should have adopted Netscape's plug-in architecture and scripts, instead of
offering ActiveX browser plug-ins or VBScript -- doesn't this qualify as
"blocking" the work of people with "real technology"?
Greg Chien
news:7plvkd$1r8e$1...@quince.news.easynet.net...
> D'Arcy Smith <nos...@itools.symantec.com> wrote in message
> news:37BC92ED...@itools.symantec.com...
machine
> > and have it work?
>
Understand mainsoft can do most of the commercial Un*xen. Do they
have Linux support? Where do you see it?
Also, I suppose D'Arcy wants universal byte-code that can be
downloaded and executed on the fly on various platforms. I am not
sure how this ideal concept meets reality, though.
Greg
Osvaldo Pinali Doederlein
> D'Arcy Smith wrote in message <37BDDF96...@itools.symantec.com>...
> "Yes, it will be happen in 2002, when Java is mature."
It is already mature enough for lots of applications, and it is getting
better very fast. Remembver that evolution is a continuum and it applies to
any technology -- Microsoft is also always selling us things that "will be
available" when "COM+ is ready" or when "Windows2000 is ready" or when
"Fahrenheit is ready" or when "NT 64-bit is ready" and so on, and you don't
claim that Windows is "immature" because it will be even better in 2002.
> Just because Microsoft is hated does not excuse palming off second-rate
> technologies that WILL NEVER WORK. Java as high level language and VM is
> never going to work as a system of interchangeable software parts.
One of the reasons why many people hate MS is second-rate technology. And
your idea of a VM making more difficult to have interchangeable software
parts of something to utterly puzzle any reasonable mind -- do you mean that
monolithic pieces of platform-specific code are any easier to mix and match?
> When the hype ends, then the proponents look like fools. That is happening
> now.
YOU are looking like fool today, and like somebody completely uninformed
about reality or in MS's pocket.
> Very few Java products are totally Java, even without the 100% Pure bull.
Didn't your mother teach you to not LIE? Please show me ANYTHING backing
this claim.
> Because of the Sun-Netscape alliance, every user cannot distribute basic
> desktop functionality. Java on the desktop basically stands for blocking
> progress.
Looks like a good scapegoat for Wintel's failure to deliver the stuff you
want to do, doesn't it? keep dreaming.
Osvaldo Pinali Doederlein
> Note that the Symantec VM in VCafe is tweaked to allow true class
> re-loading with conversion of existing instances into the new class.
> I can't say how it is done... but it is possible in a modified VM
> to get behaviour similar to Smalltalk without so much work... just
> not the standard behaviour unfortunatly.
True class unloading / reloading / mutation of objects is easy to implement
in Java VMs, you basically need to discard jitted code and, like in full GC,
stop everything to scan your heap and change the affected objects.
Greg Chien
message news:004401beebca$24252d90$4505...@mde.emn.fr...
[snip]
> Case in point: if MS doesn't fear competition, why the
> hell would they give away MSIE while their competitor was a
commercial
[snip]
If Sun doesn't fear competition, why the hell would they give away
Java/JDK while their competitor (Sun/C++) was a commercial product
with a price tag (pretty high, too), ...
Well, it's called marketing strategy. ;-)
Greg
Osvaldo Pinali Doederlein
> > Case in point: if MS doesn't fear competition, why the
> > hell would they give away MSIE while their competitor was a
> commercial
> > product with a price tag, ...
> [snip]
> If Sun doesn't fear competition, why the hell would they give away
> Java/JDK while their competitor (Sun/C++) was a commercial product
> with a price tag (pretty high, too), ...
> Well, it's called marketing strategy. ;-)
No, it is called bending the truth. Free compilers / SDKs are the norm and
the only sensible strategy for any new platform or non-proprietary language.
You could get a free C or C++ compiler when these languages were introduced,
you can download all Microsoft's SDKs for free. Java was not a competitor
of Sun's or anybody's C++ compilers by any stretch of imagination, at least
in the first years. And remember that Sun gives away a rude command-line
development kit, not really an equivalent to Visual C++ or VisualWorks or
Delphi. Add to this the hefty fee needed to license Java, and it's much
less "giveaway" than some competition (unlike when MS licensed MFC to
everybody for a song, so they could beat the much superior OWL).
James A. Robertson
power in the areas I consider critical for server deployment. So you have to get a
custom VM from a specific vendor. This introduces vendor lock - which is not
necessarily a bad thing.
Now, having been reduced to going to a sole source for this custom VM, you will still
end up being about 1/3 as productive as a Smalltalk developer, and still be every bit
as vendor dependent.
D'Arcy Smith wrote:
> "James A. Robertson" wrote:
>
> > D'Arcy Smith wrote:
>
> > Add behaviors to a running system. Add (or delete) attributes to take account of
> > changed conditions - all while the server is still up and running.
>
> Yes ... but it is far from as simple as Smalltalk. Basicaly you
> use a ClassLoader to load new .class files into the VM. There is
> no (simple) way to convert existing instances into the new version
> of the class though.
>
> It takes some planning but it can be done. Smallkalk easily wins
> in that area... and that is one of the things I would love to see
> change in Java.
>
> Note that the Symantec VM in VCafe is tweaked to allow true class
> re-loading with conversion of existing instances into the new class.
> I can't say how it is done... but it is possible in a modified VM
> to get behaviour similar to Smalltalk without so much work... just
> not the standard behaviour unfortunatly.
>
> ..darcy
D'Arcy Smith
> From: Jack Richards <jrxxx...@worldnet.com>
> > D'Arcy Smith wrote in message <37BDDF96...@itools.symantec.com>...
> > "Yes, it will be happen in 2002, when Java is mature."
Just be clear that I did not write the above.
..darcy
D'Arcy Smith
> Osvaldo Pinali Doederlein <osv...@visionnaire.com.br> wrote in
> message news:004401beebca$24252d90$4505...@mde.emn.fr...
> [snip]
> > Case in point: if MS doesn't fear competition, why the
> > hell would they give away MSIE while their competitor was a
> commercial
> > product with a price tag, ...
> [snip]
> If Sun doesn't fear competition, why the hell would they give away
> Java/JDK while their competitor (Sun/C++) was a commercial product
> with a price tag (pretty high, too), ...
> Well, it's called marketing strategy. ;-)
I never paid for GCC...
..darcy
D'Arcy Smith
> This is pretty much what I figured would happen - the standard Java VM gives you no
> power in the areas I consider critical for server deployment. So you have to get a
> custom VM from a specific vendor. This introduces vendor lock - which is not
> necessarily a bad thing.
Not sure what happens outside of the VCafe environment.
> Now, having been reduced to going to a sole source for this custom VM, you will still
> end up being about 1/3 as productive as a Smalltalk developer, and still be every bit
> as vendor dependent.
I don't frequently change code on the fly... I beleive that that is
mainly done for long running systems... right? With some for-thought
and some work you can get most (and maybe all) of what you have in
Smalltalk without modifying the VM at all.
..darcy
D'Arcy Smith
I don;t need a Universal VM - the JVM is universal enough.
Other languages have been compiled to Java bytecodes.
The requirements are that I have a component on machine X (say
Win32 - Intel) and I want to downloadn/run it on machine Y
(say Sun Solaris). Having COD/DCOM etc... on Solaris will
not help me do that.
This is why MS technologies do not compete with Java ones. They
are totaly different at a fundamental level.
..darcy
D'Arcy Smith
> D'Arcy Smith <nos...@itools.symantec.com> wrote in message
> news:37BC92ED...@itools.symantec.com...
> > Can you download a COM thing from a Win32 machine to my Linux machine
> > and
> > have it work?
> Java = working......hmmmmmm
For me it is.
> Some reading for you. http://www.mainsoft.com/
WHat was I supposed to get out of that? That they are bringing MS
technolologies like DCOM to Unix? Fine but how does that help me
_run_ my stuff compiled on Win32 on my Linux machine? I have to
re-compile. Not an option.
Next time please trim your post rather than wast everyones bandwidth.
You been taking posting lessons feom JTK?
..darcy
John Hughes
Greg Chien <p...@wwa.com> wrote in message
news:GKwv3.120$6M3....@ord-read.news.verio.net...
> John Hughes <ma...@easynet.co.uk> wrote in message
> news:7plvkd$1r8e$1...@quince.news.easynet.net...
> > news:37BC92ED...@itools.symantec.com...
> machine
> > > and have it work?
> >
> have Linux support? Where do you see it?
>
> Also, I suppose D'Arcy wants universal byte-code that can be
> downloaded and executed on the fly on various platforms. I am not
> sure how this ideal concept meets reality, though.
>
>
>
My mistake.
I spoke with Mainsoft a few weeks ago and they said they were having
problems developing it on Linux.
John Hughes
> From: Jack Richards <jrxxx...@worldnet.com>
> > D'Arcy Smith wrote in message <37BDDF96...@itools.symantec.com>...
> > "Yes, it will be happen in 2002, when Java is mature."
>
> better very fast. Remembver that evolution is a continuum and it applies
to
> any technology -- Microsoft is also always selling us things that "will be
> available" when "COM+ is ready" or when "Windows2000 is ready" or when
> "Fahrenheit is ready" or when "NT 64-bit is ready" and so on, and you
don't
> claim that Windows is "immature" because it will be even better in 2002.
>
Java is immature.
What are these applications that its mature enough for?
> > Just because Microsoft is hated does not excuse palming off second-rate
> > technologies that WILL NEVER WORK. Java as high level language and VM is
> > never going to work as a system of interchangeable software parts.
>
> One of the reasons why many people hate MS is second-rate technology. And
> your idea of a VM making more difficult to have interchangeable software
> parts of something to utterly puzzle any reasonable mind -- do you mean
that
> monolithic pieces of platform-specific code are any easier to mix and
match?
>
> > When the hype ends, then the proponents look like fools. That is
happening
> > now.
>
> YOU are looking like fool today, and like somebody completely uninformed
> about reality or in MS's pocket.
>
You look like your in Suns pocket. At least people using MS technology can
actually implement scalable, distributed systems.....TODAY!
> > Very few Java products are totally Java, even without the 100% Pure
bull.
>
> Didn't your mother teach you to not LIE? Please show me ANYTHING backing
> this claim.
>
> > Because of the Sun-Netscape alliance, every user cannot distribute basic
> > desktop functionality. Java on the desktop basically stands for blocking
> > progress.
>
> Looks like a good scapegoat for Wintel's failure to deliver the stuff you
> want to do, doesn't it? keep dreaming.
>
What has Java delivered so far? You act like its proved something.
John Hughes
> John Hughes wrote:
> > D'Arcy Smith <nos...@itools.symantec.com> wrote in message
> > news:37BC92ED...@itools.symantec.com...
>
> > > Can you download a COM thing from a Win32 machine to my Linux machine
> > > and
> > > have it work?
>
>
> For me it is.
>
>
> technolologies like DCOM to Unix? Fine but how does that help me
> _run_ my stuff compiled on Win32 on my Linux machine? I have to
> re-compile. Not an option.
>
> Next time please trim your post rather than wast everyones bandwidth.
> You been taking posting lessons feom JTK?
>
> ..darcy
re-compile is not an option? Why not?
Whats going to happen when you guys actually start deploying distributed
components using EJB's and they are all tied in with the applications
servers they were written for?
What happens when the discributed applications architecture has to change
and you wish to move the middle-tier objects onto the client?
<rest ignored>
Terry Sikes
Jack Richards <jrxxx...@worldnet.com> wrote:
All right, Jack. I don't have time to reply to all your posts, you
are another prolific one like JustTrolling. I'll just address this
one specific point because it most clearly reveals your misconceptions
about Java.
>Don't dumb down the desktop to the level of a device, and then preen and
>grin like an imp, when the real objective is to avoid PC industry
>competition.
How is Java 'dumbing down the desktop' in any way? Java 2 provides
APIs for advanced 2D rendering, 3D rendering including a scene graph,
voice recognition, encryption and multimedia. Many of these vastly
exceed the functionality available on 'raw' Win32 - and among other
things illustrate the value added to the base platform by a quality
Java implementation. Far from being a 'lowest common denominator'
approach, they represent a rich opportunity for innovative solutions.
It is equally important that these advanced frameworks are generally
well designed and used with an elegant (at least compared to C++),
object oriented language.
Sun makes the Java source code available through a quite liberal
license. There's certainly nothing comparable in the Microsoft camp.
The practice of 'open source' is certainly one of the strengths of the
Unix tradition.
Sun is moving forward with the ECMA/ISO standardization process. I
fail to see this happening with any of the Microsoft technologies you
tout.
Finally, Java is fully extensible either in 100% Java or with just
about any other language/native feature either through JNI or, with
the native compilers, even more straightforward linking techniques.
There are absolutely no restrictions to doing either, or producing
competing, clean room implementations like gcj (www.cygnus.com).
I hope this cleared quite a few things up for you.
Terry
--
tsi...@netcom.com
Terry Sikes
>> Jack Richards <jrxxx...@worldnet.com> wrote:
>>
>>
>> >Components provide multi-platform, multi-language functionality. And much
>> >more.
>>
>> OK, fine. Java components are "best of breed".
>
>Compared to what ?
Specifically, ActiveX controls. The Bean model is quite strong,
and much more elegant.
Terry
--
tsi...@netcom.com
Terry Sikes
>think he has some very interesting points. Im sure Jack can prove his
>beliefs. So far Ive yet to see anything in Java thats any good.
Closed mind because he fails to acknowlege many of the obvious
successes of Java, such as application servers. These application
servers really don't have significant competition from Redmond since
they run on hardware and operating systems with true 24/7 uptime and
massive scalability. NT's deficiencies in these areas are well known.
The advantages of servlets and JSP over the competition for Web
development are obvious to any impartial observer.
In other areas, such as embedded systems and desktop development, Java
is attractive due to its elegance as much as anything. It offers the
full power of OO with nearly the simplicity of C, much more portably
(even if fully compiled). I'm confident that performance problems and
memory consumption (neither of which are a _major_ issue now in many
problem domains) will be solved in the near term.
As important as all the above, if not more so, is the diverse support
Java has in the computer industry as well as other industries such
as consumer electronics. Look at the companies that comprise Symbian,
which have bought into Java in a big way.
>> [snip]
>>
>> >>> Since any componentized systems offers services at an any interface,
>and
>> >>> is therefore said to be a service extensible model.
>> >>
>> >>Can you download a COM thing from a Win32 machine to my Linux machine
>> >>and have it work?
>>
>> No comment, Jack?
>
>Not for Linux though. The said they were having problems with it.
>
>http://www.mainsoft.com/
What's "not for Linux though"? Who's "they"? Parse failure...
>> >>> Why are people skeptical about Jini? Basically, Sun takes these
>> >>> techologies and bastardizes them to fit the Java platform.
>> >>
>> >>Uh... no. Sun, in part, made the Java platform to allow for Jini.
>>
>> No comment, Jack?
>>
>Jini sounds like another Sun NC or JavaOS. Where is the implemention going
>to come from? Suns hype just doesnt work. I like the concept of Jini but why
>are Sun trying to make everything language dependant? Not good.
You don't know much about Jini. The main thust of Jini is to have
"hot plug and play" or "spontaneous networking" for devices connected
for any type of data transfer medium. Examples would be cell phones,
digital cameras, PDAs, stereo receivers, TVs. Again, see Symbian, or
the many companies that've committed to Jini.
BTW the "language dependent" part of Jini is there to allow devices to
carry around their own drivers...what would you propose instead?
One good sign is the consistent failure of WinCE to gain any market
share. The industry is waking up.
Terry
--
tsi...@netcom.com
John Hughes
> >close mind because he has a different point of view from you? I for one
> >think he has some very interesting points. Im sure Jack can prove his
> >beliefs. So far Ive yet to see anything in Java thats any good.
>
> successes of Java, such as application servers. These application
> servers really don't have significant competition from Redmond since
> they run on hardware and operating systems with true 24/7 uptime and
> massive scalability. NT's deficiencies in these areas are well known.
>
Java has not had success in the application server market. Its only just
beginning for Java. Where are the examples of Java application server
success? Show me a case study detailing Java being used for a real world
application?
Oh Oracle built an application server which supports Java. This doesnt make
it a success. (this goes for the other application server vendors)
> The advantages of servlets and JSP over the competition for Web
> development are obvious to any impartial observer.
>
> In other areas, such as embedded systems and desktop development, Java
> is attractive due to its elegance as much as anything. It offers the
> full power of OO with nearly the simplicity of C, much more portably
> (even if fully compiled). I'm confident that performance problems and
> memory consumption (neither of which are a _major_ issue now in many
> problem domains) will be solved in the near term.
>
> As important as all the above, if not more so, is the diverse support
> Java has in the computer industry as well as other industries such
> as consumer electronics. Look at the companies that comprise Symbian,
> which have bought into Java in a big way.
>
> >> [snip]
> >>
> >> >>> Since any componentized systems offers services at an any
interface,
> >and
> >> >>> is therefore said to be a service extensible model.
> >> >>
> >> >>Can you download a COM thing from a Win32 machine to my Linux machine
> >> >>and have it work?
> >>
> >> No comment, Jack?
> >
> >Not for Linux though. The said they were having problems with it.
> >
> >http://www.mainsoft.com/
>
> What's "not for Linux though"? Who's "they"? Parse failure...
>
> >> >>> Why are people skeptical about Jini? Basically, Sun takes these
> >> >>> techologies and bastardizes them to fit the Java platform.
> >> >>
> >> >>Uh... no. Sun, in part, made the Java platform to allow for Jini.
> >>
> >> No comment, Jack?
> >>
> >Jini sounds like another Sun NC or JavaOS. Where is the implemention
going
> >to come from? Suns hype just doesnt work. I like the concept of Jini but
why
> >are Sun trying to make everything language dependant? Not good.
>
John Hughes
When are you going to publish the 'real world' version?
Terry Sikes <tsi...@netcom.com> wrote in message
news:7pmtel$k...@dfw-ixnews9.ix.netcom.com...
> In article <7pl0om$kqg$1...@nntp1.atl.mindspring.net>,
bob@nospam
>
>
>As important as all the above, if not more so, is the diverse support
>Java has in the computer industry as well as other industries such
>as consumer electronics. Look at the companies that comprise Symbian,
>which have bought into Java in a big way.
>
Good points by tsikes.
In addition to the above, I would also add the powerfull and
comprehnsive standard API's that Java comes with.
This is for me the major selling point for Java. Ready to use,
standard packages to do almost anything, from telephony, to data
base access, to 2&3D, to media, to TV, phone cards, Jini, EJB,
serlvts, JSP, and I go on and on.
Java as a language is a fine language also. I would probably pick
Ada as a better language on pure technical issues. But a language
alone is not sufficient, one needs industry support, a large well
documented API's etc.. And when you add all these togther, Java
is the best thing we have right now, and it keeps improving.
Bob
Osvaldo Pinali Doederlein
> > > "Yes, it will be happen in 2002, when Java is mature."
> > It is already mature enough for lots of applications, and it is getting
> > better very fast. Remembver that evolution is a continuum and it
applies
> What are these applications that its mature enough for?
I guess you can get a list in Javasoft's website, or look in places like
jars.com. You may not be very impressed because one point where Java is
weak today is doing GUI-heavy, shrinkwrap applications. I guess the number
of in-house or server-side projects is much larger. But there are
interesting examples, like ObjectDomain. And you can't really belive that
the Java market is so hot, with a formidable wealth of development tools,
middlewares etc., for years, while nobody is having success with it. The
"let's kick Microsoft" sentiments may be responsible for many people buying
their first Java IDE and starting their first Java project, but if failures
multiplicate, the fun is not going to go too far.
Java is good enough for a Trouble Ticket management app, compliant with the
NMF502 spec, that we are shipping. It's a distributed system using CORBA.
I picked Java+Swing for the clients, which are pretty big and complex; and I
picked C++ for the servers because Java was not good enough when the project
started, but it was a bad long-term decision because it is today. Of
course, we have the freedom to install whatever JVM we want on clients, so
we can install the best available.
Java is mature enough so I could do a state of the art reverse engineering
framework as part of my MSc thesis. You can check this one in my site.
Less than 2 man-month do do it, and without ever using a debugger. This
experience evangelized me about how Java is great more than any Sun
marketing. I am just as good, and more experienced, in C++, and I'm sure
I'd need much more than 2X the effort to do it in C++. The program is so
heavy (due to what it does, not to Java) that I used it as a benchmark for
an article on Java performance... the memory mgmt alone would overload me in
C++. No going back to the past. Performance is excellent (comparing the
common features with a native, commercial tool, like Rose -- same level).
Code talks... bullshit walks.
> > YOU are looking like fool today, and like somebody completely uninformed
> > about reality or in MS's pocket.
> You look like your in Suns pocket. At least people using MS technology can
> actually implement scalable, distributed systems.....TODAY!
Our company knows a lot about distrbuted, scalable systems, given that we
are pioneers and leaders in CORBA technology in our continent. We have
clients whose projects include telecom, security, healthcare, scientific
research etc. Your mantra that "only MS delivers today" is something to
laugh.
A+
Osvaldo
----------------------------------------------------------------
Check these great projects!!
- Framework for Reverse Engineering / Design Patterns
- UML 1.3 metamodel Implementation for Java
- The Free, Open and Pure Java Shell
http://student.vub.ac.be/~opinalid/
----------------------------------------------------------------
Osvaldo Pinali Doederlein osv...@visionnaire.com.br
MSc student, Developer, Java & CORBA evangelist, Rainmaker, etc.
Terry Sikes
John Hughes <ma...@easynet.co.uk> wrote:
>Did you copy this from Suns site?
Nope. :-)
>When are you going to publish the 'real world' version?
Nice job at addressing the facts I presented.
Content: 0
Thanks for quoting the entire message though, so folks can see
how empty your response was.
Terry
--
tsikesentcom
bill@x
>
>Java has not had success in the application server market. Its only just
>beginning for Java. Where are the examples of Java application server
>success? Show me a case study detailing Java being used for a real world
>application?
>
NetDynamics is an application server written completely in Java.
(except for couple of small components writte in C++).
Are you saying that the NetDynamics application server is not real?
Bill.
Terry Sikes
Gemstone is another example of a top-rated Java application server. I
believe it won a PC Week shoot out not too long ago.
I'm sure the IBM offerings in this area are strong as well.
Terry
--
tsi...@netcom.com
Jack Richards
something about all the time I'm wasting doing these posts.
You and others have read into my posts things that I haven't said.
My comments about the historic failure of high level languages, even with
VMs, to be used successful in a cross platform manner, is not an original
thought on my part. In fact, it is why the industry has gone to
componentized systems like CORBA, COM and even JavaBeans.
If you don't agree, it's not going to change the large parts of the industry
that do agree.
Saying the above is not the same thing as saying Java is not a welcome
addition to the marketplace of languages. Nor is it to discount the current
success Java is having as a language of choice for business logic in the
middle tier, irrespective of its long term success, which only time will
tell, because life is full of change, especially away from language fads.
To describe a language as a fad is not to say the ones that have been fads
are not useful languages. Fads are those that were "overpromising." Many are
very useful. They simply are considered to be the vehicle to save the world
from whatever any longer.
However, Java has been a failure on the desktop. This is not my opinion, but
the considered opinion of a growing consensus in the industry. You may
disagree on either of these points, but you will hardly change many minds at
this point.
This too is in itself not a cause for concern because technologies don't
match up to expectations in certain areas all the time. This does not limit
what they may do in other areas.
But Java, or more specifically the Java hype, has an additional role it
plays. That role is its ABM role. It blocks MS as part of the ABM camp. This
in itself is not a problem either because companies butt heads all the time,
and MS needs the competition on the desktop.
The problem is the effect of Netscape's use of Java as a excuse for blocking
desktop component technologies, and what that does to desktop users.
Java's history as a device techology was the reason it could be adapted to
the thin client model that is represented in Netscape's browser technology.
And because, after this, Sun has created a component technology that is
similarly disfunctional or limited because it Java only. In particular,
because of the limitations of Java in the GUI area, a component model that
supported VisualBasic, used by millions of programmers on the desktop, would
have been expected, absence the ABM mentality driving its development.
When I say disfunctional, I mean disfuntional compared to the rich
functionality of the desktop app (often created in heavy lifting languages
like C++ and Smalltalk). So the user now cannot distribute app functionality
because Netscape, a Unix world company, has come onto the desktop, became
the market leader, and has used this dominant position to keep the desktop
in thrall to its Unix world vision of a thin client (what I call dumbed down
to the level of a device. Or, as Marc A. says, a "terminal").
Netscape could support the underlying desktop COM technology if it was not
in an alliance with the ABM camp. In fact, Sun bought Netscape simply to
ensure this device-thin client status would remain as it is, so Sun could
sell servers to service these thin clients and promote its device-based Java
platform as the "exlusive" computer web technology.
I call that a "proprietary" language platform, because even it it's based on
open standards, it is still "proprietary" in the sense that it is
anti-competitive.
To use the openness and standards mantra to enforce an exclusive technology
is a form of corruption of their basic purpose. When a market leader
establishes a standard, as often happens throughout the economy, there is no
problem if it is useful to others as well.
The problem with proprietary is when it is used in an exclusive sense. This
means that it becomes anti-competitive. The Java-language only platform,
when it is enforced on the desktop by Netscape's market leadership position,
in alliance with Sun, to block the very underlying desktop technology it is
on, which is now componentized and offers much greater distributed
functionality to the desktop than Java, is anti-competitive. Therefore, I
think it's fair to describe it as proprietary.
Knowledge workers on the desktop are blocked from distributing web pages,
via a technology, i.e. components, made for distributing these pages, simply
because Netscape, and Sun, are arrogant to believe they can get away with
this when the Java hype dies.
The rub will come as web functionality is built into every desktop app, and
distributing this is blocked. Your pretence that you can't understand the
points I'm making or that I'm just rabidly anti-Java is ridiculous. But then
I've been called a Linux zealot as well, which is also misleading, although
I do believe Linux is the real thing and Java is largely hype in terms of
promises made.
My arguments have little effect. It's what users will do, not what I say.
Desktop users are the most spoiled by great programming and great prices,
etc. It is foolish to expect they will be happy with the situation I've
described.
It is ok to be anti-Microsoft, but not so in a way that makes you
anti-desktop.
Jack
Terry Sikes wrote in message <7pmtel$k...@dfw-ixnews9.ix.netcom.com>...
John Hughes
An empty response for an empty message.
John Hughes
OK, so what is the Java technology you are using to build scalable CORBA
distributed objects? How do you do things like database pooling? Are you
using an application server?
Osvaldo Pinali Doederlein
> My comments about the historic failure of high level languages, even with
> VMs, to be used successful in a cross platform manner, is not an original
> thought on my part. In fact, it is why the industry has gone to
> componentized systems like CORBA, COM and even JavaBeans.
Giving the existence of Java, the point of CORBA et al. is that Java will
not be the ideal platform to implement everything. But the cross-platform
feature of Java works pretty well yes sir. Most of the problems I see in
WORA have nothing to do with Java's technical merits: Some pople don't like
when their menus don't use the correct font, or some other detail, and
consider Java GUI "substandard" even if it's more powerful; The latest and
greatest JVMs, top JIT compilers etc. are not yet available on all
platforms; There may be different bugs in different implementations and you
need to work around them, etc. OTOH, concerns like performance of
CPU-intensive code are gone, as well as lack of tools or libraries for
"serious work". So it only keeps getting better and better. You can't
expect all problems to be solved at once, even if Sun's marketing would tell
you that JDK1.0.2 was already perfect.
Your "historical failure" argument is moot because Java doesn't try to do
cross platform with a language. It does it with the bytecode. You insist
in your argument of treating Java as a language. Maybe it's too much
Microsoft marketing -- "just a language".
> To describe a language as a fad is not to say the ones that have been fads
> are not useful languages. Fads are those that were "overpromising." Many
are
> very useful. They simply are considered to be the vehicle to save the
world
> from whatever any longer.
Agreed. But unfortunately, massive marketing hype seems to be the only
option to popularize something new when the market is tightly monopolized by
the previous thing -- C++ in this case. Look at how Smalltalk or Eiffel are
doing with their great technologies and no good marketing / strategy.
> However, Java has been a failure on the desktop. This is not my opinion,
but
> the considered opinion of a growing consensus in the industry. You may
> disagree on either of these points, but you will hardly change many minds
at
> this point.
Java has yet to HAPPEN in the desktop. It didn't even begin. We just moved
between multiple phases where different problems would exist; at Java 1.0
times, the whole platform was just a toy and no wonder Corel failed doing
something bigger than a spinning mug on it. At Java 1.1, the AWT would mean
mediocre and buggy GUIs, and Java's core was missing a lot of important
enterprise-critical functionality, so it was only good for dumb and simple
clients, and it was not good for servers because it still lacked good
performance in addition to the libraries. Today, the performance and the
big-iron APIs are mostly in place, so Java is growing a lot in the server;
but the desktop equation is only half-solved because we have an extremely
powerful GUI, Swing, but its performance still sucks and integration with
host environments is also a concern, as well as deployment in the Applet
fashion. If you can't predict that these problems will also be addressed,
you are the one who's out of reality.
> But Java, or more specifically the Java hype, has an additional role it
> plays. That role is its ABM role. It blocks MS as part of the ABM camp.
This
> in itself is not a problem either because companies butt heads all the
time,
> and MS needs the competition on the desktop.
By the way, I don't remember reading your definition of ABM? It's a new TLA
for me.
> Java's history as a device techology was the reason it could be adapted to
> the thin client model that is represented in Netscape's browser
technology.
> And because, after this, Sun has created a component technology that is
> similarly disfunctional or limited because it Java only. In particular,
> because of the limitations of Java in the GUI area, a component model that
> supported VisualBasic, used by millions of programmers on the desktop,
would
> have been expected, absence the ABM mentality driving its development.
This is so ridiculous that justifies people asking how much does BillG pay
you for each word here. If you really think that Java adopting VB's
component model, i.e. ActiveX, as its component model, makes ANY sense, I
guess you don't know the first thing about either technology. (I know, in
YOUR perfect world we would have ports of COM in all platforms and use
something like WFC on all platforms.) But why not the opposite? Contrary
to your totally broken understanding, JavaBeans are not language dependent;
you can embed a JavaBeans into a VB app or a Word document. It's easy,
there is a bridge to do this, it lets COM clients see the bean as a COM
control. Of course you pay the hit of running the bridge including a JVM.
But if you adopt COM, you pay the hit of running COM, which may be
particularly expensive on non-Windows platforms. THERE IS NO DIFFERENCE AT
ALL, BOTH ARE BINARY STANDARDS. The Java VM also exposes a lot of standard
native interfaces (of course Microsoft tried very hard to spoil that) and
application can use them to access beans without any COM bridge; so you can
develop a C++ library that makes JavaBeans look like MFC objects or whatever
you like -- this is called language bindings, and all binary standards need
them. CORBA needs client stubs or DII code; COM needs significant lot of
work for languages without direct vtable support and other features, like
the BSTR type (so much for language-neutral); and Java needs a layer that
talks to the JVM and accesses the beans. What is the fscking difference? I
tell you -- Microsoft and COM and VB were available before, and they were
the de facto standard, so they shouldn't be challenged because they are a
holy monopoly. And they were even ubiquitous, running on so many platforms
whose names begin with "Windows"!.
In order to further clarify: a non-binary, language-based component standard
is something that induces specific language syntax or object code on the
client. For example, a conventional C++ library forces you to link against
its .o files, which are not standard even in the same platform (every
compiler has different name mangling and many dirty details, and they may
use different binary formats like COFF or ELF). A templatized C++ is even
worse, there's no way to use it from another language because the components
are in header files that can't even be compiled to stand-alone libraries, so
a binary standard cannot be defined. A really proprietary thing, like a
PowerBuilder DataWindow, is even worse, it's tool-dependent if there are no
open specs and hooks to access the component from the outside. THESE are
non-binary software components.
> Knowledge workers on the desktop are blocked from distributing web pages,
> via a technology, i.e. components, made for distributing these pages,
simply
> because Netscape, and Sun, are arrogant to believe they can get away with
> this when the Java hype dies.
You keep crying that Netscape and Java block some technology that you want,
when this technology doesn't seem to be provided by anybody else. What the
hell you're talking about, anyway, by "workers distributing web pages" I
guess you want that all people publish things, so remember that Netscape was
first offering HTML editing in the browser. And when you say "components",
remember that Java was available before ActiveX components in MSIE. And
ActiveX didn't fail to "catch on" for deployment of web apps because Sun or
Netscape blocked it. It failed because it sucks, it's unsafe and
unportable, and harder to develop too.
> It is ok to be anti-Microsoft, but not so in a way that makes you
> anti-desktop.
It is also ok to be pro-Microsoft. For the record, I love NT, MSIE, Office
and other stuff that the more radical people regard as abominations. I am a
honored Unix Hater(TM). I would use lots of MS stuff even if I had a real
choice. But it's not ok to be pro-MS in a way that's blatantly incompatible
with evidence or logic, and particularly with both.
What is your problem anyway? From your rantings about JavaBeans and wish to
have COM as the component model, I guess you invested much time or money in
VisualJ++/WFC and you feel sick that most developers correctly decided to
ignore and bury that crap?
A+
Osvaldo
Greg Chien
> Greg Chien wrote:
>
> > If Sun doesn't fear competition, why the hell would they give away
> > Java/JDK while their competitor (Sun/C++) was a commercial product
> > with a price tag (pretty high, too), ...
>
How many C++ commercial software are compiled and supported with GCC?
If you want to use GCC as an example, then where is Java GPL? Or, if
you want to limit to academic research projects? It's certainly
perfect for academia.
Greg
Osvaldo Pinali Doederlein
> OK, so what is the Java technology you are using to build scalable CORBA
> distributed objects? How do you do things like database pooling? Are you
> using an application server?
The project I mention uses C++ in the servers, it uses RogueWave libs to go
to the db. I have no experience with application servers; I have been out
of the company for an entire year (back to school) and I know that they're
now playing with that and other things like ITS (transaction server)...
actually I will be back in two weeks, and catch up with some "real world"
things that I have been tracking mostly by press releases and specification
docs :)
Greg Chien
> No, it is called bending the truth. Free compilers / SDKs are the
norm and
> the only sensible strategy for any new platform or non-proprietary
language.
What are the truth? See below.
> You could get a free C or C++ compiler when these languages were
introduced,
> you can download all Microsoft's SDKs for free.
Where did you get a free C or C++ compiler when they were introduced?
Sun sold their C and C++ compilers with "hefty" prices from day one.
I remember DECUS/C with which one need to be a member (or at least run
a PDP machine). My first Aztec C (on Apple ][+) or Lattice C on PC
were never free. Perhaps, you were talking about Small-C, but it is
small and interpretive ;-) Most people here seem to think of GCC, but
it was much later than the commercial compilers. For C++, one had to
pay pretty steep prices for a C++ "preprocessor" on any platform
(before Gnu people added C++ to GCC, of course).
> Java was not a competitor
> of Sun's or anybody's C++ compilers by any stretch of imagination,
at least
> in the first years.
It was for tiny embedded devices, I assume.
> And remember that Sun gives away a rude command-line
> development kit, not really an equivalent to Visual C++ or
VisualWorks or
> Delphi. Add to this the hefty fee needed to license Java, and it's
much
> less "giveaway" than some competition (unlike when MS licensed MFC
to
> everybody for a song, so they could beat the much superior OWL).
Any way, go to http://www.sun.com/workshop/pricelist/ to see how other
languages are priced and think what you would do if Java is one of
them on the list. I suppose Borland's big mistake was not to make OWL
running on Un*x; otherwise, MFC would never have a chance.
Greg
John Hughes
Look forward to it.....
Jack Richards
Netscape, and others of course, must be given great credit for its original
browser.
It greatly increased the power of computer users, what I call the desktop,
to "communicate" by going on the web in many millions.
It was the first user friendly, successful web middleware client. Java
playing no part in this revolution.
Now it would have been better for the user on the desktop if Netscape had
been successful in developing an "OS on an OS," in the sense of a web
middleware client that had provided web services to desktop apps (the
central benfit of legions of programmers in liberating the EDP world with
the personal computer to the benefit of many millions of professionals in
their careers by providing tools such as CAD for designers, etc.).
You see, if Netscape had provided a real competitor to the desktop OS, that
ADDED to the functionality of desktop apps by helping distribute what users
will be doing in every desktop app, i.e. save some work to a web page in the
app that incorporates the underlying app functionality and thus easily
distributing it.
Now you and the other posters seem to have to have a blind eye here. Because
this kind of points up the dismal failure of Java after all the promises
about distributed applets which would be part of a new generation of better
apps, presumably important ones, such as a stock analysis done via a related
app that could distruibute the app as a applet of some sort.
Whatever. Had this happened it would have been great. This would have been
competition that increased the functionality of the desktop. Possibly it
would have displaced MS. Regardless, it would have been good for the
consumer.
Instead, Netscape, a Unix world company, adopted Java and did not adopt a
competent component model, or any other technology that offerred web
services to app, let alone to allow the user to distribute existing app
functionality.
Netscape appears to be going to do so now, only because MS is doing it by
integrating the browser in the OS. In fact, contrary to your statements
about COM, Netscape uses some COM components to do some of its browser
functionality. So COM could easily have been supported by Netscape, since
its browser was built on Win32, etc., APIs.
Now Java is simply a blocking device that works because Netscape's browser
was the market leader, and set the standards, including JavaScript, etc. If
the only browser were MS's, then the blocking would not be possible.
I say this is anticompetitive, and therefore represents a proprietary
language platfrom scheme.
Your response is basically to say that all kinds of interoperability are
possible with binary standards, etc.
The point is that you can't use COM because Netscape does not support it. It
has achieved the Unix server world dream. It has come onto the desktop
disguished as a friend to the desktop, when actually it is helping Sun, a
Unix world company, to dumb down the desktop to the level of a device.
All kinds of excuses are made for this. A favorite one is security problems,
but no one really has security problems using MS's browser, which of course
supports component technologies.
So Java on the desktop basically represents, in important respects, an alien
anti-desktop culture that has succeeded in taking the user, in as far as
user-friendly distribution of app functionality is concerned, back to the
days of EDP.
Congratualtions!
Jack
EDP=Electronic Data Processing
AMB=Anybody But Microsoft
Osvaldo Pinali Doederlein wrote in message
<002901beec92$09753850$4505...@mde.emn.fr>...
Mark Smith
>Some pople don't like
>when their menus don't use the correct font, or some other detail, and
>consider Java GUI "substandard" even if it's more powerful;
Swing's mediocre emulation of the native environment may be a mere
"detail" in certain contexts where cross-platform support is the top
priority. For many desktop applications attempting to compete with
native versions where cross-platform isn't the top priority, it will
certainly be a handicap. And how is a Java GUI "more powerful" from
the user's perspective?
--
-M- markj...@my-dejanews.com
Osvaldo Pinali Doederlein
> Instead, Netscape, a Unix world company, adopted Java and did not adopt a
> competent component model, or any other technology that offerred web
> services to app, let alone to allow the user to distribute existing app
> functionality.
> Netscape appears to be going to do so now, only because MS is doing it by
> integrating the browser in the OS. In fact, contrary to your statements
> about COM, Netscape uses some COM components to do some of its browser
> functionality. So COM could easily have been supported by Netscape, since
> its browser was built on Win32, etc., APIs.
You have a good point here. And that was a major problem in Netscape's
competition to MSIE: their architecture sucked horribly, from 1.0 up to 4.61
it's a monster-sized monolithical app, plugins notwithstanding. The problem
is that Netscape supports lots of platforms and they made the decision of
having as much stuff as possible in a single code base; I guess this is
because they wouldn't have the development resources to maintain major
platform-specific branches. And you surely gotta rewrite most of an app
like that if you're going to model it around something like COM.
If you track the Mozilla project, you'll see that they dumped the old code
in the trash bin, and decided to have a fully componentized architecture
(like IE has since 3.0). They're doing an open-source clone of COM, called
XPCOM (cross-platform COM), I think it's not complete but good enough for
what they're doing. They will use other things, including Java and advanced
scripts/XML/etc. to implement some parts of the product, too.
(Too bad that OpenDoc flopped, btw.)
> Now Java is simply a blocking device that works because Netscape's browser
> was the market leader, and set the standards, including JavaScript, etc.
If
> the only browser were MS's, then the blocking would not be possible.
I guess you know, or you should know, that Javascript is completely
unrelated to Java. Anyway, your blocking conspiracy still doesn't work,
because 1) MSIE has 3X of Netscape's market share today, and it still
doesn't offer anything revolutionary, and the ActiveX stuff that it offers
doesn't qualify; 2) Netscape supports ActiveX with a bridge, it's called
NCompass and it is quite old; 3) Browser-specific technology was never a
problem for many people, and particularly for intranets, even when the only
supported browser was not predominant: look at the adoptions of DHTML, CSS,
XML. Your whole theory that some wonderful MS technology didn't change the
world because Netscape didn't support it is a big joke.
> The point is that you can't use COM because Netscape does not support it.
It
Yes it does.
> All kinds of excuses are made for this. A favorite one is security
problems,
> but no one really has security problems using MS's browser, which of
course
> supports component technologies.
Get real. People have security problems even with Office macros. The only
reason why ActiveX did not spawn a similar wave of viruses and other attacks
is that it failed to be popular enough. Almost all MSIE users only install
a minimal set (most often an empty set) of ActiveX components other than the
ones included in IE's official installation. If some day it becomes common
practice to hit pages with new ActiveX applets every day and install them,
everybody will learn the hard way if those cute certificates are good
enough. EXE files attached in email require at least as much stupidity or
laziness from the users to infect them, and STILL it happens often enough to
be a problem; ditto for macro viruses (remember Melissa bringing servers
down?). It is hopeless.
ActiveX is unsafe even if you check the certificates, because they install
native code in your PC and even if this native code is not malicious, it may
have bugs. As soon as a virus programmer discovers that some popular
ActiveX thing has a bug that can be exploited, e.g. with controlled crashes
to inject code, it's going to be a disaster. When you have a single
Internet app like MSIE, it's bad enough if it has a serious bug (like dozens
of bugs already uncovered and patched on it) because it's only on app,
albeit a big one, and you can hope MS will have combed all security bugs
after some maturation (idem for operating systems, and Internet-enabled apps
like Office). Now, imagine a brave new world (YOUR utopia) whene millions
of little ActiveX components are in the Web. Do you have any hope that all
those components will have no bugs that can be used for security attacks?
Now, compared that to Java. Sure the JVM can have security bugs (as it had
in the past a few times). But the JVM is a single program -- in the worst
case, a handful of programs due to ports and multiple vendors. If you run
1000 Java applets, you are still running only one native program which is
the JVM. But if you run 1000 ActiveX applets, you are running 1000
different programs, EACH ONE IS A NATIVE PROGRAM WITH UNRESTRICTED ACCESS TO
YOUR SYSTEM and a SINGLE ONE with a security bug will mean PARTY TIME FOR
VIRUS.
hungr...@my-deja.com
"John Hughes" <ma...@easynet.co.uk> wrote:
>
> D'Arcy Smith <nos...@itools.symantec.com> wrote in message
> > > > Jack Richards wrote:
> >
> > > > Can you download a COM thing from a Win32 machine to my Linux
machine
> > > > and
> > > > have it work?
> >
> >
> > For me it is.
> >
> >
> > > Some reading for you. http://www.mainsoft.com/
> >
> > WHat was I supposed to get out of that? That they are bringing MS
> > technolologies like DCOM to Unix? Fine but how does that help me
> > _run_ my stuff compiled on Win32 on my Linux machine? I have to
> > re-compile. Not an option.
> >
> > Next time please trim your post rather than wast everyones
bandwidth.
> > You been taking posting lessons feom JTK?
> >
> > ..darcy
>
> re-compile is not an option? Why not?
>
> Whats going to happen when you guys actually start deploying
distributed
> components using EJB's and they are all tied in with the applications
> servers they were written for?
>
Yes, thats where the giant java scam unravels. They are not interested
in enabling distributed computing based on component technology
(ushering in the age of industrialization for software). They wan't
you to believe they are doing that but failures like JavaOS for
Business prove otherwise. For crying out loud they can't even get
their act together on a universal java client. Why anyone believes it
will happen on their bread and butter turf of servers is beyond me.
Sun, Oracle, etc. are really mainframers at heart and
McNealy's "utility model" for network computing is nothing but a return
to the old days of attaching a underpowered terminal (JavaStation) to a
powerful proprietary host. Log in and the meter starts running and
thats what McNealy wants.
Sun is really in big trouble. They are rapidly becoming irrelevant in
the software picture with the power brokers behind Linux taking over.
Thats why they are interested in buying StarOffice.
> What happens when the discributed applications architecture has to
change
> and you wish to move the middle-tier objects onto the client?
>
> <rest ignored>
>
>
Osvaldo Pinali Doederlein
It is more precise to say that a Swing GUI is potentially more powerful --
because many complex things are very easy to do. Sophisticated controls
like trees and tables are piece of cake to use and customize. Simple UIs
requiring no real programming -- i.e. you drag a bunch of controles from
your RAD tool's palette to the form, wire events to methods and code
associated actions -- are equally easy, or they depend on the environment
rather than on the language/library. And I could mention some unique
features of Java, like layout managers; it's a joke how some native apps
don't let the user resize dialogs that should obviously allow it, or do the
resize but don't redistribute space in any reasonable way (look Rational
Rose 98i, and then look the pure java ObjectDomain, for an apples-to-apples
comparison).
Jack Richards
Osvaldo Pinali Doederlein wrote in message
Obviously you don't understand the use of the JavaScript example.The example
was used to show how JavaScript became the universal scripting standard on
the web, because Netscape was able to establish it as the market leader. The
same thing is true of Java v. COM.
You talk about market share now, but don't you understand the role of market
leaders in setting defacto standards? Once the standard is set, it is almost
impossible to change, because anyone with a significant market share can
block the change but not the original standard because everyone is using it.
Where you develop such a pompous style for stating the obvious?
I hate to SHOUT the OBVIOUS, but you may someday be a PROFESSOR IN
PROGRAMMING. Yes, no doubt "programming" nursery school children TO USE THE
POTTY.
Why not sing to them, being you are no doubt a GENIUS in many areas, not
just SHOUTING?
Osvaldi: "All togther now children, do the JAVA in the SANDBOX, that's the
proper place."
Let me get this straight: you're saying that Netscape supports COM and
therefore ActiveX, which is a part of COM?
Let's take this "roundabout" means of support or compatibility approach. By
the same standards of "loose" support, one could say that COM is
multiplatform. Then a poster cites the fact that so and so company is doing
a porting operation or whatever.
Do you think that you have the emotional ability to provide the same
standard when gaguing this kinds of "roundabout" means of support?
But we live in the age of spin, and consistency is not much of an
intellectual value anymore. It's more important to be part of the cause.
Jack
Dimitri I. Rakitine
: Jack Richards wrote:
:> Yann wrote in message <37BD2DE1...@ims.ltd.uk>...
:> >Point taken.
:> >However, why such a thread title then?
:> The title should be "Java-language-only is going to be the death of Java
:> Beans."
: What other component models allow you to write components that
: will run on different platforms without recompiling?
Visual Basic ;-)
: Serious question.
: ..darcy
cm...@singapore.sun.com
<snip>
JR> The problem is the effect of Netscape's use of Java as a excuse for blocking
JR> desktop component technologies, and what that does to desktop users.
"Blocking" - if you do not support a technology then you are effectively
blocking that technology. Funny definition.
<snip>
JR> Knowledge workers on the desktop are blocked from distributing web pages,
JR> via a technology, i.e. components, made for distributing these pages, simply
JR> because Netscape, and Sun, are arrogant to believe they can get away with
JR> this when the Java hype dies.
How did component gets involve in _DISTRIBUTING_ web pages?
Chuk - not speaking for SUN
cm...@singapore.sun.com
GC> Where did you get a free C or C++ compiler when they were introduced?
GC> Sun sold their C and C++ compilers with "hefty" prices from day one.
GC> I remember DECUS/C with which one need to be a member (or at least run
GC> a PDP machine). My first Aztec C (on Apple ][+) or Lattice C on PC
GC> were never free. Perhaps, you were talking about Small-C, but it is
GC> small and interpretive ;-) Most people here seem to think of GCC, but
GC> it was much later than the commercial compilers. For C++, one had to
GC> pay pretty steep prices for a C++ "preprocessor" on any platform
GC> (before Gnu people added C++ to GCC, of course).
The Sun compilers are actually called Workshop. They are just not merely
compilers. Workshop includes stuff like debugger, optimizers, dead lock
detection tool etc. All of these runs under XWin. Giving away these would be
akin to MS giving away VC++ IDE.
Remember what happend when Sun announced that they wanted to charge people
for using HotSpot?
John Hughes
Osvaldo Pinali Doederlein <osv...@visionnaire.com.br> wrote in message
news:001701beed1e$13857df0$4505...@mde.emn.fr...
Sophisticated controls like trees and tables? hahahaha
Shows you how powerful Swing is if you think thats sophisticated. Have you
ever used Delphi or VB?
John Hughes
<cm...@singapore.sun.com> wrote in message
news:7pqv8v$o6n$1...@new-usenet.uk.sun.com...
Giving away workshop would not be similar to giving away MS C++ as C++ users
actually use it. How many Java developers have chosen Workshop as their tool
of choice. ??!!
From what I have read the best tools are : VCafe, Visual Age, JBuilder and
some other tool called 'Kawa'??
Greg Chien
message news:001701beed1e$13857df0$4505...@mde.emn.fr...
> features of Java, like layout managers; it's a joke how some native
apps
> don't let the user resize dialogs that should obviously allow it, or
do the
> resize but don't redistribute space in any reasonable way (look
Rational
> Rose 98i, and then look the pure java ObjectDomain, for an
apples-to-apples
> comparison).
I am still waiting for ObjectDomain to run on Java 2 (version 1.2.x)
with standard Swing so that I can do an apples-to-apples comparison.
Let's see how fast it goes. For now, the question is: why is it so
hard to port from Java 1.1.x to 1.2.x?
Greg
Greg Chien
merely
> compilers. Workshop includes stuff like debugger, optimizers, dead
lock
> detection tool etc. All of these runs under XWin. Giving away these
would be
> akin to MS giving away VC++ IDE.
If I recall correctly, there were times when Sun's C/C++ compilers
(high price) were sold separately from Workshop (even higher price).
People (like me) were using make and dbg (xdb, whatever, can't
remember the real name for now) to create X Window applications (XSun,
to be more exact). Perhaps, it is the reason why people are still
using vi/emacs, make and dbg to develop programs on Un*x platforms
today. Bundling the compiler into Workshop is the current trend or it
cannot compete with development environment on other platforms.
> Remember what happend when Sun announced that they wanted to charge
people
> for using HotSpot?
And, what happens now? You skipped commenting the price list from
Sun's website. Why isn't Java/JavaWorkshop on the workshop price list
with similar price ranges as other language products? Even better,
can you bring down the other language products to the price similar to
Java and have a "leveled playing field" as people are always asking
for?
Greg
Mark Smith
>It is more precise to say that a Swing GUI is potentially more powerful --
Oh please. Desktop application users aren't going to pay for "potential"
when the reality is JFileChooser.
Ben Z. Tels
users
> actually use it.
I used MS VC++ once. And I was instantly sorry I did.
> How many Java developers have chosen Workshop as their tool
> of choice. ??!!
Who knows? Who cares?
> From what I have read the best tools are : VCafe, Visual Age, JBuilder and
> some other tool called 'Kawa'??
I disagree. Of all the development tools I've seen so far, I liked NetBeans
best. And that's still not a pinch on the ultimate in Java development tools
I've seen: MS-Edit for DOS, Notepad for Windows and joe on Unix.
--
Ben Z. Tels
opti...@stack.nl
http://www.stack.nl/~optimusb/
UIN:2474460
"The Earth is the cradle of the mind, but one cannot stay in the cradle
forever."
--Tsiolkovsky
Ben Z. Tels
> with standard Swing so that I can do an apples-to-apples comparison.
What exactly is so non-standard about the Swing that is available fr JDK
1.1.x? It's not a core API, but other than that.....
> Let's see how fast it goes. For now, the question is: why is it so
> hard to port from Java 1.1.x to 1.2.x?
It isn't. ObjectDomain feels they should keep supporting JDK 1.1.x, because
it is more widespread and they won't lose their installed base that way.
Terry Sikes
Jack Richards <jrxxx...@worldnet.com> wrote:
[snip]
>But we live in the age of spin, and consistency is not much of an
>intellectual value anymore. It's more important to be part of the cause.
>
>Jack
Exactly, Jack. You are spinning like no other, and your lack of
intellectual muscle is more than evident. All you do is keep stating
the same lame conspiracy theory without any evidence to back it up.
Even worse, you keep painting Java and associated forces as less open
than the Redmond behemoth. That is laughable, to say the least.
Your centerpiece seems to be that Netscape somehow "blocked" better
technology by pushing Java. Given that Netscape Java has never been
that great, and that ActiveX support for Netscape has been available
for quite a while (and could have been there much sooner if Microsoft
had written an ActiveX plugin for Netscape, which as I stated sometime
back, is trivial), your argument looks more full of holes than ActiveX
security.
Shouting something at the top of your lungs isn't enough to make
people believe it...it needs to make sense as well. You consistently
fail to address the specifics in other's posts, falling back to vague,
unprovable generalities and emotional language.
It doesn't work.
Terry
--
tsi...@netcom.com
Greg Chien
news:7prpae$9ps$2...@news.IAEhv.nl...
> > Let's see how fast it goes. For now, the question is: why is it
so
> > hard to port from Java 1.1.x to 1.2.x?
>
> It isn't. ObjectDomain feels they should keep supporting JDK 1.1.x,
because
> it is more widespread and they won't lose their installed base that
way.
You really believe that? If they have had a version running on Java
2, they would immediately advertise it. Besides, how hard is it to
support both 1.1.x and 1.2.x?
Greg
Ben Z. Tels
Yes, I do.
> If they have had a version running on Java
> 2, they would immediately advertise it.
But they don't. Instead, they use Java 1.1 . Or more to the point, the Swing
library that isn't part of the Java core but does run under Java 1.1 VMs.
> Besides, how hard is it to
> support both 1.1.x and 1.2.x?
Not hard at all. That's my point. There haven't been any changes to the
language or the VM between Java 1.1 and 1.2; a Java 1.1 program IS a Java
1.2 program. The difference is in the Swing libraries used; in Java 1.1, the
JFC isn't a core API. Java 1.2 changes the package name. When ObjectDomain
feels Java 1.2 VMs are generally enough available, a simple textual
replacement and recompile will bring them up to Java 1.2 .
Greg Chien
> > Besides, how hard is it to
> > support both 1.1.x and 1.2.x?
>
> Not hard at all. That's my point. There haven't been any changes to
the
> language or the VM between Java 1.1 and 1.2; a Java 1.1 program IS a
Java
> 1.2 program. The difference is in the Swing libraries used; in Java
1.1, the
> JFC isn't a core API. Java 1.2 changes the package name. When
ObjectDomain
> feels Java 1.2 VMs are generally enough available, a simple textual
> replacement and recompile will bring them up to Java 1.2 .
A simple textual replacement and recompile...Tell that to ObjectDomain
;-)
Jack Richards
cm...@singapore.sun.com wrote in message
<7pqreq$nrg$1...@new-usenet.uk.sun.com>...
blocking
>JR> desktop component technologies, and what that does to desktop users.
>
>"Blocking" - if you do not support a technology then you are effectively
>blocking that technology. Funny definition.
Not of course when the concept is used to describe Microsoft's conduct. Then
you understand it intuitively.
Basically throughout the economy, companies who are market leaders get to
set technological standards of all types. The "problem" with antitrust law
in not recognizing this characteristic of the progress in the modern economy
has been long noted.
The market leader goes for market share. The revenues from economies flowing
from the adoption of technology is used to go after market share so as to
have future revenue strength.
So market leadership, like Netscape's allows Netscape to establish the
standard. JavaScript is the best example (so diversionary issues are not
pursued to obscure the issue). JavaScript is now the standard for web pages.
A web designer cannot use VBscript, for instance, because not all browsers
support it. However, all browsers support JavaScript.
So it is with component models. Or some rough equivalent like Java.
Netscape in its ABM alliance adopted Java and blocked COM by not adopting it
on the very platform where Netscape achieved its market share.
The Windows desktop is componentized using COM. Its users are accustomed to
user-friendly, cheap and very functional apps from many talented programmers
using a variety of languages.
Enter Java. A big flop on the desktop. If Java had added to the user's
functionality, especially in distributing that functionality via the web.
This was the great promise of Java. The great reality of Java is that is a
Unix world intruder on the desktop that has not done anything for the
world's computer user's, except to keep the desktop dumbed down to the level
of a device.
Sun's motto should be "dumb as a device."
There's no begrudging Sun if it wants to be the "Microsoft of devices" as it
determined in pursuing Java.
But the collateral damage to the desktop has been great. The fact that you
can't or won't see the notion that the power of the PC should be an
expansion of its functionality via a distributed mode is a result of the Sun
Unix culture with its emphasis on selling servers rather than empowering the
desktop.
>
><snip>
>
>JR> Knowledge workers on the desktop are blocked from distributing web
pages,
>JR> via a technology, i.e. components, made for distributing these pages,
simply
>JR> because Netscape, and Sun, are arrogant to believe they can get away
with
>JR> this when the Java hype dies.
>
>How did component gets involve in _DISTRIBUTING_ web pages?
This is not my idea. The idea of distributed objects, now called components,
is not new. If fact, Sun has borrowed the language of components, pretending
these notions are indigenous to Java, without acknowledging their origin.
Web pages are the perfect container for components. The basic Sun notion
that the "network is the computer" is a great slogan for selling servers.
However, Sun with Java has not done anything for the increased functionality
of the desktop.
What do mere mortals want with minicomputers on the desktop, is the McNealy
expression because it knows what has happened to its workstation business.
Right now there is a great battle for the web server market. But the
emphasis will shift back to the desktop soon.
Sun just is blocking the road. That's why it entered the deal with AOL for
Netscape--to keep the desktop dumbed down to the level of a device, so Java
can be a big player in a world of diminished distributed functionality for
the user.
You can snip and run but you can't hide from the facts, no matter how much
in stock options you earn, ultimately at the expense of the user. Be sure to
snip this comment.
Jack
Jack Richards
Dimitri I. Rakitine wrote in message
<7pqqtd$p...@dfw-ixnews13.ix.netcom.com>...
Java is the "VisualBasic" of OO languages. ;) ;)
Jack
>
>: Serious question.
>
>: ..darcy