cocoon-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Kevin A. Burton" <bur...@relativity.yi.org>
Subject Re: [wanna laught?] Microsoft invented XML
Date Tue, 16 May 2000 06:35:38 GMT
Robin Green wrote:
> 
> >The point here is that no other company in the world, including
> >Microsoft, has done more to *crush* Java than SUN.
> 
> ??? !!!
> 
> In what ways have Sun "crushed" it? It doesn't look very "crushed" to me.
> They must have spent millions on all the products and APIs they have
> developed for it - many of which they give away free of charge.

When was the last time you used "Microsoft Office" for Java?  Java has
0% percent penetration on the client and certainly doesn't have majority
market share on the server side.
 
> If they had cancelled all further development - yes, that would be crushing
> it. If they had written Solaris-only APIs - yes, that might be described as
> crushing it. But they evidently have not.

Where is WORA?  Doesn't exist.  Essentially Java means Solaris, Windows,
some Linux and then every other OS is on their own.  
 
> >There were some very
> >smart people initially involved with Java.  Then the laywers and
> >marketing guys got involved and screwed things up.
> 
> Um, a technically superior product often fails without sufficient marketing
> (dare I say it, even overhyped marketing at times). Remember OS/2?

Nope.  I would say that a superior product always wins.  By definition. 
OS/2 failed because it was inferior.  It was inferior because tehnical
specs/limitations aren't the only thing that makes a product superior.

Any time OSS is superior it has a significant advantage because the
marketing people that can kill a technical product (in the above case
that would be IBM) don't exist and any company can make the correct
choice.
 
> No? Exactly! ;)
> 
> Okay, word of mouth can do a lot. Sometimes marketing is unnecessary. But
> marketing doesn't necessarily "screw up" a product, either. Where is your
> evidence that marketing or lawyers have "screwed up" Java?

Have you read the license to the JDK?  :)
 
> >The point is that with the lawsuit against MS SUN has basically said
> >that they don't want anyone to innovate in the Java space.
> 
> Nonsense. They have implied no such thing. What about Pizza, what about GJ,
> GCJ, TowerJ? 

YES!  All of which have *NOT* licensed Java from SUN so they can't be
sued !

> You can literally embrace and extend Java, AS LONG AS YOU DON'T
> TRY TO PASS IT OFF AS COMPATIBLE WHEN IT ISN'T! It is obvious, in my
> opinion, that Microsoft deliberately broke RMI etc. to put a dent in Write
> Once Run Anywhere, and deliberately made their J++ documentation misleading
> to obscure the fact that J++ is designed to generate hard-to-port,
> Windows-bound code. Even if the latter is not legally actionable, the intent
> is clear, at least to me.
> 
> Kevin, this is really weird. Are you trying to claim that Microsoft was
> wonderfully "innovative" with J++? "Innovate in the Java space" - that
> sounds like marketroid speak to me.

Yes.  their RNI (JNI implementation) was much easier for Windows
developers to use.  Also at the time J++ was clearly a better IDE than
anything on the market. (but let's not argue over that point :)
 
> >  This has not
> >been a good thing for Java.
> >
> >The point is that if SUN were to pick a good license, maybe require the
> >code to be republished, then MS would have had to release the code and
> >then their embrace and extend philosophy wouldn't have worked.
> 
> Why wouldn't it have worked? Publishing the source code doesn't guarantee
> compatibility.

Sure it does :).  SUN could bridge the code in their JVM or provide
developers a way out.  Microsoft's empire exists (well, one major
reason) because all of the Win32 API is closed and it is impossible for
you to take a Win32 app and port it easily.  If you have the code you
can bypass this strategy and have global apps.
 
> Nor does open sourcing it, either. Only enforced, non-open-source contracts
> can do that, as far as I can see, because the open source definition does
> not allow the imposition of compatibility requirements (see Clause 4), IMHO.
> [IANAL.] Sun have gone almost as far as they can with the SCSL, in my view.

Right.  I do not think that products like Linux should be licensed like
this.  But the specs/technologies on which Linux is based should be
provided/developed within a standards body.  I have no problem with SUN
being a standards body to write up specs but the core JVM should be OSS
and the community should be able to do whatever they want.

If it doesn't break the spec then you should be able to do WHATEVER you
want with the JVM.  A good example of this is TowerJ.  Tower was not
allowed to license SUN's Java because they were working on a compiler
which SUN didn't like because of political/marketing reasons. 
Technology should never be limited by marketing/legal.
 
> > > Which OpenSource project has
> > > ever gone to court over licenses, how about Sun, and how many times has
> > > MS (and others) ignored licenses, to crush opponents?
> >
> >Open Source wouldn't change this fact. If SUN has an Open Source JVM and
> >someone broke the license they would have the right and will to sue.
> 
> If it enforced compatibility it wouldn't be open source, technically and
> practically.

Correct.  But with the source public you could change it back to being a
standard. 

> Having a mishmash of Java VMs, all with different features, would be a
> disaster. 

We already have that.  The situation with Collections in JDK 1.1 vs 1.2
is clear.  I would have given a BIG -1 to this design if it was proposed
in an open manner.

The point is that there shouldn't be multiple JVMs.  Just like the Linux
kernel it is a standard code base.  IBM might have a flavor of Java but
it would be OSS and any vendor could take it back.

> Okay, there are different versions of the platform, but with
> versions 1.0 to 1.3 at least you know where you stand, and you have 99.9%
> backward-compatibility.

Not really.  The SUN JVM crashes flat when I use a specific
multithreading/sychronization strategy.  The IBM JVMs do not crash.  If
the SUN JVM was OSS I would just fix it :)
 
> Getting Cocoon to work with different bad implementations of different
> Servlet APIs is bad enough - 

This is over now.  The problem is that they were closed source before. 
Now they are OSS.  Happy days :)

> if Java was a non-standardised platform it
> could be orders of magnitude worse! Let's not regress to ugly
> implementation-specific hacks.

Right.  The code should be in one CVS and things will have mini forks
(just like Linux distributions are all slightly different) but we would
all be moving towards one standard goal.
 
> > > Those questions should be asked before requesting Java to become
> > > OpenSource. After all, Open Source is a method to improve code, and you
> > > are free to do so, and your work could be for the gain of all others.
> >
> >If you don't believe in Java and Open Source, why are you here?

> I suggest we bring this discussion to a quick close; it is not very relevant
> specifically to cocoon.

I agree.  But it is an important issue.  I am writing a paper on this
right now and open discussions are good :)

Kevin

-- 
Kevin A Burton (burton@apache.org)
http://relativity.yi.org
Message to SUN:  "Please Open Source Java!"
I just patented "one click e-mail", when you hit the "reply" button you 
own me 50 cents.

Mime
View raw message