www-jcp-open mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Wade Chandler <hwadechandler-apa...@yahoo.com>
Subject Re: [VOTE] New ASF/JCP Policies
Date Sun, 15 Jul 2007 12:41:19 GMT
--- Steve Loughran <stevel@apache.org> wrote:
> Rory Winston wrote:
> > Hangover of VJ++? Really? How so?
> > 
>...
> The problem was, the negotiated license wasn't
> adequate to deal with 
> these. it didn't say "you cant add stuff" and
> although there was a 
> requirement to redistribute everything in java.* and
> javax.*, putting it 
> up on an FTP site was, microsoft argued, adequate. I
> ended up writing an 
> ActiveX control that forced the download; anyone who
> hit the page would 
> get a copy installed.
> 
> This is why, today the T&Cs of the java and jcp
> licenses are so strict about
>   -adding stuff to the java, javax packages (no
> additions)
>   -taking stuff away (the completeness requirement)
>   -compliance: you can't just comment out the RMI
> stuff and throw 
> connection refused exceptions
> 
> It's not just about retaining control/profit, its
> stopping anyone from 
> taking control away from Sun/the JCP.
> 

I don't think one can say Sun/JCP and it have the
intended meaning. Too many companies involved with
their own JSRs and have influence to say Sun tries to
force some great totalitarian control. Why do you
think it is hard to get all the companies to agree on
this or that model? (All the companies have different
investments they want to assert some type of control) 

The reason for the changes not being wanted is that
you or I would write a "Java" application and it would
not be a "Java" application. You could not run it on
any Java runtime if there were so many differences
etc. So, instead of having developers have to deal
with all of that the changes came about to protect the
base. 

Have you ever written a C++ library (dynamic or
static) and tried to use it in another binary
environment or even compile in another compiler? It
isn't very easy if the IDE you used ties a UI or some
other utilities into design time as usually there will
be special keywords or undefined things which you will
have to have.

I happen to like the idea of embedding a JVM with an
application, so I think as long as the application
doesn't claim to use "Java" it would be something good
to change about the JVM and runtime implementation and
 distribution policies. Sun is hung up on the JVM/JRE
being a system component such as Perl and this is the
main problem. 

I think we can have both and it work perfectly well
and be flexible enough for everyone without forcing a
one size fits all model on all implementations. That
way a system level implementation has to match the
specification 100% and the others can do anything they
want unless in an embedded environment where there is
no chance of other developers using the platform (like
a router or something) in that case it would not
matter as the implementation would only be exposed to
the developer. This is something the JCP groups could
try to work on.

Wade

==================
Wade Chandler
Software Engineer and Developer

Netbeans Community and Dream Team Member:
http://wiki.netbeans.org/wiki/view/NetBeansDreamTeam

Check out Netbeans at:
http://www.netbeans.org

Mime
View raw message