Suggestions for Java Generics Semantics in Java Version 7.0 _ Resurrected
kirillcool at yahoo.com
Wed Jun 17 17:28:47 PDT 2009
Just to add a clarification to the question on commit rights. While BSD ports, XRender pipeline or portable 2D backends are all great projects that take the platform to new directions, i don't see them touching the core of the platform - in the way the generics reification can.
From: Kirill Grouchnikov <kirillcool at yahoo.com>
To: jdk7-dev at openjdk.java.net
Sent: Wednesday, June 17, 2009 5:21:17 PM
Subject: Re: Suggestions for Java Generics Semantics in Java Version 7.0 _ Resurrected
>> The usage of words "valuable" and "evidently" is very subjective in this
>> context, since i haven't seen any evidence of true community feedback,
>> or any sign of a transparent decision / development process in JDK 7.
> I think that it's pretty self-evident that the JDK 7 development
> process is a lot more transparent then any previous JDK's before.
> This discussion wouldn't have been possible in the bad old days.
I have written about the transparency / opacity of the JDK process last
November at  and pretty much nothing has changed. I am not working
on OpenJDK 7, nor do i want to see your "nothing wrong" comment as an
implied lessening of the opinions of the overwhelmingly vast majority of
Java developers who don't contribute code to OpenJDK 7. I don't have hard
numbers, and i never saw Sun publish them, but compared to millions
of developers and "billions" of devices - how many non-Sun employees have
contributed code to OpenJDK 7, and how many of those have not worked on
Java prior to JDK 7? In fact, is there any single developer with commit rights
to the code base that has not priorly worked at Sun on Java?
I am a close follower of what happens in Swing / AWT / Java2D. I see only one
contribution in that area - the XRender pipeline. Everything else is decided
and designed behind closed doors. Then we get the final result without any real
discussion on what we - as the users of the product - would have liked to see.
The shaped / translucent windows API is the prime example of the functionality
that was added without any prior public discussion on what will be supported
and how the API will look like.
If you want more examples - i have plenty. The withering JSR 296 (App
Framework), the officially dead JSR 295 (Beans Binding) have both left the community
for many many months wondering what is going on, waiting for the next
big conference to see the same slides and hear whether or not they will make
it in the final JDK 7 distribution. What about the JavaFX JMC media layer?
Are there any plans to include that in JDK? May be there are, may be there are not,
but i'm certainly not going to get any answer in a public forum. The same
goes about JWebPane - with more than a year of promises and still no concrete
information on the availability, licensing and JDK version support. I can go on
and on about missing components, modern layouts and Nimbus designer,
but i think you see what i'm trying to say.
About contributing to OpenJDK 7. I follow quite a few Swing related
discussions and i see an excruciatingly slow and exhausting process for
even the simplest things - such as removing warnings. It goes on and on for weeks
and months, with outside contributors repeatedly reminding Sun engineers that they
are waiting for answers. What good is an open mailing list when you need to wait for
weeks to get an answer? What good is a source repository when you need to send a
webrev and wait at the mercy of Sun engineers to check them in?
Yes, i do have quite a few ideas that i would like to see implemented in core Java -
specifically for Swing. Some are big, some are small, some are breaking backwards
compatibility, and some are extending the client side Java. Speaking realistically, i see
zero chance of having these ideas making it into the core Java. The process is too
closed, too long, too convoluted, and absolutely not open to democratic decision making.
More information about the jdk7-dev