Re: Engineyard

On Tue, Jul 28, 2009 at 8:19 PM, john casu<john@...> wrote:
> I wish you guys nothing but good things, with your move to Engineyard, but
> having done similar moves myself, in the past, I'm a little concerned as to
> the future.   And, I hate to be the party pooper, but these questions need
> to be asked:

I think most of these have been answered in some form elsewhere, but
for those on the list...

> Why did you leave Sun?  Did you get the impression from management that
> jruby was destined to be shelved, or at least deprived of resources or a
> long term commitment that you would need to do your work or make jruby
> successful?

The primary reason for leaving was the high level of uncertainty
involved. On the one hand, we were getting absolutely no indication
that JRuby either would or would not be funded. We weren't even
allowed to talk about JRuby's future because of regulatory concerns.
That makes it difficult to get anything done and certainly difficult
for us to present JRuby to prospective users. On the other hand,
EngineYard was eager to bring all three of us on board, consider
adding more resources, and explore offering real commercial support
options. They'd committed to providing JRuby options in their cloud
offering, they had the two most active Rails core committers on staff,
and they were very interested in bringing Ruby to the Java world.

In that situation, I doubt anyone else would have made a different choice.

> Didn't Oracle think of the whole jruby/rails/oracle combination as being
> strategic, from their POV ?

We have no idea what Oracle thought, because we were all forbidden by
law to communicate. Even if we had asked Oracle about their plans
(which we did, early on), they would be unable to tell us.

> How does your move from Sun to Engineyard affect your ability to work with
> the JVM people?  The "j" in "jruby" is the entire reason for jruby's
> existence, imho, with the ability to interface with the jvm, and  directly
> access all (or most) of the classes and methods available in a jvm
> installation.

I don't think it will harm our relationship with the JVM guys at all.
We're still going to be working closely with them to make sure the new
Java 7 features like invokedynamic work well with JRuby, and we'll be
sharing with them information about how JRuby is utilizing the JVM so
they'll be able to keep improving it. And I'm also hoping to attack
the "J" side of JRuby over the next couple months, filling in all the
integration gaps we've been missing. I'd even say we'll spend (or at
least I'll spend) more time working on the "J" side than I did while
at Sun, but still keep the "Ruby" side moving forward as we have in
the past.

> Does your move affect the direction of jruby development, in the sense of
> some development tasks being prioritized over others?  For example,
> de-prioritizing even further the integration of non-rails related classes
> like Java3D, etc..  Yes, there are some of us who don't give a flying rats
> ass about Rails (maybe we care a little).

I think the emphasis on Java integration is actually increased by the
move. In order for us to really bring Ruby to the Java world we need
to make sure the "J" side is top-notch. And admittedly, there are gaps
in our Java integration. So we're going to fill those gaps, improve
performance of both Ruby code and Java integration code, and finally
bring JRuby parity with other options on the JVM as far as Java
integration goes. My prototype Hibernate and JUnit4 annotation support
is just a taste of things to come.

- Charlie

---------------------------------------------------------------------
To unsubscribe from this list, please visit:

    http://xircles.codehaus.org/manage_email