So far

Nate touched on some of the features the newer versions of Java provide for concurrency at an OCI Java Lunch a couple of months earlier. Eric’s follow on Java Lunch concurrency topic nurtured the planted seed, but the BIG THANKS goes to Charles (I like calling him Charlie, with NO idea what he thinks) for sending out an email that provided the tipping point. I’m repeating facts (one of my patterns)… on to the show…

The screen shimmers and waves to a clear picture of six people sitting around a conference table. Some with books, others with laptops. Charles provides introductions and greetings, one attendee is even on the other side of a speaker phone working from home. The lunch time study group is REALLY on it’s way.

Chapter two of JCIP “Thread Safety” propones that the issue is about “managing access to state… shared mutable state”. If our programs don’t manage this correctly they are broke. That’s a capital period. The idea that’s trying to be communicated is that, what we used to believe about visibility of variables and objects, the order of execution of code, even the values that a variable has no longer is as simple as it was. Specifically when programs threads start running concurrently.

For me the moment of ‘Oh no, now what am I going to do’ came when it was pointed out the increment operation (i++) is not atomic, that it is a read, modify and write. The chapter went on to provide several insights into ways I need to think differently if I’m writing code for multiple thread access. Covering the topics of Atomicity, Locks, Guards, Liveness and Performance I was able to quickly grasp the authors’ definitions.

The book clearly shows me code that is broke and steers me to solutions on what I can do to make it correct. I am MOST grateful for the bits of humor, there is just not enough levity in this world. I enjoyed the discussion around the table and the time went by quickly. I was in a bit of a hurry, my tee time was quickly approaching. Slow fade…

What is This?

Back on March 29, 2007, Charles Sharp sent this email to his fellow OCI/Advantage employees:

Greetings All,

Now that Eric has whetted your appetite for Java’s take on concurrency (or thrown you into a panic due to your understanding level), I thought this would be a good time to see if there is an interest in getting a  study group together to go through the book, Java Concurrency in Practice by Brian Goetz et al.

The study will be done in two parts.  The book has 16 chapters.  Part one of the study would cover chapters 2 through 9.  Part two would cover 10 through 16.  We’ll decide about Part two after we see how Part one goes.

Today I’m soliciting people to participate in Part one.

There will be eight meetings.  (The group might decide to have a pre-study organizational meeting to discuss and argue about format changes in the usual uninteresting fashion.)

Study meetings will be held during lunch hour (something like 11:30 to 12:30 with a half-hour overrun buffer to 1:00) once a week at the OCI office.

Due to other meetings through the week, I’m proposing either Monday, Wednesday, or Thursday as the meeting day.

The format of a meeting, as is always the case with a group of experienced wonks, is open for discussion.  Here’s a first pass to start that discussion.

Everyone reads the week’s chapter before the meeting.  (From my perspective, this is non-negotiable.  If you don’t have time to read a chapter a week, you just don’t have time to sign up for the study group.)

Each week one person will be assigned the chore of providing code that illustrates the concepts discussed in the week’s chapter.  I expect most of the attendees relate to code examples as well as anything.  So please note that part of what you’re signing up for is to provide this illustrative code for one of the meetings.

After the code presentation and attending discussion, I would like to leave time to do a quick “round the table” where each attendee has a chance to pitch in some point (or points) from the book that didn’t get covered.

And that’s pretty much the format plan.  In summary, we’re agreeing to meet once a week for eight weeks and do the homework.

If that sounds like your idea of fun, please stay away from me.

If that sounds like something that would be useful and constructive and a good leg up in understanding some crucial concepts, then give me a shout via reply email.

And thus, the study group was formed. We first met last week on Monday, April 16, and have met every Monday since then! Enjoy.