Hacker News new | past | comments | ask | show | jobs | submit login
"A JVM Does What?" (jeremymanson.blogspot.com)
166 points by fogus on April 6, 2011 | hide | past | favorite | 20 comments



If you're interested in this, you might also be interested in his high-scale-lib library: lock free maps and other data structures for the JVM.

http://sourceforge.net/projects/high-scale-lib/


Is there a non video version of this?


And can we please get back to tagging videos with [video] at the end of the title?

  If you submit a link to a video or pdf, please warn us by appending [video] or [pdf] to the title. [0]
This is the second all-video submission today (the other being the Job's interview) that wasn't tagged. It's not that big of a deal but when your on a mobile connection and you spend a minute waiting for a video you can't even watch to finish loading it can be really annoying.

[0] http://ycombinator.com/newsguidelines.html


Slides from his Java Symposium presentation which probably covers similar material:

http://www.slideshare.net/javasymposium/a-jvm-does-that-pres...

Title is actually: A JVM Does That?


For what it's worth, I don't think I could possibly read as quickly AND effectively as he talks. What a great speaker! I learned a lot.


I haven't watched the video yet, but will be shocked if it contains ideas that it would take 52 minutes to absorb in written form.


Anyone care to post a short jist on what this 52min video includes ?


"Here are some common misconceptions about the JVM and how it works, and a few interesting titbits. These are some things people want in, or to believe about the JVM. Here are some things which we would do to fix the odd bits, and make things a bit nicer."

The slides are here: http://www.azulsystems.com/blog/wp-content/uploads/2011/03/2...


Considering GC and concurrency issues, this talk is good for people making language choices on the JVM. In the end, he said he personally thinks STM is not a good way to program concurrency. And, I think immutable data is good for GC.



I went to the version of this talk he gave at Java Symposium; his argument was that it can work for Clojure, where immutability and functional purity resolve some of the problems. Perhaps these discussions moved his position slightly.

Or maybe he just learned Clojure people are pedantic and it's more time efficient to agree rather than argue with them.


Apparently Project Fénix https://fenix-ashes.ist.utl.pt/ at the Instituto Superior Técnico in Lisbon somewhat changed his mind about STM last May: http://www.azulsystems.com/blog/cliff/2010-06-04-or-how-i-go...



Or look at how STM is handled in Haskell, where it's arguably more pure.



Its by Cliff Click, from Azul Systems. (the "pauseless GC" guy :)


He says he doesn't like bytecode for expressing programs, but he never mentioned what else would work better. I'm intrigued, does anyone know more?


i was wondering about that too. he later mentioned that the format is bad and that .net bytecode has a better format, so i think maybe he didnt mean bytecode in general but the format of jvm-bytecode.


Very interesting talk for the Java geek in me. Side note: blogspot's UI has gotten really old.


Re: side note, that's why they announced this: http://buzz.blogger.com/2011/03/fresh-new-perspectives-for-y...




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: