Episode 108 Comments Thread

[Dick Wall]

Leave 'em here...

[Joshua Marinacci]

Here is the link for AB5k:

The AB5k website:

http://www.ab5k.org/

The AB5k blog:

http://desklets-news.blogspot.com/

The AB5k Google Code project:

http://code.google.com/p/desklets/

A more indepth discussion of AB5k on my blog:

http://weblogs.java.net/blog/joshy/archive/2007/03/ab5k_widgets_in.html

Thanks everyone. Let me know what you think.

- Josh

On Mar 21, 2007, at 5:29 PM, Dick Wall wrote:

Leave 'em here...

- Blasting forth in three part harmony!

[David Bates]

I couldn't find a link in the show notes to the User Friendly google
phone, so here it is:

http://ars.userfriendly.org/cartoons/?id=20070318

On Mar 22, 12:29 am, "Dick Wall" <dickw...@gmail.com> wrote:
> Leave 'em here...

[andrew.bruce.law]

Hi Dick,

Any links for the new Google Podcast yet? Sounds interesting...

Another great one BTW. Made the weekly flight home "fly" by... ;-)

Andrew

[Dick Wall]

Hi Andrew

The first Google podcast has been recorded, but there are a few more
steps before it is ready. For one - it has to go through legal to make
sure we aren't saying anything we shouldn't (no one is going to bother
suing the Java Posse - I hope, but Google is another matter with those
deep pockets). We also have to set up a blogger account, burn the feed
and add it to iTunes, but none of these things should take more than a
week or so. The interview with Crazy Bob is excellent and I am really
looking forward to getting this podcast out there.

Thanks for the interest - we will be sure to talk about it on the
Posse when the Google podcast does go out.

Cheers

Dick

On Mar 23, 5:06 am, "andrew.bruce.law" <andrew.bruce....@gmail.com>
wrote:

[kohlerm]

Hi all,
I just read Brian Goetz's article about Java performance myths at

http://java.sun.com/developer/technicalArticles/Interviews/goetz_qa.html

Also I agree with some of his points such as that some of the C
programming rules just don't work for Java, but I also disagree with
some other statements.
He explains for example that realistic tests are needed in order to
decide whether something should be optimized or not.
I think that is true to some extent, but on the other side I daily see
that a lot of problems could be avoided with running some simple test
or by taking some simple rules about resource consumption (memory)
into account.
If you for example know that for one click in your application you wil
create 100Mbyte of garbage objects for a single click of a single
user, you can be pretty sure that your application will not scale up
well with an increasing number of users.

The same is true to some extent for the memory consumption of your
java objects. Also the memory layout of the Java objects is not
specified by SUN, you can compute the memory consumption of your java
objects on the SUN VM, by either using the profiling API for by using
these rules: https://www.sdn.sap.com/irj/sdn/weblogs?blog=/pub/wlg/5163.

Hi also explains how cheap object allocation really is, which is true,
but he doesn´t mention that you still have to be careful because in a
highly multithreaded server application you may end up with a lot of
objects being promoted to the old generation which can kill your
performance because full GC´s will be triggered.

Regards,
Markus Kohler (https://www.sdn.sap.com/irj/sdn/weblogs?blog=/pub/u/
6389)

[Chris Nappin]

On Mar 22, 1:29 am, "Dick Wall" <dickw...@gmail.com> wrote:
> Leave 'em here...

Re: Google Guice

Just to briefly jump to Spring's defence, (and before the speed up
figures start to be mis-quoted out of context throughout the web), the
figures are the once-off cost of pre-instantiating objects as your
application starts up. I don't believe there is any implication that
runtime performance would be any different.