29 April 2011

[JavaSpecialists 191] - Delaying Garbage Collection Costs

[JavaSpecialists 191] - Delaying Garbage Collection Costs
In modern JVMs, the autoboxing cache size is configurable. It used to be that all values from -128 to 127 were cached, but nowadays we can specify a different upper bound. When we run the code with -XX:+AggressiveOpts, it simply increases this value.

25 April 2011

The Business and Technology of Low-Latency Trading « A-Team Group

The Business and Technology of Low-Latency Trading « A-Team Group

Interesting talking at the A-Team conference, this week.  Lots of focus on equities but managed to squeeze a little FX in there ;).  Enjoyed a presentation by Corvil discussing the art of measuring latency.

23 April 2011

The Five Traits That Get You Promoted to CEO

The Five Traits That Get You Promoted to CEO
# Passionate curiosity: Relentless questioning and being infectiously fascinated with everything around you, human nature in particular
# Battle-hardened confidence: Overcoming—and even relishing—adversity. CEOs most often ask job candidates how they've dealt with failure in the past.
# Team smarts: More than just being a team player, understanding how teams work and getting the most out of the team (in sports terms, being a playmaker)
# A simple mindset: Being concise, simple, and clear in your communications
# Fearlessness: Comfort with the unknown and taking calculated, informed risks; also, seeing opportunities and being proactive about positive change

22 April 2011

Touch Calendar

Discussion on code coverage

100% Code Coverage! | Javalobby. The fundamental point is you need to look at more than just code coverage. If you haven't though hard enough about your test cases you could have a system that is well covered (in the code coverage sense) but is not well tested.
If you have 100% coverage you don’t know if your system works, but you _do_ know that every line you wrote does what you thought it should.

Thus, code coverage and test are completely different things. Taken to the extreme, we could code-cover the entire application, that is achieve 100%, and not test a single line because we have no assertion!

Not all of our code is neither critical nor complex. In fact, some can even seen as downright trivial. Any doubt? Think about getters and setter then. Should they be tested, even though any IDE worth his salt can generate them faithfully for us? If your answer is yes, you should also consider testing the frameworks you use, because they are a bigger source of bugs than generated getters and setters.

10 April 2011

ScaleBase Database Load Balancer

ScaleBase makes some bold claims about transparent sharding, no single point of failure and complete transparency to the application layer. I wonder what happens if you need a transaction between data on different shards?

09 April 2011

Java Floating-Point Number Intricacies Summary

Centos 5.6

The CentOS team is pleased to announce the availability of CentOS 5.6. Major changes in CentOS 5.6 compared to CentOS 5.5 include:

ext4 is now a fully supported file system
libvirt was updated to 0.8.2
bind was updated to 9.7 and supports NSEC3 now.
ebtables was added
php53 is available as a php replacement.
System Security Services Daemon (SSSD) has been added.

Other upgrades include newer version of several wireless drivers, Samba3x, ghostscript, LVM, mod_nss, subversion and gcc, plus others.
Update process from 5.5:
  • yum clean all
  • yum update glibc\*
  • yum update yum\* rpm\* pyth\*
  • yum clean all
  • yum update mkinitrd nash
  • yum update selinux\*
  • yum update
  • shutdown -r now