Toluu

Enfranchised Mind

Robert Fischer and Brian Hurt on Punditry, Programming Languages, and Other Religious Issues

Experience of a Freemason: Thoughts a Few Years InJanuary 5

Editor’s Note: For more on Freemasonry from this blog, see the Freemasonry category.

---

By popular demand from the followers of my Twitter stream, I’m cataloging my journey into Freemasonry. At this point, I have a vantage point of being involved for a few years, and so I have a bit of perspective on the events. From that perspective, I’ll say that all in all it has undoubtedly been worth it, although not in all the way I expected.

---

After leaving college, I hit the real world and discovered it to be a cold and lonely place. In college, I had been active in the ΒΚ chapter of ΘΧ, and for all the ups and downs that fraternity life offered, it was an excellent sense of community. I had something that I could work for and with: an organization whose ideals I admired and where people were in the same place as life as me, but where people had very different ideas and engaged in a dialog in those ideas. Furthermore, my brothers in ΘΧ really helped me grow up and become a better man — they held me accountable when I was being an idiot, and the opportunity to be their recruitment chair was one of the key formative experiences in my self-identity as a leader.

It didn’t take me too long after leaving college to realize that I wasn’t going to work well in the role of “alumnus who always hangs out at the house�

Useful Grails Integration Testing Snippet: Delete All Instances of Domain ClassesJanuary 3

Just thought I’d share this quick:

import org.codehaus.groovy.grails.commons.ApplicationHolder   class FooTests extends GroovyTestCase {   def sessionFactory   void setUp() { (ApplicationHolder.application.getArtefacts("Domain") as List).each { it.newInstance().list()*.delete() } sessionFactory.currentSession.flush() sessionFactory.currentSession.clear() } }

That code deletes all the instances of all the domain classes. Yes, the transactional aspect of Grails integration tests should make sure everything is rolled back properly. But there’s always something that sneaks into the database one way or another (often as the result of exception handling or other weirdness), so I like to be thorough. I had the same problem back when I was developing Rails. If you have that problem, feel free to have at the above code.

Also note the usefulness of Spring dependency injec

The problem with STM: your languages still suckJanuary 2

So, I’m reading Brandon Werner’s post on STM, about why Software Transactional Memory hasn’t caught on yet. There are three problems with STM that make it a CS Researcher Full Employment Act, allowing many to try and none to succeed in implementing STM for “mainstream” languages like Java, C#, and C++. They are:

1. What happens when you access a piece of transactional memory from outside a transaction? This allows for race conditions to happen, even on transactional memory. Worse yet, these race conditions can often reveal to the programmer the inner workings of the implementation of STM- meaning that it’s now possible to write code that works on one implementation of STM, but not on another.

2. What happens when you access a non-transactional variable or perform an IO from within a transaction? If you start a transaction, then call the function launch_missiles, then abort the transaction, what happens- is Moscow still vaporized?

3) The performance implications of transactional memory. As programmers, we’re used to assignments being cheap. Basically, if we see the code:

x = y;

we assume it gets compiled into code something like:

movl %eax, (%ebx)

i.e. a single instructi