I Built His Cage

Announcer: Now for quick hits and commentary on software development topics from around the web, the EIP web-ring brings you the stigmatized spawn of a refactory, MoffDub, and Helltime!

• Watts S. Humphrey of the SEI takes a step back in the July/August issue of CrossTalk and tries to figure out why large software projects almost always fail. The big takeaway for me from this lengthy article is that Watts first says that managers aren’t the problem, but rather the management method: managers like to “Manage By Walking Around”, and this works because they can see, feel, and hear the work being done as it is being done; he goes on to say that because software is different because it is essentially imaginary, these typical techniques do not work, and managers end up sticking their noses where it doesn’t belong and generally pissing everyone off; finally, Watts says that the engineers must manage themselves since they know the work best and can give the most accurate reports and status!

  So who needs managers? I swear, every time I make a joke about those people (managers), it comes true.

• To see how the new kids on the block are looking at us OO old-timers, take a read at Isaac Hodes’ CopperThoughts blog, where he covers Java for Clojure programmers. Aside from the accurate description of classes in functional terms (“A class is a bundle of methods (functions which act on the class) that can serve as a data type.”), it is interesting to see the syntax Clojure uses for Java methods; it seems that a method name is passed to a function as data, kind of like Smalltalk. It harks back to my playing around in various functional languages to force-fit OO where it didn’t really belong.
• Dan McComas, web developer for BayCitizen, fears becoming an old web developer because he cannot think of any he has worked with recently. He figures that most of the people that would have been 50+ developers are now managers. Now, he contemplates abandoning the virtuous path of code creation. Don’t do it, Dan. Don’t do it.

Announcer: You’re reading the EIP web-ring.

Announcer: Now for quick hits and commentary on software development topics from around the web, the EIP web-ring brings you the stigmatized spawn of a refactory, MoffDub, and Helltime!

• In my once-pined-after industry of game development, GameDev.net contributor Andrey Karpov offers an epic post about real-life 64-bit problems in C code. I could only get through the first 20 examples before my VM-shielded brain started shutting down.
• The site DeveloperArt.com complains that programmers are hired as if they were lying children, and its author seems to resent this treatment to the point of refusing a job offer. Sorry to break this to you, author, but FizzBuzz is reality. It’s our own damn fault that employers screen us as if our industry was full of fools, because it is. To wit:
• Shut up and write your unit tests, says Robert Diana, JavaLobby poster on DZone. This post is a timely find, as a fellow developer-turned-evangelist at my place of employment expressed her bewilderment that one of the teams she is advising is “melting down” at the thought of writing unit tests. Call it the Agile version of “shut up and code”.

Announcer: You’re reading the EIP web-ring.

Announcer: Now for quick hits and commentary on software development topics from around the web, the EIP web-ring brings you the stigmatized spawn of a refactory, MoffDub, and Helltime!

• Ray Chen recounts the old new thing about garbage collection and has an unusual take on the definition. Garbage collection is considered to be the simulation of infinite memory. I suppose that makes sense from your program’s point of view, and I guess from the program author’s point of view: you get to write code as if you had an infinite amount of memory and never had to clean up after yourself. Ray uses this concept to draw the bottom line: don’t assume anything about the timing of garbage collection, particularly of your finalizers in .NET.
• The offering at simple-talk on database version control by Mike Mooney is epic, and very worth the read. Predictably, I am not aware of any version control of our databases at my place of work, and I certainly never did any version control on The Project’s SQL Server. While it seems silly, this post is good at breaking down the essential differences between code and data that make versioning the latter a non-trivial task. Interesting that he concludes, like I have, that developers should have their own, private, isolated copy of the database.
• Ending the Helltime show is a cautionary don’t-reinvent-the-wheel chide from the folks at ElegantCode.com, pleading that you don’t write that millionth XML parser and just use something built-in. The methods cited seem slightly involved, and maybe that is how it is for parsing. In particular, I don’t like how the classes are usually written for you, but if they are just data structures, what’s an extra layer of mapping between friends? John Sonmez never did cover building XML, but I will in a post in the near future, since the problem is currently my main focus at the old cube farm. Stay tuned.

Announcer: You’re reading the EIP web-ring.