Archive for Helltime

Helltime for October 11

Posted in Helltime with tags on October 11, 2010 by moffdub

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!

Battling a cold and lack of connectivity between physically-separated Erlang nodes, I bring you abbreviated Helltime.

  • Uzelezz gives a two-part post on test principles, a good restatement and expansion upon Uncle Bob‘s advocacy of clean tests, something I have been trying to practice myself.
  • A post on the MuleSoft blog gives a good introduction for why and when Kanban should be introduced, but the post dances around a basic issue: how do you determine the WIP limit?
  • The official Google blog announces the inevitable, that they will soon be driving all of our cars. Hell no. I see first hand how software is written and I’ve meet a lot of the people who write it. In no way do I want these people involved in how many left turns I make into oncoming traffic.
Announcer: You’re reading the EIP web-ring.

Helltime for October 4

Posted in Helltime with tags on October 4, 2010 by moffdub

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!

  • Michael Stonebreaker beeps out a communication of the ACM on why Enterprisey shops aren’t biting at NoSQL. Three cited complaints are low ACIDity, low-level query language, and low (no) standardization. From what I understand, NoSQL and its cohorts have a shot at the latter two. ACID properties are a property of the database system, and some, like MongoDB, let you choose how consistent you want to be (at the expense of other guarantees).
  • Omar Al Zabir thinks that my shop breaks one of the ten commandments of caching: caching configuration settings. Indeed, most of our configuration settings, be them Spring properties or the legacy route, are loaded once per JVM, and any change to them requires a server restart. We even had a production issue that was solved by a mere restart. It’s no joke.
  • Matthew of “You The User” embodies the Peter Principle with gems like this:

    If you aren’t being challenged in your job – leave.

    If you aren’t sweating it at present and feeling under pressure – leave.

    If you think you understand everything that you are doing where you are currently working – leave.

    More shocking is that nobody in the comments section has brought this up. If the comments section lacks the obvious – leave.

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

Helltime for September 27

Posted in Helltime with tags on September 27, 2010 by moffdub

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!

  • Daniel Jalkut has a post on his Red Sweater blog about masking failures in a system, and it echoes some of the biggest complaints against the Null Object Pattern. Fail as early as possible. Just make sure your error handling will log the exact line of code causing the problem.
  • Interestingly enough, CrossTalk‘s September/October issue has some hidden flame bait in an article about static analyzers:

    Embedded systems expert Jack G. Ganssle advocates doing inspections before testing because inspections are 20 times cheaper than writing tests: “It is just a waste of company resources to test first

    The rest of the article is wordy but makes the point that a good static analyzer is written for humans, and since certain problems are undecidable, an ideal tool will automate what it can and provide information for what it cannot and let the human review.

  • Fellow WordPresser Luke Palmer is all philosophical as he considers what is the fundamental unit of a computer program. The assertion: programs are made of ideas.

    Yes, this is true. It also sounds very ethereal. But it can be said about most things, can’t it? This isn’t to say that I agree with the original idea, that fundamentally, programs are made of instructions. I’m not so sure that the question even makes sense, but put my back against a wall and I’d say that programs are made of abstractions. There isn’t a line of code that people write today that isn’t aided by ignoring some underlying detail.

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

Get every new post delivered to your Inbox.