I pushed out a new version of Engineer – 0.3.1 – this weekend. While 0.3.1 is a minor release to fix a couple of major bugs that slipped through, 0.3.0 was a pretty major release. A majority of the new stuff is under the covers or developer-focused, so it might not seem like a big deal – but it is. With this release I think Engineer is stable enough for me to start hyping it a bit more, so you can anticipate that over the next few months.

The full release notes go into more details about the specifics, but there are two major features I added that are particularly interesting from a development perspective: the new plugin architecture, and the Jekyll/Octopress compatibility work.

For whatever reason, I’ve seen this relatively old article from Steve Bennett linked to by several people over the last day or so. It’s definitely worth a read, and I agree with a lot of it. But I think Bennett misses the mark in a few places.

First, the things I agree with:

Yes, the command line and official Git documentation suck.

I’ve complained about this before. Software engineers like to believe that if we know how to do one thing, then we can work out how to do all similar things quickly by applying that knowledge. After all, software engineering is advanced problem solving and critical thinking, so applying previous knowledge to current problems is something we’re pretty good at.

Frankly, the Git command line breaks this all the time. Some commands need some flags; others don’t. Conceptually similar operations require different top-level commands (not just flags) depending on what you’re doing that operation on.¹ It’s madness, and I think it’s safe to say that the command line isn’t designed; it’s merely implemented.

Note: I’ve actually had this sitting in my drafts folder for awhile but with the recent release of GitHub for Windows it seemed like a good time to get my act together and post it.

The Beginning

My first experience with true source code management was with CVS. I was in college, and the HawkTour project was going into its second semester. It was becoming clear that with a new batch of undergraduate students coming in the next semester we needed a solution to keep everyone from Hulk-smashing their way through the code wreaking havoc and bringing general chaos.

The Unix server we used had CVS installed, so I learned the ropes then gave new students a crash-course in using it, as well as how to check things in and out using the Eclipse addins. This was the first time almost every student I encountered had used any source code management software, which in retrospect is a travesty. I really hope college courses include source code management tools these days.