Can We Have Too Many Tools?

Vim is a great editor. There are so many plugins that make it even better and increase my productivity. But can there be a saturation point where it's not worth finding the next plugin to shave a keystroke off of a command? I've been trying to find the sweet spot for the right amount of tools for the job.

Bigger Toolbox

Having more developers to talk to and work with, I have been exposed to different methods of development. Yan in particular, has been a tremendous Vim resource. He's made it his focus to optimize working with this editor as much as possible, and it's quite impressive. I've adopted some of his ideas, no longer content to be "good enough" with Vim.

Janus has been a tremendous help. It exposed me to command-T, which is now my favorite thing ever. I no longer use a buffer explorer because it's easier to just find it with command-T than search through the buffer list. I also rarely use NERD tree now, since I was using that mostly as a convenience to open project files. I'll still use it for looking in directories, but it's not open by default anymore.

From Yan, I've install git-grep and mapped K to search for the word under the cursor, and that's been such a pleasant, and faster experience, than using vimgrep. Yan is also cleaning up a plugin for rspec integration that provides some nice wins.

Too Big?

But how far do I take this path? I could continue to add plugins and map commonly used keystrokes to further increase efficiency, but when do I start to get diminishing returns? For example, I could map a letter to :GitGrep and save six keystrokes, but I haven't found the need to do that yet. Usually when I'm searching for something, I'm thinking about what I want to search for while typing the command, so I'm effectively multitasking and not wasting time with all that extra typing. Sometimes slowing down can be a good thing to allow that planning. Plus each new mapping or plugin is something new to learn, and it can occasionally become overwhelming with all the new options.

Just Right, For Now

I'll definitely continue to improve my Vim Fu, but I'm not in a hurry to continually add to my repertoire until I've mastered what I currently have available. I've come this far on a basic setup, and Vim has been around long enough that there have been many others who have gotten by with less, so I'm not going to stress out about not optimizing every single key I type.

Pair Programming Feedback

It's been a few weeks of pair programming with the new guys, and it's been so enjoyable and valuable. Here's what I've experienced.

Impproved Toolset

Since there are multiple people working on the same computers, we need a shared toolset. Some of us like RubyMine, while others prefer the awesomeness of vim. Since everyone prefers one of these two, we can set them both up on each machine and switch editors as needed. Yes, it would be optimal for pairing to use the exact same environment and setup like Pivotal Labs, but you can pry vim out of my cold, dead hands (hard drive?).

It's easy to keep ones toolset static after getting used to that way of doing things, so getting others who have different processes allows us all to question our own and change those parts that are lacking. I've finally come around to Janus, a suite of vim customizations and plugins, and it's been great. Command-T is almost worth it by itself. Yan also introduced me to git grep, which has been much faster than vimgrep.

Improved Code

The value of the discussion-aspect of pairing becomes immediately obvious when there's a disagreement. Trying to argue ones position on a subject exposes how much one really understands about that subject. If I can't explain why we should use technique x to my pair, do I really know enough to say for sure that we should? It's a great didactic tool for all parts of the development process - pretty much anything that we would write is subject for debate, and the discussions are invaluable for coming up with the best solution.

More Fun

Having more people in the office has made the place much more lively and interesting. Since we're such a small startup, any new person is going to have a major impact on the physical space of the office and the dynamic of the group (going from four to five is kind of a big deal). Our new interview process seems to have worked because it's been a pleasure working with each new team member, and everyone gets along with everyone else.

Exhaustion

Pair programming is fun, but it's tiring to work so intimately and intensly with someone all day. The productivity has gone up a lot because of the focus, the breaks have become more separate (not just reading an article on the computer but getting up and moving away from the damn thing), and I'm definitely ready to stop coding by the end of the day. It's a good exhaustion, a feeling that much was accomplished and comes with a certain buzz that's different than when working by myself all day. It's a new stress, but I'll adapt to it and get even better.

We're in a good place, and we're ready to continue writing great software.