Do your tools make sense?

A couple of days ago I was talking to a friend and he asked me what some of the fundamental things were that caused me to change jobs. After thinking about it for a while, I think I can start to answer it in reverse - what new things I have seen that I hope every organization I work for from here on out will have.

The first thing that came to mind is just doing things that make sense. This seems like it might be obvious and something that people do anyway, but I assure you, it is not. It's not to the extent that I named this very blog after that very thing - does having a cement cutting board that will dull your knives really make sense? Yet I used one every day for months.

One of the biggest tangible things I can think of is using tools that make our jobs as developers easier. Making our jobs easier = happy developers. We use IntelliJ instead of Eclipse since you get a complete polished package with awesome refactoring tools. We (mostly) use RubyMine instead of TextMate for the same reasons. We use git instead of subversion because you get a much more powerful feature set. We use github instead of our own server so we have less operations overhead. We use rails instead of Spring/Struts/Django/.NET/etc. because we feel it gets the job done easier and faster.

The list goes on and on, and it's always a work in progress, but that's what's so great about it. If a new tool comes along that we feel will significantly make a difference, I guarantee you it will show up on peoples' machines right away.

The point of this isn't to say "you need to use tool xyz", but simply to say that it feels great to work for a place that is willing to invest in a toolset that lets you simply get work done.

It always baffled me before when I was told that I couldn't use a certain tool that had a cost involved (often under $100) because it was too expensive. If that tool had saved me even half an hour of time total the company would have broken even at the rate they billed me for.

One of the biggest examples of this I can think of involved code reviews. At my last company we had always tried to code review everything but had trouble making sure we were diligent about doing them. One day I decided that we needed something better. I started researching code review tools and found one in particular that looked really promising. There was a 30 day trial so I got it up and running in about an hour and we started using it on my team with great success.

The issues started when we tried to convince Those With Money to spend $500/seat on this tool. I completely understand wanting to make sure that you invest your money wisely, but the battle that ensued over this tool, when it had rave reviews from any engineer that used it, was lengthy and ended up with everyone upset and exhausted by the end.

The funny thing is, if this tool could prevent everyone using it from writing one bug every 3 months the company would have broken even on the cost in the first year of using it. (These numbers are based on their calculations.) We had already seen results much better than that in the first 1 month trial, and had already expressed other benefits such as knowledge transfer, reduced time bringing someone new onto a team, and cross-team reviews to name a few. Yet still the battle was fought every day for weeks until a decision was finally made.

(By the way, if you are doing code reviews, I highly recommend using Code Collaborator. It is well worth the price tag and was called out by others as the single biggest improvement in our engineering process that year.)

Probably the craziest situation that I have ever seen happened at another company I was working with. They had a development team of about 10 people, and all the engineers were given mediocre laptops as their sole development machines. To top it off, they were running a heavy webserver (Weblogic), had only 2GB of memory, and had 5400 RPM hard drives with software encryption used on every read/write.

To give you some data points - publishing code changes to the Weblogic server on my dev machine, a fairly pimped out Mac Pro, took about 12 seconds. To publish changes on their laptops took around 7 minutes. Really. I timed it. Twice.

This means that _every_single_code_change_ they did took them 7 minutes before they could see and test it. It was amazing to see their management continually shoot down requests for new machines and then complain about how unproductive this team was!

I guess what I'm really getting at, is working for a company that honestly supports you in getting your work done is an amazing thing. Does it cost more money up front? Yes. Will it pay off in the end? Absolutely. Is it awesome to get to use tools, languages, and frameworks that the developers like? Oh yeah.

So, until next time, keep your toolkit close and your knives sharp.