I am an impatient man (if my sporadic typos, and last minute edits late at night on my blog didn’t make it obvious already). Recently this has started to stick out to me much more, especially in regards to programming. More often than not I would find myself just writing code haphazardly, then let the compiler/interpreter point out the errors for me, some not so obvious, others more obvious than I would have liked. This is a rather reckless approach to take when it comes to development. I find that it prevents me from actually thinking about what I am doing, and results in me producing tools and code that are in unsatisfactory state.
Take mgrt for example, just today I added some small features to it. Nothing to special, just storing the revision queries in the database and better supporing MySQL. And that was what stuck out to me. The support for MySQL in this tool was completely broken. It failed to parse the timestamps from the database, and didn’t even have SSL supported properly either.
This came down to two things, the first being that I developed this tool for myself, and I’m more of a PostgreSQL person, and the second being that I was impatient with the development of this tool. I never took the time to properly test it using the different databases it claims to have supported. There were other things that were poorly implemented too, such as the SSL support for MySQL. I have fixed these today, however it still bothers me that I saw fit to release this tool, to little fanfare, in such a broken state. Had I taken my time with it, then this could have been avoided. Even the feature that I added today, whereby the queries are now stored in the database, is something I initially considered when I started writing it. However I neglected it due to my impatience.
Going forward, I do not expect to be the sort of person who will write code that will compile/run on its first time. But it’s an exercise I would like to practice. Write some code, or make some modifications to a code base, but before compiling or running the program I should stop and think about the changes I’ve made instead of letting the compiler/interpreter hold my hand.
I realise that this post doesn’t have much substance, and in retrospect might come off as being common sense. But, the fact that I showed off mgrt to some people in such a broken state really bothered me. I’m dissapointed that I didn’t hold myself to a higher standard. On the bright side, I appear to be the only one using this tool so the only thing damaged was my self-esteem. No doubt if someone else used it for a MySQL database they would have opened up an issue on the GitHub repository in no time.