August 20, 2008.
Several months ago--lets take a stab and say five--I was working on the backend for my blog. Lots of little details being hand crafted by an absent-minded but well-intentioned craftsman. Today, I present one of the smallest, most enduring, and perhaps endearing of these absent-minded decisions.
Not long after a long time ago I began running into an irritating error. I'd graduate an article from the draft bin to the edited bin. I'd glance at it, check to make sure no one was watching, and immediately graduate it to the published bin (I feel guilty, because I was--for several moments--so certain that I was going to actually start proofreading my material. It turns out that now I don't even spell check everything closely because Safari doesn't spellcheck as well as Firefox and I don't have a spellchecker installed in Emacs. Quality control at its lowest, indeed). So, what I should have said is that I'd tried to graduate it to the published bin. Because, instead of blue success, I got the really helpful (and really red) generic error message for situations I wasn't specifically watching out for:
Your request made it to the server, but failed for unknown reasons.
Well, I thought to myself, fuck. Then I clicked the button a couple more times and stared at the rejection notices stacking up at the top of the screen. Then I did what many a confused programmer has done: I copied and pasted things one at a time into a new copy of the article and tried to get it to publish. Paste in the text. Does it publish? Yep. Select the tags, flows and author1, then try to publish. Yeah, still working. Add a title... ah, damn, that kills it. What the hell is wrong with this title?
Does it have some special character that is causing problems? Okay lets strip out everything except for letters. Grr. It doesn't work. Try it with a new title. Now it works? At that point, I was mumbling obscenities under my breath. What the hell was going on? Then I remembered the key to this puzzle: I had arbitrarily capped the length of titles at ninety-five characters in the code that verifies an article is publishable. I went back to the code and looked at it. Yep. Capped at 95. So I did the most logical thing.
I started keeping all my titles under ninety-five characters.
To this day, the editor will flash red failure at me whenever I create a title that goes over the limit. It would take about a minute to fix the bug, but before I get there I start deluding myself about how the limitation inspires creativity, and how structures forces us to think in new directions. So, the ninety-five character limit endures for a bit longer.
That is a boldfaced lie, the author functionality didn't work properly five months ago, but lets not be picky. Actually, the whole tense of this article is a lie. I rediscovered this bug--having completely forgotten about it in the meantime--at least four times.↩