Everything is intertwingled
• 405 words • ~2 min read ⏲
I've wanted to replace the scripts with which I create this website for quite some time now. My most important goal was to be able to build the site with a single command. Generating the RSS feed turned out to be a bit of a tyranical edge case. Within the last week I got to the point where I am satisfied enough with the result to finally ditch the two old scripts start using it for this website.
For those who read my site via the feed, this is also the reason why some old articles might have popped in your reader up all at once. The old script was heavily depending on a certain regular structure of the directory, and I would have needed to run the rss generation script for each category and then somehow properly merge the resulting xml files, which I found to be too tedious, so the weeknote category of the site was left out (which was another driver for the rewrite).
There is a great xkcd about how much time you may spend on automation before yielding negative returns. If all the tedious stuff took about 5 minutes per post and I'll write once a week, about 21 hours of programming amortize themselves over the course of five years. But even if programming a custom SSG turns out to be uneconomically from the point of timesaving, it still was fun, which is enough reason to do it. I've put both the the generator, which I named intertwingle, as well as the source of my website on github.
On to a completely different subject. This week at work, I heard that upper management is looking into establishing tools that create metrics from our repositories and issue tracker to be reported to them. I don't know whether it is merely a conincidence that a month or so ago an article about measuring developer productivity by a big management consultancy made the rounds. Kent Becks response in two articles (Part 1 and Part 2) is quite good. In a debate on Linkedin Kent also delivered a soundbite, that sums it up up nicely: McK is offering false certainty to replace honest ambiguity
. Anyway, I think it is rather obvious why advice on software engineering from a company that mainly sells political capital and powerpoint presentations (even if the latter is Turing-complete...) needs to be taken with more than just a grain of salt.