It is a good day when Steve Yegge has a new rant to read.
Yegge is a veteran software engineer whose career runs strangely parallel to mine. We overlapped for a short time at Amazon in the early 2000’s, and a few years later at Google. More recently we both worked for companies enabling mobile payments in Asia. We’re both opinionated bloggers (each of whom has name-dropped the other), we’re both Emacs partisans, and we’re both anguished by how Google’s technical superiority is matched by utter cluelessness in product design and marketing.
Where Yegge outshines me by far is in his entertaining, informative, impassioned, and dead-on-accurate rants. His most famous one is probably his Platforms Rant, which was meant to be Google-internal only but made headlines when it was posted publicly by mistake. In that one he implored Google to invest more effort into making its products, which were increasingly “walled gardens” with inflexible feature sets dictated by competitors, into platforms that would allow others to build onto them, the way Amazon was doing. This rant came in the early days of Google+, when many of us within Google were expressing concern over its product design and the lack of any useful APIs that would allow an open ecosystem to develop around it. Ironically, his rant was a Google+ post, and it was the product design, in part, that led to its being misposted publicly. Also ironically, Google+ is now dead—arguably from the very causes Yegge and I and others identified back then—taking his Platforms Rant post with it. (However, it’s preserved in other forms around the net; just google [yegge platform rant].)
In his latest rant he again improves on one of my own frequent refrains: that Google keeps giving you shiny new things and then keeps yanking them away. Like me, he’s a user of Google Cloud Platform products; like me, he is increasingly frustrated by how often those products require you to rewrite your own code to adapt to Google’s changes; and like me, he is entertaining abandoning Google Cloud Platform for this reason, in favor of the more stable (if less technically excellent) Amazon Web Services platform.