A Former Science Club Member Critiques Google, THE Science Club

August 17, 2020

I truly enjoy posts from former insiders at giant technology monopolies. Each of them hires from the other. The meta-revolving door spin is fascinating to watch. Some get tossed out of the mechanism because of some career negative factor: Family, health, mental orientation, or some other exogenous, non-technical event. Others go through a Scientological reformation and realize that the world of the high-technology nation-states is a weird place: Language, food customs, expectations of non-conformist “norm” behavior, and other cultural suckerfish. What gives me a chuckle are revelations like Tim Bray’s or Steve Yegge’s “Dear Google Cloud: Your Deprecation Policy is Killing You.” In my opinion, Mr. Yegge’s thoughtful, calm, and “in the moment” essay about the GOOG is more intriguing than the Financial Times’ story that reports Google has predicted the end of the world as we know it in Australia. What? Australia? Yep, for those receiving this warning from the Oracle at Mountain View their life amidst the kangaroos will mean no “free search” and — gasp! — curtains for “a dramatically worse” YouTube. Search can’t get much worse, so the YouTube threat means angry kids. Yikes! YouTube. Will Australia, a mere country, at the wrong end of a Google phaser strike?

Back to Mr. Yegge: In his essay, the phrase “deprecation treadmill” appears. This is the key insight. Googlers have to have something to do it seems. The bright science club members interact via an acceptable online service and make decisions informed by data. As Mr. Yegge points out, the data fueling insights may not be comprehensive or processed by some master intelligence. He notes that a Bigtable storage technology had been running for many years before any smart science club member or smart Google software noticed. (So much for attention to detail.)

Mr. Yegge points out that

One is that running a Bigtable was so inconsequential to Google’s scale that it took 2 years before anyone even noticed it, and even then, only because the version was old. As a point of comparison, I considered using Google Cloud Bigtable for my online game, but it cost (at the time) an estimated $16,000/year for an empty Bigtable on GCP. I’m not saying they’re gouging you, but in my own personal opinion, that feels like a lot of money for an empty [censored] database.

This paragraph underscores the lack of internal controls which operate in real time, although every two years could be considered near real time if one worked in a data center at Dialog Information Services in the mid 1980s. Today? Two years means a number of TikToks can come and go along with IPOs, unicorns, and Congressional hearings live streamed.

Mr. Yegge also uses a phrase I find delicious: “Deprecation treadmill.” The Google science club members use data (some old, some new, and some selected to support a lateral arabesque to a hotter team) to make changes. Examples range from the Dodgeball wackiness to change in cloud APIs which Mr. Yegge mentions in his essay. He notes:

Google engineers pride themselves on their software engineering discipline, and that’s actually what gets them into trouble. Pride is a trap for the unwary, and it has ensnared many a Google team into thinking that their decisions are always right, and that correctness (by some vague fuzzy definition) is more important than customer focus.

I wish to point out that Mr. Yegge is overlooking the key tenet of high school science club management methods: The science club is ALWAYS right. Since Mr. Yegge no longer works at the Google, Mr. Yegge is WRONG. Anyone who is not a right-now Googler is WRONG. A failure to understand the core of this mindset cannot work at the Google. Therefore, that individual is mentally unable to understand that Mother Google’s right-now brood is RIGHT. Australia and the European Union, for example, do not understand the logic of the Google. And they are obviously WRONG.

How simple this is.

Mr. Yegge points out how some activities are supposed to be carrier out in the “real world” as opposed to the cultural norms of the techno-monopolies. He writes:

Successful long-lived open systems owe their success to building decades-long micro-communities around extensions/plugins, also known as a marketplace.

This is indeed amusing. Google delivers advertising, and that is a game within a casino hotel. I think the online advertising game run by the Google blends the best of a 1950s Las Vegas outfit on the Strip and the mythical Hotel California of the Eagles’ song. Compare my metaphor with Mr. Yegge’s. Which is more accurate?

Google’s pride in their software engineering hygiene is what gets them into trouble here. They don’t like it when there are lots of different ways to do the same thing, with older, less-desirable ways sitting alongside newer fancier ways. It increases the learning curve for newcomers to the system, it increases the burden of supporting the legacy APIs, it slows down new feature velocity, and the worst sin of all: it’s ugly. Google is like Lady Ascot in Tim Burton’s Alice in Wonderland:

Lady Ascot: Alice, do you know what I fear most?

Alice Kingsley: The decline of the aristocracy?

Lady Ascot: Ugly grandchildren.

Mr. Yegge’s point is a brilliant one: The Google wants its customers to operate like “here and now” Googlers. But customers do not understand and they are WRONG.

The disconnect between the Google and mere customers is nailed in this statement by Mr. Yegge:

But after all these years, Google Cloud is still #3

Yes, Google does make decisions based on data. Those decisions are RIGHT. If this seems like a paradox, it is obvious that the customer is once again proving that he or she is not capable of working for Google. Achieving third prize in the cloud race is RIGHT, at least to some real Googlers. For Mr. Yegge, the crummy third place ranking is evidence of the mismatch between the techno-monopoly and cloud users and, I might add, Australia and the EU.

Mr. Yegge points out what may be a hint of the tension between the Google science club and its “wanna be” members. He writes about a Percona-centric, ready-to-use solution. He calmly points out:

Go ahead, I dare you. Follow the link and click the button. Choose “yes” to get all the default parameters and deploy the cluster to your Google Cloud project. Haha, joke’s on you; it doesn’t work. None of that [censored] works. It’s never tested, starts bit-rotting the minute they roll it out, and it wouldn’t surprise me if over half the click-to-deploy “solutions” (now we understand the air quotes) don’t work at all. It’s a completely embarrassing dark alley that you don’t want to wander down. But Google is straight-up encouraging you to use it. They want you to buy it. It’s transactional for them. They don’t want to support anything.

DarkCyber looks forward to Mr. Yegge’s next essay about the Google. Perhaps he will tackle the logic of reporting an offensive advertisement to the online monopoly. That process helps one understand a non-deprecation method in use at the A Number One science club. The management method is breathtaking.

As Eugène Ionesco noted:

“Realism falls short of reality. It shrinks it, attenuates it, falsifies it; it does not take into account our basic truths and our fundamental obsessions: love, death, astonishment. It presents man in a reduced and estranged perspective. Truth is in our dreams, in the imagination.”

The “our” is Google’s reality. If you are not a “here and now” Googler, you cannot understand.

Stephen E Arnold, August 17, 2020


Comments are closed.

  • Archives

  • Recent Posts

  • Meta