First, a brief note on how I hold the convictions I'm about to recklessly pontificate upon.
I hold my ideas about how to actually live Catholicism well, rather strongly. This is because the memory of my trials in Carmel provides me with the same kind of conviction I experienced when I presented to my engineering teammates the tests I had run on my brand new algorithm before sending it to production: "You have no reason to trust that my idea is good on its own, but there's proof it can withstand some rigor."
Unfortunately, I think it's pointless to try to detail the level of dedication under suffering my faith had to withstand to stay in Carmel, and stay Catholic since then. I'll have to leave it to your imagination. Nonetheless, if I could convey those validations in a meaningful way, I happily would.
But anyway, speaking of tests.
One reason I'm skeptical that gradient Catholicism is the only real kind is how that gradient is manifested: the Catholic test suite, if you like.
My problem with the Catholic test suite hinges on this Biblical concept: "Thus sayeth the Lord God: 'Stop acting like your shitty tests are helping with or really telling you anything about how well the code is running.'"
In non-Kathleen translations of these and similar verses, "your shitty tests" are visible signifiers of faithfulness (sc. "sacrifices and burnt offerings" or "the law" in "the power of sin is the law"), while the code is a man's mind and heart.
Basically, there exists an emphasis among gradient-Catholics that we should have not only good code, but test coverage: very often, all the test coverage.
In this view, to the degree a virtue can't be exposed to the evidence-based process we learned to love so well in Logic 101 -- to the degree it isn't externally verifiable -- it risks irrelevance. If its unit test is failing, then it's for all intents and purposes nonexistent. This method prioritizes the surface-ability of virtue, the provision of a strong logical interface. According to this view, unless there is a clear and rational excuse not to, code should be provable by means of an agreed-upon interface: ✓ : success, ✗ : fail.
Anyone who has written a program even a few files big knows that having tests is good and important for sustainable future development. They also know that to aim for 100% coverage, and craft code to optimize for this ahead of flexibility and performance, would actually be totally horrible and ruin the program.
Why? I mean, I think you can probably imagine it. All I'm saying is, in Carmel I felt viscerally how the cost of building and correcting the output of perfect (or sometimes any) tests just isn't worth it under some conditions. This sense matched beautifully with the one I found again years later among a good engineering team under pressure.
After all, tests can show all sorts of results whose significance is permanently in doubt. Meanwhile, being the entirely separate concern that it is, the code runs. And it can run flawlessly. I'm certain this insight is broadly known among those who have stressed their Catholicism under demand for a certain type of extraordinary, constant loving action: I'm thinking primarily of single mothers or fathers, or parents of disabled children.
In Catholicism, we were taught to optimize for test coverage -- external signifiers of virtue -- by high, ancient authorities, and by our mothers and fathers. I now firmly consider this to signify a good effort tarnished by one other hidden, essentially animal, and effectively ruinous motivation: to be able to communicate among ourselves the passing state of our code: to reveal the vital ✓.
✓ and ✗ is survival. ✓ and ✗ is acceptance and inclusion in the herd, or not. It's old, it's ancient of days; it's an instinct without which the human race could never have persevered through countless winter starvations. Our subconscious minds apply it to non-faith communities and faith communities with the same primal simplicity and force. My mother greeted the friend-pattern of one nun's habit, and guarded me from the no-pattern of another nun's beige pants, by literal instinct.
But this instinct is outdated. We no longer need to stay accepted in our tribe to survive. In the past 100 years or so, sanitation and technology have lifted us a whole new distance away from the terror and heartbreak of death that once defined our lives on earth so much more uniformly.
Once, we had no choice but to weave the imperative ✓ and ✗ into our Catholic design. Today, we have a choice. And that's why it's just plain wrong to insist upon keeping it as much of a design priority as it ever was. It ruins Catholicism in exactly the same way this mindset would be expected to ruin a computer program.
If it's so obvious that this design pattern is flawed, why don't most gradient Catholics figure that out?
That's my next post.
I like the analogy of complete test coverage to the law, and how adherence to the law was a sign of visible submission to the community just as how adherence to the complete coverage was a sign of complete submission to the corporate coding manual, albeit neither necessarily being productive of the most happiness / best code.
ReplyDeleteI'll go further, I think. If the tests are a sign of submission, it is *best* if they hurt. It means that you're doing them just for doing them. You can trust someone who obeys the Law to keep doing it, because it's such a pain that if they weren't committed they wouldn't.
Like, someone who cuts off their brother bc they are gay? That must hurt; we can trust them to follow the rest of the law. It's like killing someone to join a gang, or covering up a police misconduct before being in the "real" club, or lying about something where some group is obviously wrong before the group lets you in.
It's a general instance, I think, of costly signals being prereqs for group membership. At least if I'm picking up what you're putting down.
---
What's more confusing to me -- if Carmel is a test, *what* does the test tell us? What constitutes passing? What constitutes not passing? What do know because of a pass or fail, exactly? I'm still unclear about this.
1. If Carmel was a test of Catholicism, what about the Catholicism was it testing? You come through at the end and know that the Catholcism is -- what? What is being tested?
Oops included draft bit
ReplyDeleteOh cool, I'm glad you liked it! Yeah, I think you get my drift pretty well.
ReplyDeleteHmm, I think I could have been clearer about the nature of the Carmel-test. The "test" in this context was not so much the test-suite variety, with specific outputs, as mentioned elsewhere in this post. The output is a more generalized kind of pass/fail: what I had in mind was the kind of test I and my teammates performed all day one day on that algo of mine before sending it in production, a big, fun, cooperative effort to do everything we could think of to break my little baby (not that I'm attached or anything).
To switch analogies, if I were designing (instead of code) a bridge of straws, rubber bands, and masking tape between two school chairs, Carmel was the part where you pile all the books in the classroom on top of it. If you come out the other end with an intact bridge, pass. If not, fail. The Carmel result was the bridge didn't collapse, i.e. Catholic.sh didn't crash. The weight of books on the bridge was the weight of sadness that fell solely on the shoulders, so to speak, of my underlying assent to Catholicism as true. (That's not to say that's literally the point of Carmel; just one effect. Shit got tested.) So to finish/modify your sentence, "I come through at the end and still think that Catholicism is true." Does that make sense?
Yep!
ReplyDelete"I come through at the end and still think that Catholicism is true."
Agreed that this makes sense; the experience of Carmel did not dissuade you from thinking that Catholicism is true.
I this test pretty much is about your psychological state, though. As we discussed earlier: if you go through boot camp, exit, and think the Marines are great, it tells you about your psychological state; it doesn't tell you about the Marines. But we both agree with this anyhow.
Yeah, I think we agree.
ReplyDeleteJust a note to anyone viewing: I did update this post quite a bit in light of the above comments, to try for greater clarity on that topic.