Comment by knorker
10 days ago
So you're saying there's no right and wrong decision on whether to make your own OS because Meta has not decided whether it's a public corporation with fiduciary duty or a kindergarten for OS hobbyists to have unproductive wasteful but fun projects in?
So that's like saying there's no right or wrong answer on whether to leave your 15th floor apartment a normal morning via the stairs, or by jumping out of the bathroom window. Yes, actually, there is. Unless your goal is to go splat.
"Will this make the product better", "will this make the product deliver sooner", "will this be a net positive to the company's bottom line" are all objective, though hard, questions.
> Very related to https://en.wikipedia.org/wiki/Is%E2%80%93ought_problem
Nonsense. The company already has many many oughts already. Otherwise words mean nothing, and the "right answer" to an interview question about hashmaps is to strip naked and start humping the trash can.
And on my tennis analogy, it's objective that one way or the other allows a gives player to win more games. So we can call it "better". Even though you can still argue whether you "ought" to try to win.
> So you're saying there's no right and wrong decision on whether to make your own OS because Meta has not decided whether it's a public corporation with fiduciary duty or a kindergarten for OS hobbyists to have unproductive wasteful but fun projects in?
No, because it's subjective whether your interpretation of "fiduciary duty" precludes them from developing an OS. I think it does not, for example.
> Unless your goal is to go splat.
Yeah, goals are subjective too.
> Otherwise words mean nothing
Words do mean nothing, they are pointers to ideas. And for you "objective" points to a different idea than it does for most people.
> No, because it's subjective whether your interpretation of "fiduciary duty" precludes them from developing an OS. I think it does not, for example.
We can disagree in our best assessment if "developing an OS" is a good strategy for achieving our goals.
But no, that is not subjective. There's a right answer. We can try to use our experience and place subjective weights on probabilities, but if we could A/B history, there's definitely an objective right answer.
Boiling down "building an OS" to the math of spending dollars on OS development and the return on this investment in dollars, this is objective.
In your estimation of how much it'll cost, you turn to experts who use objective and subjective resources to try to estimate the true objective numbers.
> for you "objective" points to a different idea than it does for most people.
Me, the dictionary, and the world disagrees with you.
> It is very much not an objective discussion if you are discussing whether it makes sense to develop a new operating system.
I don't know what kind of discussions you have, but if you were in the room would you be saying "I just have this gut feeling that we need to develop an OS"?
I would hope that instead you would be making objective arguments such as "The cost of this OS development, because we control the hardware and will need drivers for video, audio, bluetooth, and inertial sensors, is (based on similar projects here) about $X in labor cost. The benefits are (just making this up) battery times because we won't need a write path and cache coherence on the storage system, which per this benchmark consumes Y% of clock cycles. And this should cause our product to be better to the tune of $Z based on these previous product statistics"
Now, an objective counter argument could be that the cost isn't $X, it should be $X2, because the people working on this have an opportunity cost. And there's another cost $W, which is delayed time to market. And $X in any case would need to be compared with patching the Linux kernel for the needed features.
And so on, and so on.
It's also an objective discussion to appeal to your experience, or experience of others: "Look, when company X, Y, and Z did this, the eventually abandoned that approach. It never worked for those companies. Why would it instead work for us? How is the problem space different?".
A subjective discussion would be "I just feel like it'll be cleaner if we had our own OS, and that should allow us to leapfrog competition", being countered with "well Linux is battle tested and just works".
"I don't think anybody cares about battery times" — subjective. "Game Boy Classic was good for 30h play time, whereas Game Boy Color only for 10. Here's sentiment analysis on the Game Boy forum indicating with so-and-so error bars that people did care for that use case, and people returned their puchases indicating battery life at a rate of so-and-so higher for Color" — Objective. — "Well, we're not making Game Boy" — Subjective.
If you're having the latter discussion, then you're not actually having a discussion whether it makes sense to develop a new operating system. Whatever you decide will just have the rationale "it felt right".
Having seen Carmack's talks and writings, it's my assumption that his case against writing an OS was not "just trust me bro, I'm experienced". It was likely painfully specific both about the technical complexities, and how those are connected to the bottom line (the end goals).
Your argument seems to be one against the idea of expertise. Experts need to be able to communicate objectively, but the other end of the spectrum where it appears you are, is that "it's all subjective, your data is as good as my guess, there's no such thing as truth".
Subjective is of course not wrong. My examples above may imply that, but just because a visionary can have a subjective vision that predicted the future correctly, doesn't make objective counters not objective. Just objectively wrong.
"You have a one-in-a-billion chance to win the lottery" — "Well I felt I would and I did!" — "Congrats. I was not wrong though".
> Words do mean nothing, they are pointers to ideas.
Are you a philosophy student? Seems a bit entry level.