4 ms·

Who gets to define merit? Say a certain style was more likely to pass code review. And that style was more often associated with people with CS degrees, vs self

Who gets to define merit? Say a certain style was more likely to pass code review. And that style was more often associated with people with CS degrees, vs self-taught programmers. Even though it's a stylistic choice, functionally equivalent, one is "better" than the other and is introducing bias. It's very easy to do, especially because consistency is important for legibility, but a stylistic choice nonetheless. I think a decent, practical solution to some degree are prettifiers/code-formatters and linters. Because the style is codified, and trivial for anybody to apply.

sanxiyn 8y ago
You should follow the project's style when you start contributing. If style is pointed out in code review, fix it.
- guitarbill 8y ago
  Well yes, I agree. It's a bit of a toy example, simplistic, but hopefully one we can all relate to. It does illustrate the issue of using "meritocracy" without anything else. It's like saying "we follow a style guide", and then not providing that style guide. Obviously sub-optimal. (The example breaks down since providing a style guide is doable, whereas providing a "merit guide" is hard/impossible? In that case, is it better to remove that? I don't know.)
bluecalm 8y ago
Just because it's hard to accomplish pure meritocracy doesn't mean we can't strive for it or have it our stated values. It has an advantage that we can reject all efforts which contradicting the principle and could possibly undermine the project.
derefr 8y ago
Usually calling something a meritocracy implies that it also has public bars for promotion that you can constantly self-test against and so iterate toward. A common example is a marital-arts dojo: every student can take a belt (rank) promotion test at any time, and as many times as they like; and it is crystal-clear what such a test entails (usually, being able to beat someone who is already at the given skill-level, with both participants using a defined subset of the taught skills that the target ranking "expects.") Thus, even without taking the official test, students can simply ask their peers who have reached that skill level to spar with them, and in-so-doing iterate toward being of that skill level themselves. For a more hypothetical example: a meritocratic public schooling system would be one where there're no "years" of education, but rather "ranks" (or just a big unified "tech tree" of topic-units to study), with tests to attain new ranks/unlock new topic-units; and where each student can take each test an infinite number of times (presumably with procedurally-generated tests that resist answer-memorization.) In such systems, you (ideally) will first very quickly equilibrate to the rank in the system that your initial level of merit allows you to reach, and then will continue forward through the system at the rate you're willing to grind to increase your merit.

sanxiyn 8y ago

You should follow the project's style when you start contributing. If style is pointed out in code review, fix it.

guitarbill 8y ago
Well yes, I agree. It's a bit of a toy example, simplistic, but hopefully one we can all relate to. It does illustrate the issue of using "meritocracy" without anything else. It's like saying "we follow a style guide", and then not providing that style guide. Obviously sub-optimal. (The example breaks down since providing a style guide is doable, whereas providing a "merit guide" is hard/impossible? In that case, is it better to remove that? I don't know.)

bluecalm 8y ago

Just because it's hard to accomplish pure meritocracy doesn't mean we can't strive for it or have it our stated values. It has an advantage that we can reject all efforts which contradicting the principle and could possibly undermine the project.

derefr 8y ago

Usually calling something a meritocracy implies that it also has public bars for promotion that you can constantly self-test against and so iterate toward. A common example is a marital-arts dojo: every student can take a belt (rank) promotion test at any time, and as many times as they like; and it is crystal-clear what such a test entails (usually, being able to beat someone who is already at the given skill-level, with both participants using a defined subset of the taught skills that the target ranking "expects.") Thus, even without taking the official test, students can simply ask their peers who have reached that skill level to spar with them, and in-so-doing iterate toward being of that skill level themselves. For a more hypothetical example: a meritocratic public schooling system would be one where there're no "years" of education, but rather "ranks" (or just a big unified "tech tree" of topic-units to study), with tests to attain new ranks/unlock new topic-units; and where each student can take each test an infinite number of times (presumably with procedurally-generated tests that resist answer-memorization.) In such systems, you (ideally) will first very quickly equilibrate to the rank in the system that your initial level of merit allows you to reach, and then will continue forward through the system at the rate you're willing to grind to increase your merit.