This post is apropos to a recent topic "PR process killing morale and productivity" [1].
The Google code readability program is talked about very little outside of Google, and as far as I know, is replicated nowhere outside.
Within Google, it is a cult. There is never any discussion or action taken to walk any part of it back. The bar simply gets ever higher, year after year. The C++ style guide itself is massive and authoritarian. Are you working on an embedded system? Too bad, you must follow best practices meant for high volume production services to the T.
When I got C++ readability 4 years ago, it felt like a hazing. It was an utterly opaque system, accountable to no one, where a random readability reviewer had a 10% chance on each review of insisting that you make some major refactor of code you didn't write.
Even your level of progress was inscrutable, being measured in little "+"'s next to multiple dimensions. Occasionally, you'd get another "+" and wonder when you'd finally be done. And then one day, after many months, thousands of lines of C++, hundreds of review comments, the readability program hive mind opaquely decided that your hazing was done, and you are now free to write whatever code you want, whether it adheres to the style guide or not.
Some languages at Google have more than a year of backlog to even enter the program to start getting readability reviews and make progress towards readability. The answer is always to "add more readability approvers", yet the bar for that is even higher, and the time demands of being such an approver are incompatible with most teams at Google.
In any case, I don't intend to litigate Google's readability program in a public forum. But I provide it as an example of how "style guides" can and will go too far and consume far more resources than necessary. And how there is naturally no review process or escape hatch to roll back excess in this area.
The Google code readability program is talked about very little outside of Google, and as far as I know, is replicated nowhere outside.
Within Google, it is a cult. There is never any discussion or action taken to walk any part of it back. The bar simply gets ever higher, year after year. The C++ style guide itself is massive and authoritarian. Are you working on an embedded system? Too bad, you must follow best practices meant for high volume production services to the T.
When I got C++ readability 4 years ago, it felt like a hazing. It was an utterly opaque system, accountable to no one, where a random readability reviewer had a 10% chance on each review of insisting that you make some major refactor of code you didn't write.
Even your level of progress was inscrutable, being measured in little "+"'s next to multiple dimensions. Occasionally, you'd get another "+" and wonder when you'd finally be done. And then one day, after many months, thousands of lines of C++, hundreds of review comments, the readability program hive mind opaquely decided that your hazing was done, and you are now free to write whatever code you want, whether it adheres to the style guide or not.
Some languages at Google have more than a year of backlog to even enter the program to start getting readability reviews and make progress towards readability. The answer is always to "add more readability approvers", yet the bar for that is even higher, and the time demands of being such an approver are incompatible with most teams at Google.
In any case, I don't intend to litigate Google's readability program in a public forum. But I provide it as an example of how "style guides" can and will go too far and consume far more resources than necessary. And how there is naturally no review process or escape hatch to roll back excess in this area.
[1] https://news.ycombinator.com/item?id=42274003