On 04.02.2015 12:31, Patrick Georgi wrote:
Am 2015-02-03 11:38, schrieb Carl-Daniel Hailfinger:
The current development guidelines didn't appear overnight, but were refined over time without a lockdown on the wiki.
And yet they're mostly useless.
The CoC wikipage is now unlocked. Let's see what comes of it.
Thanks!
I changed the wording a bit, reordered some stuff and replaced one of the homegrown definitions with the one used by the UN which is hopefully not objectionable.
There is another pending proposal (not by me) which IMHO would be a nice addition:
https://www.coreboot.org/Talk:Code_of_Conduct#Some_words_about_assessing_cod...
Some words about assessing code quality
Proposal: Add to the list in the chat etiquette something like: 'Code that you are changing wasn't perfect (or you wouldn't change it). However, try to avoid assuming that it was written by monkeys, or that it must have always been broken. It's likely that it used to work and it's likely that your new work is necessary and important because circumstances changed where the code didn't.'
Pros: Remind people that there's a coder behind the code. Assuming bad intent or stupidity (and doing so publicly) is as much a statement about the code as about the coder.
Cons: Is that micromanaging things?
The "Mailing list and chat etiquette" section of the CoC already has two related items:
- People who intentionally insult others (users, developers,
corporations, other projects, or the coreboot project itself) will be dealt with. See policy below.
- We are dealing with hardware with lots of undocumented pitfalls. It
is quite possible that you did everything right, but coreboot or its tools still wont work for you.
Admittedly some code has always been broken and nobody may have noticed before, but I really do like the "avoid assuming it was written by monkeys" part. Quite a lot of code in coreboot (especially hardware-specific stuff) has been copy-pasted and changed until it worked well enough. Some cleanups have removed code of that type, but there probably still is such code that remains. In flashrom, we do notice from time to time that the remaining really old code was written in an age where some classes of bugs simply weren't widely known and thus the coder could not have known how to avoid them. Sometimes, enough code was written the wrong way to have zero net effect. Yes, head-scratching and exclamations of WTF?!? happen whenever we try to fix or clean up that code. Still, back then the code contribution (patch) seems to have been considered a net positive, otherwise it wouldn't have been merged. For me, this is mostly a question about respecting our elders who didn't have our modern tools and still got things to work.
Opinions?
Regards, Carl-Daniel