Sharp: Closing a door [LWN.net]

Sharp: Closing a door

Posted Oct 6, 2015 1:04 UTC (Tue) by ssmith32 (subscriber, #72404)
In reply to: Sharp: Closing a door by jhardin
Parent article: Sharp: Closing a door

But it is pretty much useless. How about - this code triggers an oops, under commonly encountered condition X so we need to fix. Saying it's crap is just as useless as the management speak. I'm really tired of people defending lazy and annoying comments by saying they're marginally more useful than other lazy and annoying comments. I disagree that vague, hyperbolic and hostile comments improve things more than just plain vague comments.

to post comments

Sharp: Closing a door

Posted Oct 6, 2015 3:20 UTC (Tue) by bronson (subscriber, #4806) [Link] (2 responses)

If only the real world were quite so simple. Some code is so sweeping and nebulous that you just can't reason logically about it, short of writing a book anyway. Email-sized arguments therefore get boiled down into emotionally-charged aesthetic observations. devfs, X, capabilities, Mir, HAL, systemd, etc etc etc.

Agreed, vitriol and hype are useless and should just embarrass the author. But sometimes lazy arguments are the best we can do with the time that we have.

This is not to excuse lazy arguments, of course. We can and should do better. I just wouldn't like to see anyone err too far on the other side and demand quantitative proof before removing a difficult subsystem.

(hm, speaking of nebulous arguments boiled down into unsubstantiated emotional response...)

Sharp: Closing a door

Posted Oct 6, 2015 3:58 UTC (Tue) by ssmith32 (subscriber, #72404) [Link] (1 responses)

Then you say this code is so large and difficult to follow, it's time to start over. Still a lot clearer than "crap". Crap code could mean anything.

Sharp: Closing a door

Posted Oct 6, 2015 4:33 UTC (Tue) by bronson (subscriber, #4806) [Link]

But lots of merged code is large and difficult to follow. Scheduler, I/O, VM, DRM, ...

Sharp: Closing a door

Posted Oct 6, 2015 18:27 UTC (Tue) by dskoll (subscriber, #1630) [Link]

I agree with this. I quite often review code produced by developers who work for me and I have never and would never say "this code is crap." I'd say something like "this code can fail in situation XYZ" or "this code needs refactoring because there's cut-and-paste" or whatever... but never anything insulting.