Clan x86

General Forums => Entertainment District => Topic started by: Lead on August 27, 2008, 01:25:58 PM

Title: truth about code review's
Post by: Lead on August 27, 2008, 01:25:58 PM
(http://i33.tinypic.com/bhdqgh.jpg)

agree / disagree?
Title: Re: truth about code review's
Post by: iago on August 27, 2008, 02:09:47 PM
Agree.

Wait, that domain is blocked at work, so all I see is a 'X' :)
Title: Re: truth about code review's
Post by: Camel on August 27, 2008, 02:47:29 PM
Quote from: iago on August 27, 2008, 02:09:47 PM
Wait, that domain is blocked at work, so all I see is a 'X' :)

Top:
The only valid measurement of code quality: WTFs/minute

Left:
Door labeled "code review" with two speech lines coming from it with "wtf" at the end. Below the door is the annotation, "good code"

Right:
Same as left, with extra more "WTFs," and the annotation reads "bad code"
Title: Re: truth about code review's
Post by: Hdx on August 27, 2008, 07:51:19 PM
Agree, But its sad most the code I deal with goes on the right.
Title: Re: truth about code review's
Post by: Sidoh on August 27, 2008, 08:18:07 PM
I don't think I've ever attended a formal peer review.

At work this summer, we had "inspections", but the process was much more well-defined.  Reviewers would mark defects before the meeting independently, the person would either fix them or make a comment saying the issue should be discussed during the meeting.  It kept out the long, drawn out debates that result in pretty much nothing.  It was a fairly painless process.  The meeting for my code, which was about 10k lines, only took 45 minutes -- and that was with the principle engineer there. :)
Title: Re: truth about code review's
Post by: iago on August 27, 2008, 10:24:31 PM
Quote from: Sidoh on August 27, 2008, 08:18:07 PM
I don't think I've ever attended a formal peer review.

At work this summer, we had "inspections", but the process was much more well-defined.  Reviewers would mark defects before the meeting independently, the person would either fix them or make a comment saying the issue should be discussed during the meeting.  It kept out the long, drawn out debates that result in pretty much nothing.  It was a fairly painless process.  The meeting for my code, which was about 10k lines, only took 45 minutes -- and that was with the principle engineer there. :)
Heh, we had something similar at Symantec, where the document I wrote would go to an editor, then come back for changes, then we'd discuss it all. But, there was one difference -- you weren't allowed to argue with the editor, no matter how wrong he was (especially when it came to technical stuff). I tried to decline a change once, and he came to my desk to yell at me (even though he was clearly wrong, to the point where somebody changed it back during the final review.. justice! :) ).
Title: Re: truth about code review's
Post by: Sidoh on August 28, 2008, 12:40:10 AM
Yeah, I suppose that sort of makes sense since its his job to correct, pretty much.  The meetings we had were all engineers and the occasional manager who was mostly there to make sure everyone was working, lol
Title: Re: truth about code review's
Post by: Joe on August 30, 2008, 04:47:52 PM
Quote from: HdxBmx27 on August 27, 2008, 07:51:19 PM
Agree, But its sad most the code I deal with write goes on the right.

:)
Title: Re: truth about code review's
Post by: Hdx on August 31, 2008, 07:07:54 PM
Quote from: Joe on August 30, 2008, 04:47:52 PM
Quote from: HdxBmx27 on August 27, 2008, 07:51:19 PM
Agree, But its sad most the code I deal with write goes on the right.
:)
I'm getting better.
And 90% of the WTF?s are because I don't comment. AT ALL.