From: Jonathan Gordon <jdgordy_at_gmail.com>
Date: Wed, 15 Sep 2010 23:55:14 +1000

On 15 September 2010 23:39, Rafaël Carré <rafael.carre_at_gmail.com> wrote:
> 2010/9/15 Jonathan Gordon <jdgordy_at_gmail.com>:
>> When is it read? users only read it when it happens (or soon (<6m)
>> afterwards ) to see what development they might have missed, probably
>> looking for keywords of interest to them. Devs (speaking from my
>> experience here) do the users thing, but when the message is actually
>> important is when you are trying to track down a possible breaking
>> change, and even then (as I said a few replies ago) the message isnt
>> even the first thing I look for.) And then when I find a possible
>> change I look at the diff before rereading the message (because I know
>> it cant possibly explain everything).
>
> It happens frequently to me that I'm reading the full log of a particular file
> or folder (with or without the diff), and especially when it's an area that I
> don't know.
>
> Please think that all your logs can and will be read in years from now and
> any info which could help the future devs to understand the code is nice.
>
> It's quite frequent that you are trying to understand why a line of code was
> committed 8 years ago and you wish the committer had explained better
> what he did. (ok perhaps not very frequent but it did happen to me and
> other people in different projects)
>

This is another problem entirely which is impossible to fix unless we
require commit message reviews. enabling log changing isnt going to
fix it either (except for when it is blatantly wrong and completely
accidental). This is no different to trying to write the actual code
so it is understandable 8 years later. "you" understand it now, and
"you" will understand it then so whats the big deal, right?
Received on 2010-09-15