On Wed, Dec 21, 2005 at 12:10:43PM +0000, Alistair Crooks wrote:
> We're getting into human factors here, so some points to consider:
> 
> 1. the original diff author (Doug McIlroy?) could have used "+-" but
> chose "><" for a reason
> 
> 2. the visual difference between '>' and '<' is much greater than
> that of '+' and '-', making >< changes much more visually
> obvious at a glance. In addition, there is very little visual
> difference between the '+' and '-' characters, leading to
> much more concentration needed to differentiate old from new.

I changed the format because gdt@ requested it, since he finds the
+- format easier to read. It seems there are different views
colliding here. Perhaps other people want to provide their opinion
on the subject? After all, I read the diffs when I create them, so
I don't really care either way.

> 3. unified diffs were designed to keep context, but compress the
> amount of text in the resultant diff (I believe I read that
> somewhere, I could be wrong). In these examples, context
> does not help information dissemination - it hinders it.

Actually, neither the previous nor the current version contains
any context, so this is a moot point, since nothing has changed
in this respect. Or do I misunderstand the point?

Cheers,
 Thomas