[Bf-funboard] UI talk + session at bconf
davidj at gmail.com
Thu Sep 26 05:26:32 CEST 2013
On Wed, Sep 25, 2013 at 5:30 PM, Brecht Van Lommel <
brechtvanlommel at pandora.be> wrote:
> Ok, that's a good suggestion. I do think in practice the separation
> between UI review and code review isn't so easy to make.
Can you explain why you think this, as I see them as easily separable.
A UI proposal is going to need feedback from people who deeply understand
blender code (to figure out what is practically possible), but it should
contain nothing more than the description of what would change (with
screenshots), how it would change, what other UI similar mechanisms exist,
and why this is a good solution. If we can't agree on this, there is no
point in looking at code.
Just as a trivial example to make the point, the "UI design review" for
this tiny trivial UI change I made (which was accepted!) does not require
looking at the patch code. You can see everything you need to decide
whether this is a good usability change merely from the descriptions and
screenshots... Whether or not the patch correctly *implements* this change
in an acceptable way is an entirely different issue -- and should be
handled during code review.
> The developer does need to understand the patch deeply, assuming the
> external developer did is not a good assumption in my experience :)
I'm sure both cases occur. My experience so far is the opposite.
I spent time deeply understanding the spline handle-set code, and the "just
bugfix" version of my patch is sitting in part because the committers I've
tried to get to look at it don't know that code. I was told on IRC this is
15 year old blender code, the implication being there is waryness of
touching it. For a committer to understand this patch as well as I did
when I wrote it, they'll probably have to spend the same amount of time I
spent learning, changing, and testing the code when I wrote the patch.
I don't mean that to come across as some kind of self-aggrandizing, as I've
only contributed a *stupidly* *small* amount of code to blender, and this
patch is also trivially small. I'm merely trying to make the point that for
a committer to understand that patch as well as I did when I wrote it is
basically going to cause them to duplicate the research effort which went
into me making the patch in the first place. Very inefficient.
More information about the Bf-funboard